The industry is debating the wrong question

Walk into any corporate travel conference this year and the argument you will hear is about autonomy. Should an AI agent be allowed to rebook a flight on its own? Change a hotel? Approve an expense under a certain threshold? Entire panels are built around where to draw that line.

It is a reasonable debate. It is also premature by several years for most of the companies having it, and a survey published this year makes that uncomfortably clear.

GBTA research conducted in March 2026 with 269 travel buyers across the United States, Canada and Europe found that only 12 percent of travel programs have a consolidated view of their own program data from a single source. The same study found that 58 percent of buyers say AI has had little or no impact on their programs so far.

Those two numbers are not independent findings. The second is a consequence of the first.

Corporate travel data consolidation is not a prerequisite the industry has been slow to address. It is the thing the industry has skipped entirely while moving on to a more interesting conversation about agents.

What the same survey wants

The gap becomes stark when you look at what those buyers say they want from AI.

In the same GBTA study, 92 percent expressed interest in predictive analytics for spend forecasting.

Another 89 percent wanted automated disruption management and rebooking. Eighty five percent wanted AI-powered traveler support, and 83 percent wanted conversational booking.

Every one of those capabilities depends on the system having a reliable, current, unified picture of the program. Spend forecasting on fragmented data produces a forecast of the fragments. Automated rebooking without knowing a traveler’s actual itinerary, policy tier and preferences produces a rebooking somebody has to fix manually.

The study also found 63 percent citing lack of consolidated reporting as a specific pain point, and 61 percent saying managing travel across regions is a challenge. Buyers are not unaware of the problem. They are describing it accurately and then asking for capabilities that require it to be solved first.

Where the foundation actually cracks

If the data foundation is thin, you would expect the tools built on it to work well in the simple case and fail in the complicated one. That is precisely what a second study found.

Christopherson Business Travel surveyed 123 US travel management, procurement, operations and finance professionals for a white paper called The Booking Breakpoint. Only 20 percent said their booking tool supports their company’s travel needs very effectively. Just 11 percent said more than three quarters of trip modifications get handled successfully through self service. And during unplanned disruption, the moment a travel program most needs to work, only 10 percent said their tools support travelers very effectively.

Read those three numbers in order and they describe a specific failure shape. The tool works at the point of purchase and degrades from there. Booking is a single transaction against clean inputs. A modification requires knowing what was already booked, what policy allows, what the traveler is entitled to, and what alternatives exist. Disruption requires all of that at once, under time pressure, across multiple suppliers.

Christopherson’s CEO Josh Cameron framed the underlying issue in a way I think is exactly right: travelers experience one connected trip, while the systems serving them treat air, hotel, car and support as separate transactions. The traveler’s experience is continuous. The data is not.

The loop this creates

Here is where the two studies stop being separate observations and start describing a cycle.

When a tool fails a traveler during a modification or a disruption, the traveler does not file a complaint and wait. They solve their own problem, using whatever channel works.

Every one of those trips leaves the managed channel. Which means the program loses the data from it. Which makes the consolidated view thinner. Which makes the tools worse at handling the next modification. Which sends more travelers outside the channel.

The industry calls the result a shadow travel stack, and usually frames it as a compliance problem. It is worse than that. It is a self-reinforcing loop where each rotation degrades the thing everyone says they want to build AI on top of.

The part nobody connects

There is a second consequence of travelers leaving the managed channel, and it has nothing to do with data quality.

I wrote recently about a breach at a group of UK airports in which the records of roughly 8.7 million customers were accessed. The detail that mattered was where the data came from: not a booking tool, not a travel management company, but airport WiFi sign-ups, car park reservations, lounge access and fast track purchases. Every one of those is a transaction a traveler makes personally, outside any system their employer contracted or reviewed.

So a traveler who goes around the managed channel is doing two things simultaneously. They are removing a record from the program’s data foundation, and they are creating a record somewhere the company has no contract, no retention schedule, and no visibility.

Data quality and data security are usually owned by different people, budgeted separately, and discussed in different meetings. The shadow stack is one problem producing both, and it is not being managed as one problem anywhere I have seen.

What doing this properly looks like

There is a useful counterexample, and it is worth studying precisely because it is unglamorous.

When Business Travel News named Moderna’s Jennifer Steinke its 2026 Travel Manager of the Year for building an AI system her travelers use daily, the coverage focused on the AI. The more instructive part of the story sits earlier in the timeline. Before any vendor was involved, in October 2024, she manually uploaded a year and a half of program data into spreadsheets. Only after that foundation existed did she bring in a data specialist to normalize it and feed it back through an API.

Evan Konwiser of Amex GBT made a related point in a Forbes interview with Jeff Fromm: the breakthrough interface gets the attention, but the durable value usually sits in the plumbing underneath. Proprietary context, clean data, and the infrastructure to act on both in real time.

There is also a security argument for getting the foundation right rather than layering agents onto a fragmented one. Researchers from Stanford and Georgia Tech showed in work presented at ACL in 2025 that AI agents completing web-based tasks, including travel booking, can be manipulated by content designed to fool a machine rather than a person. An agent operating on incomplete context has fewer internal checks against being wrong, and fewer still against being deceived.

The commercial reading of all this

I want to note something about the three studies I have cited, since it affects how they should be read.

The GBTA research was produced in partnership with a booking platform, a hotel group and a travel management company. The Christopherson white paper concludes that booking technology and service need to work together, and Christopherson is a travel management company that sells service alongside technology. Each study’s conclusion lands comfortably on its sponsor’s product category.

None of that makes the underlying data wrong. Vendor-funded research is a large share of what this industry knows about itself, and refusing to read it would leave us knowing less. But it does mean the useful move is to take the numbers and set aside the conclusions.

I should apply the same standard to myself. I cofounded VOLL, a corporate travel and expense management platform, so I have a direct commercial interest in companies concluding that their data foundation needs work. I have written before about how the real cost of a travel program hides in the places that never reach an invoice, and this is the largest example of that I know.

The regional gap in all of it

One last observation about all three studies.

The GBTA sample covers the United States, Canada and Europe. The Christopherson sample is entirely American. The Amadeus sample is American. Latin America appears in none of them.

That means the consolidation figure for this region is simply unknown. It could be better than 12 percent, and there is a reasonable case that it might be: Latin America has repeatedly adopted new travel and payment technology faster than the markets that invented it, often skipping intermediate stages other markets are still working through.

But that is an argument, not a measurement. The most useful thing I took from three days at a global industry convention this year was how much of what the industry treats as settled turns out to be a US finding that traveled. Somebody should measure this one directly, here, before the region inherits a diagnosis that was never about it.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Not a lobbying operation

Most people outside Latin America, when they hear “industry association,” picture a lobbying vehicle. An organization funded by suppliers, arguing supplier positions to regulators, occasionally hosting a conference where the sponsors outnumber the buyers.

Alagev, the Latin American Association for Corporate Travel and Events Management, is built on a different premise, and the difference is structural rather than rhetorical. Founded in 2003, it is a nonprofit that deliberately puts corporate travel managers and their suppliers inside the same room, on the same footing, as a single membership rather than two constituencies with opposing interests. Membership is free for travel and event managers, which is not a marketing gesture. It is the mechanism that keeps the buyer side of the room from being priced out of it.

I sit on Alagev’s board, and I serve on its technology and innovation committee. I want to be direct about that bias before going further, because everything below is written by someone with a stake in the organization succeeding. What follows is still worth writing, because the thing Alagev turned out to be best at is something almost no one outside Brazil realizes exists.

The unglamorous work that turned out to matter most

Trade associations usually measure themselves by events, memberships, and the occasional policy win. The most consequential thing Alagev does is considerably less exciting than any of that.

In partnership with FecomercioSP, the São Paulo state federation of commerce and services, the association produces the LVC, a monthly survey measuring what Brazilian companies actually spend on business travel. Not a projection, not a sentiment index. Actual monthly spend across airfare, lodging, vehicle rental and ground transport.

That series now runs long enough to say things nobody could say about this market a decade ago. Brazilian corporate travel closed 2025 at 147.8 billion reais, the largest figure in the sector’s history, up 6.3 percent on the prior year. January 2026 came in at 12 billion reais, a record for a month that is traditionally the industry’s quietest. March reached 18.2 billion, putting the first quarter near 48 billion. By the end of May, the year had accumulated close to 85 billion reais, up 6.2 percent against the same stretch of 2025.

Two numbers that do not agree, and why that is fine

There is a wrinkle worth being honest about, since it comes up whenever someone tries to size this market.

The LVC measures domestic spend on travel services, in reais, month by month. A separate global measure, the Business Travel Index produced by the GBTA, sizes Brazil differently: it projected the Brazilian market at 35.8 billion dollars for 2026, roughly 183 billion reais, ranking Brazil among the world’s ten largest corporate travel markets, with growth of 13.8 percent. That figure was presented at the opening of the GBTA convention in Chicago this year, and it made Brazil the fastest-growing market among the world’s fifteen largest, ahead of Japan at 10 percent and both the United States and Germany at 6.7 percent.

Those two numbers do not reconcile cleanly, and they are not supposed to. They use different scopes, different currencies, and different methodologies for what counts as corporate travel spend. The useful move is not to pick the bigger one. It is to understand what each is good for. The LVC tells you the shape of the market month to month from inside Brazil. The BTI tells you where Brazil sits relative to everyone else.

Both matter, and until recently Brazil had reliable access to neither. I have written before about how the best global research in this industry is openly honest about where its sample thins out, and Latin America sits squarely in that thin zone. Alagev’s contribution has been to build the domestic half of that picture, month after month, for years, so the region stops being estimated from a distance.

What the numbers are actually saying right now

Reading the LVC series alongside other Brazilian sector data produces a picture that is more nuanced than the headline growth suggests, and Alagev’s own leadership has been notably careful about not overselling it.

Executive director Luana Nogueira has pointed out that a meaningful portion of recent growth is being driven by rising prices rather than by more trips being taken, which she has described as a less ideal scenario than genuine demand expansion.

Jet fuel cost increases pushed airfares up sharply earlier this year, and some companies have responded by shifting shorter routes to ground transport and vehicle rental without reducing the underlying need to travel.

The lodging data points the same direction. Hotel occupancy in Brazil moved from 62.4 percent in May 2025 to 63.4 percent in May 2026, close to flat, while the average daily rate rose 5.4 percent over the same period. That is the same pattern I have written about on the airline side: revenue growing through price rather than volume.

An association that only wanted good headlines would report the record and stop. The fact that Alagev’s own executives volunteer the less flattering reading of their own data is, to me, the clearest evidence the organization is doing the job properly.

Why the technology committee, specifically

I chose the technology and innovation committee deliberately, and the reason is a risk I think this market is running without talking about it enough.

Artificial intelligence is arriving in corporate travel management faster than most travel programs can absorb it. The companies with dedicated technology budgets, internal data teams, and vendor relationships will absorb it fine. The concern is everyone else: the travel manager operating alone inside a mid-sized company, with no analyst, no data engineer, and no realistic way to evaluate whether a given AI tool is useful or noise.

That is not an abstract commitment. Alagev’s education arm runs a program called Primeira Classe, bimonthly masterclasses for travel managers across Brazil, built in partnership with VOLL, the corporate travel and expense management platform I cofounded, which is also one of the association’s supporting members.

The subject matter runs from pricing and competitive dynamics to governance and supplier ethics. The format is deliberately practical rather than promotional, because a masterclass that turns into a product demo teaches nobody anything and costs the association its credibility.

What longevity actually buys

Twenty three years is a long time for a trade association in a market that was, for most of that period, treated as a cost center rather than a strategic function.

The value of that longevity is not sentimental. It is that the LVC series exists at all, that it has enough history behind it to distinguish a genuine trend from a seasonal blip, and that when the association publishes its outlook for the year ahead, it is drawing on a record long enough to be worth reading rather than a guess dressed up as a forecast.

It also buys the thing that is hardest to build and easiest to lose: enough trust from both sides of the market that a buyer and a supplier will sit at the same table and say what they actually think. I have written about what three days at a global industry convention taught me about how much of this industry’s real progress happens in rooms rather than in press releases.

Building and maintaining those rooms, for two decades, in a region where the market itself was long dismissed as peripheral, is the actual achievement.

Brazil is now among the ten largest corporate travel markets in the world and, by at least one credible global measure, the fastest growing of the major ones. That did not happen because an association willed it. But the fact that anyone outside Brazil can now verify the claim, with numbers rather than assertion, is very much because one existed.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Finally, an actual conversation

I had seen Jennifer Steinke present at GBTA conventions in Dallas, in Atlanta, in Denver, always from a few rows back, always taking notes. This year, in Chicago, I finally sat down and talked with her directly.

You notice it within minutes of being near her. There is a cognitive sharpness that is obvious even in casual conversation, and a depth of knowledge about corporate travel that goes well past program mechanics into how the entire industry actually works underneath. You also feel the weight she carries. Running the travel program for a company the size of Moderna is not a small job, and it shows in how carefully she talks about the decisions she has made.

I caught her on a panel that week, alongside another travel manager and a vendor building AI for travel agencies, discussing where artificial intelligence is actually heading in corporate travel. Both women on that stage were sharp, and the exchange was one of the more substantive conversations on AI I have heard at an industry event this year. One line from Steinke stuck with me. As I recall it, she said something close to this: change is inevitable, and she does not want to be caught flat footed by it. She would rather already have her own tests, her own validations, her own experience in hand by the time agentic AI adoption becomes something every program has to do.

That is not the language of someone reacting to a trend. It is the language of someone who has decided the trend is coming regardless of her opinion of it, and who would rather spend the runway before it arrives building evidence than spending it forming an opinion.

What she has actually built

The scale of what is behind that attitude became clear when Business Travel News named her its 2026 Travel Manager of the Year at an award ceremony in Chicago in early August, the same city where I finally met her weeks later. The system she built, which she calls Tami, for Travel and Meetings Intelligence, sits behind a natural language interface and coordinates three separate partners: an AI-first booking tool called BizTrip.ai, a data platform called Cerebri.ai, and a mid-size travel management company, Cain Travel, that handles fulfillment.

BizTrip.ai itself launched in July 2025 at the GBTA convention in Denver, one of the very conventions where I had watched Steinke present from the audience. Skift’s coverage of the launch noted that Moderna was already piloting the tool at launch, alongside travel management company Cain Travel, with backing from Andrew Ng, the Google Brain cofounder. Steinke was not an early adopter reacting to a finished product. She was in the room helping shape it before most of the industry had heard the company’s name.

The concept she built on top of that infrastructure is what she calls Policy of One. In an earlier profile, BTN described how the system assigns real monetary value to the perks tied to a traveler’s loyalty status: what lounge access is worth against a meal budget, how upgrade probability factors into whether a slightly pricier fare is actually the cheaper choice once the loyalty math is done, how a baggage fee gets absorbed by status. BTN’s vice president of content, Elizabeth West, put it simply while presenting the award: Steinke has not just implemented a program, she has provided a model for the industry.

What that model replaces is the blunt instrument most travel programs still run on, the lowest logical fare rule that technically saves money on paper while quietly making experienced travelers resent the program enforcing it. Steinke’s system can show a traveler, and their finance team, exactly why a fare that looks two hundred dollars more expensive is actually the better total-cost choice once status benefits are priced in.

The parts that do not make the highlight reel

The version of this story that circulates in industry recaps tends to stop at “she built an AI travel assistant.” What actually happened underneath that headline is less flashy and more instructive.

Steinke did not start with a vendor pitch. She started in October 2024 by manually uploading a year and a half of program data into spreadsheets, building the raw material for what she wanted Tami to become before any AI vendor was involved. Only once that foundation existed did she bring in a data specialist, Cerebri AI, to curate and normalize it, feeding it back into Tami through an API. Cerebri’s own CEO, Jean Belanger, has said the goal was giving the system access to up-to-the-minute data, since an agentic booking tool is only as good as the data foundation underneath the conversation layer.

That foundation now does things a standard travel policy cannot. When a Moderna traveler books a flight to a destination they rarely visit, say Warsaw, Tami can compare the fare against Moderna’s own historical average for that route and flag it if it comes in unusually high, suggesting an alternate day or a connecting flight instead. Steinke has described herself as a believer in internal benchmarking over generic market rates, on the logic that what matters is what a trip should cost this company, based on this company’s own travel pattern, not an industry average that has never seen Moderna’s data.

The system also reaches into hotel sourcing in a way most managed travel programs do not attempt. Moderna has two primary hotels near its Cambridge headquarters, one within walking distance and a second, cheaper option requiring an Uber or shuttle. Rather than simply mandating the cheaper property, Steinke worked with that hotel to build in extra incentives, additional loyalty points, food and beverage perks, so Tami can show travelers exactly how much they personally save the company by choosing it, alongside what they gain for choosing it. Kelvin Woo, director of national sales at the Hyatt Regency Boston Cambridge, one of Steinke’s supplier partners, described the relationship as one where she has been explicit about what she needs and the hotel has learned to anticipate it before being asked.

There is also a small human detail buried in the BTN coverage that says something about how the tool actually lands with travelers. As the pilot expanded past its initial twenty users toward the more than two hundred using it today, Steinke noticed travelers referring to Tami by name, the way you would talk about a colleague. One traveler told her Tami had become one of her best friends, and specifically praised being able to drop a conference agenda or an email into the system and have it handle the booking without further instruction. That is a strange thing to hear about a piece of enterprise software, and it is also exactly the outcome Steinke was building toward: an assistant good enough at understanding context that using it stops feeling like operating a tool.

She is not building this alone in the industry

It would be easy to read Steinke’s award as a story about one unusually capable person outpacing everyone else in her field. The same ceremony that named her Travel Manager of the Year also recognized three Best Practitioners, and looking at what they were rewarded for makes clear this is a broader inflection point in corporate travel, not a single outlier.

Google’s head of travel, Rob Bray, spent the past year working through the technical detail of integrating a large, complex travel program onto a new joint platform from Amex GBT and Concur, pushing on content and NDC connections to bring consumer-grade booking parity to a corporate tool. McDonald’s global travel manager, Koreen Barbara, took the opposite approach from consolidation, deliberately keeping two travel management partners rather than one, so she could place a majority of the program on newer, innovation-forward platforms while those tools were still maturing toward full global capability. Adobe’s Kathy Burdge inherited a contract with a travel platform still facing industry skepticism and, according to the vendor’s own head of travel marketplace, effectively became the enterprise voice pushing that product toward the reporting and workflow improvements that made it credible for larger, more complex programs.

None of those three stories is really about AI as a feature. They are about travel managers willing to sit inside a vendor relationship long enough, and specifically enough, to shape the product rather than simply adopt it. Steinke’s version of that same instinct just went further, because she had already spent nearly a decade building the internal discipline, the data, the sense of what her travelers actually needed, that made her ready to co-build a tool from the ground up rather than wait for one to arrive finished.

The throughline nobody mentions

What most coverage of Steinke’s award misses is that this is not a sudden pivot toward artificial intelligence. It is the same three priorities she has been building around for close to a decade, executed with tools that finally exist.

Back in 2017, when Steinke ran the corporate travel program at Dycom Industries and won a different industry recognition, the pillars she had already organized her program around were safety, service, and savings, with a duty of care communication network built specifically to protect travelers during disruption across dozens of company locations. There was no AI available to her then that could do what Tami does now. What she had instead was a clear model of what a travel program owed its travelers, built years before the technology existed to fully deliver on it.

That is the part of her story I find most instructive, more than the AI itself. The technology changed. The underlying judgment about what a good travel program protects did not.

Why this matters beyond one company

I have written before about the difference between augmentation and automation, a distinction I picked up hearing about at Stanford this year: augmentation is AI that extends human judgment, automation is AI that replaces it. Steinke’s system is a clean example of the former. The traveler still decides. The system just makes the tradeoff visible in a way it never was before, in real time, at the exact moment the decision is being made.

That has a direct implication for how a company should think about building or buying its own version of this. The hard part was never the language model. It was building a data foundation clean enough, and a loyalty and policy structure detailed enough, to feed it something worth acting on. Steinke spent years accumulating that foundation before the interface existed to make use of it.

Every regional dataset I work with at VOLL, the corporate travel and expense management platform I cofounded, points to the same lesson from a different angle. Latin American companies adopt new travel technology quickly, often faster than their US counterparts, but the data discipline underneath it tends to arrive later. Steinke’s decade-long build suggests the order matters more than the speed. The data foundation has to come first, or the AI layered on top of it has nothing real to work with.

This is already showing up in Brazil, not just in theory

The same pattern Steinke is proving out at Moderna is starting to show up, on a smaller scale, among corporate clients in Brazil. At VOLL, we launched a marketplace of narrow, single-purpose AI agents last year called VOLL SmartHub, covered at launch by Brazil Journal, a piece of sponsored content we published rather than independent reporting, so I want to be upfront about that distinction. One of those agents, RatesAudit, continuously tests whether hotels are actually honoring the negotiated rates a company signed, since a rate that looks locked in on paper can quietly disappear from availability without anyone noticing until someone pays full price.

Cogna Educação, one of the clients running RatesAudit, published its results through the Brazilian trade outlet PANROTAS: 92 percent of its bookings now issue online, a traveler satisfaction score of 93, and more trips taken without an increase in budget. None of that is Steinke’s scale of system-building. It is a single narrow agent doing one job continuously that no finance team was ever going to check by hand. But it is the same underlying bet: that the useful version of agentic AI in corporate travel is not a general assistant, it is a specific, auditable worker assigned to the one blind spot a human team could never watch closely enough.

Back in Chicago

I have written before about what three days in Chicago taught me this year, and about a number that ended up on a screen there. Meeting Steinke in that same city, after years of watching her from a conference audience, felt like a fitting place to finally close the loop.

Watching someone build toward a vision for a decade before the tools exist to realize it is rare enough to notice when it happens. Watching them then win the industry’s biggest individual award for finally getting to use those tools is rarer still. Steinke’s answer to a technology arriving faster than anyone expected was not to catch up. It was to already be ready, because she had never stopped building toward it.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Two industries, the same headline

Every airline earnings call this year has featured some version of the same sentence. Delta reported that in both the first and second quarters of 2026, revenue from premium seating matched revenue from economy sales, a threshold that would have sounded implausible a few years ago.

American Airlines said lie-flat and premium economy seats grew more than twice as fast as Main Cabin seats in the first quarter. Qantas posted premium cabin revenue growth of 15 percent for fiscal 2026, twice the pace of its economy cabin.

None of that is happening in isolation. A Brazilian hotel industry report published this month shows a parallel pattern on the ground. In the FOHB’s July data, covering a sample of 570 hotels across the country, the Upscale category was the only segment with positive occupancy growth, up 0.5 percent, while Economic and Midscale both lost occupancy. Average daily rate rose in all three categories, but the direction of demand only moved one way.

Put the two industries side by side and the same sentence describes both. The top of the market is filling up. The rest is not.

What airlines are actually reporting

The scale of the shift in aviation is easier to see once you separate revenue growth from unit revenue growth, which is what the airlines themselves are doing on these calls.

Former Delta president Glen Hauenstein put the change in blunt terms last year: premium products used to be loss leaders, offered as free upgrades to loyal flyers, and are now the highest margin products the airline sells. Delta’s June quarter results back that up directly, with premium revenue up 17 percent year over year, a full point in front of the airline’s overall unit revenue growth.

American’s first quarter filing goes further on the mechanics: the airline is retrofitting existing aircraft with more lie-flat and premium economy seats precisely because that capacity is outperforming Main Cabin, not because premium travelers are simply paying more for the same number of seats.

Analysis from McKinsey, cited in recent aviation coverage, found that on the busiest transatlantic routes, business class tickets alone can generate revenue close to what the entire economy cabin brings in, despite occupying a small fraction of the aircraft’s floor space. That is the commercial logic behind every retrofit announcement this year. A business class seat that used to be a courtesy is now closer to the product an airline is built around.

What the hotel data adds

The FOHB numbers describe something narrower than the airline data, a single month in one country, but the direction matches closely enough to be worth taking seriously.

Occupancy fell 1.4 percent in the Economic category and 1.7 percent in Midscale in July, while Upscale gained 0.5 percent. Average daily rate rose in every category, so this is not simply a story about hotels charging more across the board. It is a story about where the additional demand is actually landing.

One caveat is worth stating plainly. The FOHB sample mixes corporate and leisure demand, and does not isolate business travel specifically. I am not claiming this is a corporate travel data point in the way the airline figures more directly are, since airlines report managed corporate revenue as its own line. What the hotel numbers add is a second, independent confirmation that premium demand is broad enough to show up in an entirely different industry, measured by an entirely different methodology, in the same month.

Why this is happening on both sides at once

The most common explanation for the airline shift is post-pandemic behavior change: travelers who experienced a premium cabin once, often through the price collapse and free upgrades of 2020 and 2021, found it difficult to go back. That explanation is well supported for aviation. It does not, on its own, explain why a Brazilian hotel report would show the identical pattern in the same month, since hotel guests were never handed the equivalent of a free business class upgrade during the pandemic in the same systematic way.

A more useful explanation covers both industries at once. Corporate travel budgets absorbed a real cost shock this year. Domestic airfare in Brazil, for a Monday morning departure, now runs 59.4 percent above the identical route on a weekday afternoon, and I have written before about how Brazilian airfare stayed expensive even when input costs should have worked in the country’s favor. Faced with a base cost that keeps climbing regardless of category, a traveler or a travel manager choosing where to spend the marginal amount of budget has a real incentive to spend it on the seat or the room that delivers the most usable time back, rather than spreading a smaller increase evenly across every trip.

That reframes premium growth as a rational response to cost pressure rather than a simple upgrade in taste. If the base fare or base room rate is rising anyway, the relative cost of stepping up to premium narrows, and the productivity argument for doing so on a long flight or an important stay gets easier to make internally.

What this means for a travel program

For a company negotiating corporate rates, this pattern changes what the negotiation is actually about.

The traditional target of a corporate travel policy is the discount on the base fare or the standard room rate. If demand and pricing power are concentrating at the premium end, the base rate discount is negotiating over a shrinking part of the total spend picture.

The practical response is not to ban premium travel, which tends to fail quietly through executive exceptions anyway. It is to make the tradeoff explicit rather than implicit: which routes or stay lengths justify the premium cost on productivity grounds, and which do not, decided in policy rather than discovered after the fact in an expense report. I have made a version of this argument before about how the real cost of a travel tool is what happens after the invoice. The same logic applies to premium cabin and room policy. The sticker price is the smallest part of the decision.

A pattern still missing its Latin American data

The airline data in this piece comes almost entirely from US and Australian carriers reporting in US dollars. Every regional dataset I have worked with at VOLL, the corporate travel and expense management platform I cofounded, suggests Latin American travel behavior diverges from that baseline often enough that it should not be assumed to track it here.

The FOHB hotel data is the first piece of evidence in this specific pattern that comes from inside the region rather than being extrapolated from the US market. It is one month, one country, and one industry, which is not enough to call a regional trend. But it is a start, and it points to a question worth someone answering directly with airline data from Brazilian and Latin American carriers: is the same premium migration happening in the seats, not just the rooms.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Where the number actually comes from

If you have sat in enough conversations about expense management, you have heard some version of this claim: an expense report costs a company around 58 dollars to process, and considerably more when it has to be corrected.

That number has a specific origin. It comes from a study the GBTA Foundation published in partnership with HRS in October 2015, based on 533 travel managers surveyed globally that September and October. The scope was narrow and clearly stated: the time and cost of processing an expense report for a business trip with an overnight hotel stay.

A report that goes through cleanly costs 58 dollars. A report that comes back costs 110 and takes nearly forty minutes of human time. That gap between the clean path and the corrected path is where the real money sits, and it is the part most companies never quantify.

The aggregate is the part that gets attention

The study also found that companies process an average of 51,000 expense reports a year. Applying the error rate to that volume produces the figure that made the study travel: about half a million dollars and close to 3,000 hours spent annually, per company, doing nothing except correcting mistakes in paperwork that documents decisions already made.

Three thousand hours is more than a full time position. It is a job nobody posted, nobody interviewed for, and nobody defends in a budget meeting, because the cost is spread thinly enough across enough people that it never lands on a single line.

The study also noted something counterintuitive that rarely gets quoted alongside the headline: larger companies, with more employees and higher annual spend, took less time per individual report than smaller ones. Scale helped here, which is not the usual direction for administrative process.

The eleven year problem

Here is what bothers me about how this number gets used, including by people in my own industry.

It is from 2015. The fieldwork happened in September and October of that year. Every time someone cites 58 dollars per expense report in a 2026 pitch deck, they are citing a survey conducted before most of the current expense platforms existed in their present form, before mobile receipt capture was standard, and before instant payment reshaped how business travelers actually pay for things.

I am not arguing the number is wrong. I am arguing that nobody knows, because the industry has largely stopped measuring it. A figure gets repeated until it feels like a fact, and the repetition itself becomes the reason nobody goes back to check.

That is a strange state of affairs for a cost that, by the study’s own math, runs to roughly half a million dollars a year at a single large company. It is the kind of number that should be re-measured every few years, and instead it has been quoted for a decade.

Why the Brazilian version of this problem looks different

Payment behavior in Brazil has moved faster than the process built around it, and that gap is where a lot of the friction comes from.

Looking at payment data across more than 850,000 monthly users on our platform, 64.6 percent of transactions on the corporate prepaid card are now Pix (digital instant wires). Only 35.4 percent still run on credit. Nobody ran a change management program to make that happen. Business travelers simply paid the way they already pay in their personal lives, and the corporate card followed.

The timing of that spend follows the clock of a meal rather than the clock of a workday. Close to 60 percent of everything paid happens between lunch and dinner, and the most frequent merchant category is restaurants, followed by gas stations.

Put those two facts together and you get a specific kind of reconciliation problem. Payment is instant, informal, and happens at the exact moment someone is eating or filling a tank. The documentation of that payment happens days later, at a desk, from memory. A process designed around a physical card and a printed receipt was never built for a payment method that clears in three seconds at a lunch counter.

I have written before about how Latin America adopts fast and formalizes late. Expense reporting is the clearest example I know of. The payment layer here leapfrogged. The reporting layer did not.

The cost that never appears on the invoice

Every conversation about expense software eventually becomes a conversation about license price. That is the number that goes into the procurement comparison, and it is the smallest number in the equation.

I made a version of this argument about the corporate travel ROI problem: the tools that look cheapest on a vendor invoice are frequently the most expensive once you count what it costs internally to make them work. Expense reimbursement is the purest case of that math.

A company can save forty thousand reais a year on a cheaper platform and lose several times that in reconstructed receipts, bounced approvals, and the finance analyst hours that go into chasing them.

The relevant question is not what the software costs. It is what a clean report costs versus a corrected one, and how many of each the company produces per month. Most companies cannot answer that, because nobody measures the second half.

Where AI helps (and where it does not)

The obvious move is to point AI at this problem, and it is a reasonable move. Categorizing a receipt, matching a transaction to a trip, flagging a line that falls outside policy, these are exactly the kinds of narrow, repetitive tasks where automation earns its cost quickly.

But the Amadeus “Tailored Horizons” survey found something worth pausing on. Seventy six percent of business travelers say they trust AI to assemble a travel option for them. Only forty four percent trust it to spend within the company’s budget. That thirty two point gap sits precisely on the boundary between recommending and committing, which is also the boundary this problem lives on.

There is a research reason for that caution too. Researchers from Stanford and Georgia Tech showed, in work presented at ACL in 2025, that AI agents completing web based tasks can be manipulated by content designed to fool a machine rather than a person. An agent that automatically approves reimbursements is an agent that someone eventually has an incentive to fool.

So the useful framing is narrower than full automation. AI can collapse the thirty eight minutes by removing the reconstruction work: pulling the transaction, matching it to the trip, reading the receipt, pre-filling the report before the traveler ever opens it. What it should not do, at least not without a check, is silently approve the spend at the other end. The cost being attacked here is the correction loop, not the approval itself.

Evan Konwiser of Amex GBT put it well in a Forbes interview with Jeff Fromm this month: the goal is fewer manual steps, not a new interface to learn. In expense management that means the report should mostly be finished before anyone sits down to write it.

What this changes for a travel program

The practical implication is a shift in what a travel and expense program measures.

Most programs measure spend: how much went to air, how much to hotel, how much above policy. Almost none measure process cost: how many reports required correction, how long each took, what that time is worth. A program that only measures spend will keep optimizing the visible number while the invisible one grows quietly underneath it.

The companies that get this right tend to start with an unglamorous exercise. Count the reports. Count how many came back. Multiply the corrected ones by what a corrected report actually costs. The number is usually large enough to fund the fix several times over, which is the part that surprises people. This is rarely a case of needing to justify an investment against an uncertain return. The return is already being spent, just in a form nobody put on a slide.

Thirty eight minutes does not sound like a crisis. Multiplied across a company that produces a few thousand reports a month, it is a full time job that nobody hired for and that produces nothing except documentation of decisions that were already made.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Where the data actually came from

The detail that matters most in the Manchester Airports Group breach is not the number of people affected, though that number is large. It is where the data was sitting.

The BBC reported this week that criminal hackers accessed records belonging to roughly 8.7 million customers of MAG, the group that owns Manchester, East Midlands and London Stansted airports, and demanded a ransom the company says it refused to pay. The compromised records included contact details, postcodes and vehicle registration numbers.

None of that came from a booking engine, a travel management company, an expense platform, or a corporate travel program of any kind. The bulk of it was email addresses collected when passengers signed up for WiFi inside the terminals.

The more detailed records, the ones including vehicle registrations, came from people reserving car park spaces, arranging lounge access, and buying fast-track through security.

The exposure surface nobody maps

Ask a travel manager to describe their program’s data risk and you will usually get an answer about the booking tool, the expense platform, and the TMC. Those are the systems under contract, the ones with a security questionnaire on file and a data processing agreement someone in legal reviewed.

Now list what actually happens on a single business trip that touches none of those systems. The traveler connects to airport WiFi using their work email address, because that is the address they have memorized and because the sign-up form asks for one. They book airport parking through the airport’s own website, entering the license plate of a company vehicle. They buy lounge access at the gate because the connection is tight. They pay for fast-track because the security line is long and the meeting is not moving.

Each of those creates a record, held by an entity the employer has no contract with, containing information the employer would classify as sensitive if it lived anywhere else. A work email address is a directory entry. A postcode plus a name is enough to identify a person. A vehicle registration tied to a corporate fleet is an asset record.

MAG stated that no passenger safety or aviation security was compromised, and that the affected system held no banking or payment details. Both statements appear to be accurate, and both are also beside the point for a company assessing its own exposure.

Payment data is the easiest category to protect and the fastest to remediate, because a card can be cancelled in an afternoon. An email address tied to a name, a home postcode and a vehicle cannot be reissued.

The industry already lived the rehearsal

Corporate travel has a recent, uncomfortably direct precedent for what happens when the information layer fails, and it is worth revisiting because it was not a breach at all.

The travel industry felt this in an almost literary way. The Global Business Travel Association’s annual convention opened on 22 July in Atlanta, Delta’s home city, with the airline still working through the aftermath days later. Many of the roughly 5,800 business travel professionals expected at the event did not make it in time for the opening session. The people whose job is managing corporate travel disruption spent that week personally stranded by one.

Nobody was hacked. No data was stolen. The lesson was about dependency: how much of modern travel runs on a small number of systems, and how little visibility any single company has into the resilience of the layers beneath its own.

Read alongside an incident where data actually was taken, the two form a complete picture of what information security means for this industry. One end of the spectrum is confidentiality, what gets out. The other is availability, what stops working. A travel program is exposed on both, usually through infrastructure it does not own and cannot audit.

Why the stakes rose with AI

There is a specific reason this class of breach has grown more consequential, and it has to do with what the stolen data enables rather than what it directly reveals.

Phishing used to be a volume game. Generic messages, obvious errors, a low hit rate that still worked because the cost of sending was near zero. What a dataset like this one provides is the raw material for something considerably more targeted: a message that knows the recipient’s name, their employer’s email domain, the airport they travel through, and in some cases the car they parked there.

That specificity is what makes the warning MAG issued to its customers, to be careful with unexpected emails, calls and messages, harder to follow than it sounds. The whole point of a message built on real details is that it does not look unexpected.

There is a second layer to this that the corporate travel industry has not fully priced in. As AI agents take on more of the booking and expense workflow, they inherit the same trust problem, and they may be worse at it than people are.

Researchers from Stanford and Georgia Tech showed, in work presented at ACL in 2025, that AI agents completing web-based tasks including travel booking can be manipulated by content specifically designed to fool a machine rather than a human. A human who has learned to distrust a suspicious pop-up carries a decade of hard-won pattern recognition. An agent processing a page does not automatically have that.

Combine a breach dataset accurate enough to make a fraudulent message look legitimate with an agent authorized to act on messages, and the failure mode stops being an employee clicking something they should not have. It becomes a system acting on instructions it had no basis to question.

The governance question this lands on

Evan Konwiser of Amex GBT made a point in a Forbes interview with Jeff Fromm that reads differently after a week like this one. Discussing what it takes to build traveler profiles rich enough for AI to personalize a trip, he argued there is no shortcut around the unglamorous requirements: security, governance, compliance, and robust deletion.

Deletion is the one that usually gets skipped, and it is the one this breach turns on. An email address collected at a WiFi captive portal in 2023 has no operational reason to still exist in 2026. It exists because deleting data requires someone to decide it should be deleted, build the process to do it, and accept the marketing cost of a smaller list.

For a corporate travel program, the practical version of that question is uncomfortable and worth asking anyway. How many entities currently hold traveler data belonging to your company, and how many of them do you have a contract with? The gap between those two numbers is the exposure that no security questionnaire has ever covered.

What a certification actually commits you to

There is a reasonable objection to everything above, which is that security certifications are widely treated as procurement theater. A logo on a vendor’s website, a PDF in a bid response, a box someone ticks.

That reading is fair for some standards and unfair for ISO/IEC 27001, mostly because of what the standard demands in the areas companies least enjoy addressing. It is not a checklist of technical controls. It requires an organization to map its own risks, implement proportional controls, and monitor them on a continuing cycle, which means answering the question of what data exists, where it lives, who can reach it, and on what schedule it gets destroyed. Retention and deletion are not optional annexes in that framework. They are the part an auditor returns to.

At VOLL, the corporate travel and expense management platform I cofounded, we were the first Brazilian corporate travel management solution built on proprietary technology to obtain the international certification, which the trade press covered in 2024. Others in the market have since certified as well, which is the outcome anyone serious about this should want.

I am not raising it as a credential, and a certification is emphatically not a guarantee against incidents, since nobody credible in security would claim otherwise. I raise it because the process changes what a company knows about itself.

The uncomfortable discovery in that kind of audit is rarely a missing firewall. It is a dataset nobody remembered was still being kept.

The more useful application of the standard is on the buying side rather than the selling side. A travel manager evaluating suppliers can ask a certified vendor concrete questions with verifiable answers: what is the retention schedule for traveler profile data, what is the response time commitment when an incident occurs, what happens to the data when the contract ends, who inside the vendor can access a traveler profile and under what approval.

Those questions are answerable under a certified management system and largely unanswerable without one. That is the practical difference, and it matters more than the logo.

What a program can actually do about it

The instinct is to write a policy prohibiting travelers from using airport WiFi, and that policy will be ignored, in the same way that policies against premium cabins get quietly routed around by executive exception. A rule that makes a traveler’s day materially worse does not survive contact with a delayed connection.

More workable options exist, and most of them cost very little. What follows is what I would put in front of my own team, offered as one operator’s view rather than as security advice from a security professional, which I am not.

1. Start with the captive portal problem, since it produced the bulk of the records in this case. Giving travelers a mobile data allowance removes most of the reason to connect to airport WiFi at all, and it costs less per trip than most programs assume. Where terminal WiFi is genuinely necessary, a simple instruction not to use the corporate email address at sign-up captures a meaningful share of the exposure at no cost whatsoever.

2. Then map what your program does not control. Most travel managers can list their contracted systems from memory and have never written down the uncontracted ones: airport parking accounts, lounge memberships, fast-track purchases, hotel loyalty profiles, rental car programs, airline accounts. That list is the actual exposure surface, and building it is an afternoon of work that almost nobody has done.

3. Add retention questions to supplier reviews, not just security questions. Every vendor will confirm they encrypt data. Far fewer can tell you how long they keep it, what triggers deletion, and what happens to a traveler profile ninety days after a contract ends. The answer to that last question is frequently that nobody has thought about it.

4. Treat traveler profile data as a live inventory rather than a permanent record. Passport numbers, dietary information, emergency contacts and loyalty credentials accumulate in travel systems and are almost never pruned. A profile for an employee who left three years ago is pure liability with no operational value.

5. Build the disruption plan and the breach plan as one exercise. The 2024 outage and this breach have almost nothing technically in common, and yet the first hour of a company’s response to either is identical: who decides, who communicates to travelers, and through which channel when the usual channel may itself be compromised or offline.

6. Finally, run the drill before it matters. Most companies have a duty of care protocol for physical incidents and nothing equivalent for information incidents, despite the second being considerably more likely in any given year.

None of this is glamorous, which is exactly why it does not get done. I have argued before that the real cost of a travel program lives in the parts that never show up on an invoice, and this belongs in that category. There is no line item for data that has not leaked yet.

It is worth noting that the travel managers doing the most advanced work in this field have been treating the traveler’s full journey as their responsibility for a long time. When BTN named Moderna’s Jennifer Steinke its Travel Manager of the Year this year, the recognition was for AI work built on a data foundation she spent years assembling deliberately. That same discipline, knowing precisely what data exists, where it sits, and why it is being kept, is what makes a program resilient rather than merely modern.

The regional footnote

One last observation, and it is the reason I keep returning to this theme.

This breach happened in the UK, was reported by British media, and is being assessed by a British regulator. The 2024 outage originated with a US vendor and grounded aircraft on four continents. Neither is a local story, because the infrastructure underneath corporate travel is not local. The exposure exists wherever business travelers move through airports, which is everywhere.

Latin America has been ahead of the curve on adopting new travel technology, and that pace has come with real investment in the governance underneath it. The certification work I described above happened in Brazil, in a Brazilian market, driven by Brazilian corporate clients who were asking the right questions of their suppliers before regulation forced anyone to. That is worth saying plainly, because the default assumption in global industry coverage runs the other way.

I have written before about how much of this industry’s real learning happens in rooms rather than in press releases. This is a case where the learning is available for free, from someone else’s incident, before it becomes anyone’s own. The airports involved will absorb the regulatory consequences. Everyone else gets a map of an exposure surface they were not looking at, at no cost, which is the cheapest security audit a travel program will ever receive.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

The landing

If you have never flown into Congonhas, the approach is disorienting the first time.

The aircraft descends over a dense residential grid, and for most of the final minutes there is no airport visible at all, only apartment towers. Then the runway appears, abruptly, between the buildings. Close enough that you can see into the windows of the apartments alongside it, close enough that people living there can presumably see you. Foreign visitors post about it constantly, usually with some variation of the observation that the plane seems to be landing in the middle of a neighborhood.

That is because it is. Congonhas is not an airport with a city that grew toward it. It is an airport that was surrounded, decades ago, and never moved. It handles over 20 million passengers a year from a footprint wedged into the urban fabric of a city of 12 million.

The visual is memorable enough that it tends to be the whole story people tell about the airport. But the more interesting thing about Congonhas is not where it sits. It is what flies out of it.

A distribution that does not look like the others

Downtown airports exist for a specific reason. They are close to the business district, they take short haul traffic, and the time saved getting to them justifies the operational compromises they impose. Sao Paulo has Congonhas. New York has LaGuardia. London has City. Buenos Aires has Aeroparque.

They are usually described as versions of the same idea. Looking at weekly flight frequencies, they are not.

Pulling route data from Flightradar24 in August 2026, the busiest route out of Congonhas is the shuttle to Rio de Janeiro’s Santos Dumont, at 368 flights per week. The second busiest, to Brasilia, runs 167. Then Porto Alegre and Belo Horizonte at 160 each, Curitiba at 138, Salvador at 95, Recife at 85, Florianopolis at 84, Goiania at 67.

LaGuardia is busier at the top but flatter overall

LaGuardia’s leading route is Chicago O’Hare at 227 flights per week, followed by Toronto at 162, Atlanta at 159, Boston at 154, Dallas Fort Worth at 125, Miami at 104, Charlotte at 101, Washington National at 94, Denver at 89, Nashville at 86.

The first place figure is large, but the drop to second is far gentler than Sao Paulo’s. More importantly, the list spreads across a continent. Chicago, Atlanta, Dallas, Denver and Miami are not a single business corridor. They are the hubs of an entire national network, and LaGuardia connects to all of them at similar intensity.

The tail is also thicker. LaGuardia’s tenth busiest route runs 86 flights per week. Congonhas’s ninth runs 67. The American airport sustains meaningful frequency much further down the list.

London City operates on a completely different scale

London City’s busiest route is Amsterdam at 57 flights per week, then Zurich at 37, Milan Linate at 31, Edinburgh at 30, Glasgow at 25, Luxembourg at 23, Nice at 22, Dublin at 21, Ibiza at 19, Berlin at 19.

The entire airport runs at a fraction of the frequency of the other two. Its busiest route would not crack the top five at Congonhas. What it does instead is serve a tight set of European financial and administrative centers, plus a couple of leisure destinations, at consistent low volume.

That is a different design entirely. London City is not a shuttle airport with a dominant corridor. It is a convenience airport for a specific professional population moving between a handful of cities, and its route list reads exactly like that.

Aeroparque mixes business and leisure in a way the others do not

Buenos Aires Aeroparque leads with Bariloche at 102 flights per week, then Cordoba at 92, Mendoza at 87, Iguazu at 66, Sao Paulo Guarulhos at 56, Santiago at 51, Neuquen at 51, Salta at 47, Rio de Janeiro at 47, Tucuman at 41.

Bariloche and Iguazu are tourism destinations. So, substantially, is Mendoza. The Argentine downtown airport carries domestic leisure traffic at the top of its list in a way neither Congonhas nor LaGuardia does, and the distribution across its top ten is the flattest of the four.

That matters for the comparison, because it shows the route profile is not determined by the format. It is determined by what the country’s internal travel demand actually looks like.

Why the São Paulo concentration exists

Two things explain the shape of the Congonhas list.

The first is that Sao Paulo and Rio sit roughly 400 kilometers apart, which is the distance where flying beats driving for a same day business trip but does not require a wide body aircraft or a long block time. That is close to the ideal shuttle distance, and both cities have a downtown airport positioned to serve it.

The second is that Brazil’s corporate activity is more geographically concentrated than the American equivalent. LaGuardia connects New York to five or six comparably important business centers. Congonhas connects Sao Paulo to one dominant counterpart and then a set of considerably smaller ones. That is not an operational choice by any airline. It reflects where the country’s corporate demand actually sits.

Guarulhos, the larger Sao Paulo airport by volume, confirms this from the other direction. Its busiest route is Porto Alegre at 125 flights per week, then Santiago at 118, Recife at 104, Curitiba and Belo Horizonte at 96 each, Rio Galeao at 82, Salvador at 72, Brasilia and Fortaleza at 68 each. The distribution is flatter, international destinations appear immediately, and no single route dominates.

The two Sao Paulo airports are not a large one and a small version of it. They do structurally different jobs, and any analysis treating Sao Paulo as a single aviation market will miss that.

What concentration does to price

This structure has a direct commercial consequence, and it shows up in fares.

Comparing domestic tickets issued for Monday morning departures against the same routes Tuesday through Friday, in the same 6am to noon window, Monday averaged 2,095.22 reais against 1,314.73 reais. That is 59.4 percent more for the identical trip taken a day later, across 930 paired domestic routes and 11,874 Monday morning issuances between August 2025 and July 2026.

Demand concentration in time and demand concentration in geography compound each other. When a large share of a country’s corporate air travel wants the same corridor in the same few hours of the same day, the pricing response is sharp rather than gradual. An airport whose top route carries twice its second place has very little slack to absorb a demand spike.

I have written before about how the corporate travel market here is growing in price rather than in trips, and about how Brazilian airfare stayed expensive even when input costs moved in the country’s favor. Structural concentration belongs in that explanation alongside fuel and exchange rates, and it gets discussed far less.

What this says about reading global aviation data

None of this appears in a global corporate travel report, and the reason is instructive.

Aggregate aviation analysis tends to compare airports on passenger volume, on-time performance, or total destinations served. Those measures put Congonhas and LaGuardia in roughly the same category. The shape of the route distribution, which is where the actual difference lives, rarely gets compared at all.

Latin America keeps showing up in the data I work with at VOLL, the corporate travel and expense management platform I cofounded, as a region whose structure diverges from the global average rather than tracking it. This is a small example, but a clean one. Four airports built for the same purpose, and four different answers to the question of what that purpose actually requires.

A company managing travel out of Sao Paulo is operating in a market with a single dominant corridor and very little structural slack. A company managing travel out of New York is not. Those are different planning problems, and running the same playbook for both is how a travel budget gets surprised.

The runway between the apartment buildings is the part visitors remember. The route list is the part that shows up on the invoice.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

A number that resets every seven months

METR, an independent research organization, has spent years measuring something more useful than how smart an AI model is. It measures how much time of human work an AI agent can complete on its own before it needs a person to step in. That measure, the length of task an agent can handle unsupervised, has been doubling roughly every seven months, consistently, for several years running.

Seven months does not sound dramatic until you do the math. It means what an agent can do alone this year roughly doubles again by next year, and doubles again the year after that. A governance policy written eighteen months ago was written for a tool that, by the numbers, no longer exists. Most companies are not behind on adopting AI. They are behind on rewriting the rules for what they already adopted.

I spent time at Stanford this year around exactly this question, and the conversations that stuck with me were not about which model performs best. They were about what happens once the model is good enough that the bottleneck moves somewhere else entirely.

Why the interface gets the headline and the data keeps the value

Evan Konwiser, chief product and strategy officer at Amex GBT, put this well in a Forbes interview with Jeff Fromm this month. His framing was that the breakthrough interface gets the attention, but the durable value usually sits in the plumbing underneath it: proprietary context, a functioning marketplace, clean data, and the speed to execute on all three at once.

That is a governance point disguised as a product point. Anyone can eventually license a capable model. What a competitor cannot simply license is a decade of transaction history, the context of who travels where and why, and the infrastructure that turns that context into a decision in real time rather than a report next quarter. The model is a commodity faster than most companies expect. The data is not.

This reframes what governance is actually protecting. A governance policy focused only on which AI tool employees may use is protecting the wrong layer. The layer worth protecting is the proprietary data feeding it, who can access it, what it gets used to decide, and how a company proves that later if it has to.

Adoption happens inside the workflow, or it does not happen

Konwiser’s interview also pointed at where AI adoption in travel is actually landing: inside tools people already use, not as a separate destination. Integration with Concur, Claude, Microsoft Teams, Google Chat, and Slack matters more than a standalone app ever could, because the win condition is fewer manual steps, not a new interface to learn. An agent that notices a meeting on the calendar, books the flight, checks for a conflict, and reconciles the expense without anyone opening a new tab is the actual goal line.

That detail matters for governance too. If adoption happens inside existing workflow tools, oversight has to live there as well. A policy that only governs a dedicated travel app misses every booking that happens through a chat interface or a calendar integration instead.

What an AI native organization actually looks like

I wrote earlier this year about what three days in Chicago taught me and about the number that ended up on a screen there. Both pieces were circling the same idea from different angles: the companies pulling ahead are not the ones with the most AI pilots running. They are the ones that changed how a decision gets made, not just which tool touches it last.

An AI native organization, in the sense I mean it, is not defined by how many agents it has deployed. It is defined by whether a person can explain why a decision was made, quickly, without reconstructing a chain of tool calls after the fact. Konwiser’s “show your work” idea belongs here as much as it belongs in a conversation about traveler trust. Explainability is not a customer facing nicety. It is the actual mechanism of governance, the thing that lets a compliance team, a finance team, or a regulator ask a question and get a real answer.

There is also a research reason to take this seriously rather than treat it as caution for its own sake. Researchers from Stanford and Georgia Tech, presented at ACL in 2025, showed that AI agents completing web based tasks, including travel booking, can be manipulated by fake pop-ups built specifically to confuse a machine rather than a human. An organization that cannot explain how an agent reached a decision also cannot easily tell the difference between a good decision and a manipulated one. Governance and explainability are not separate workstreams. They are the same workstream, described from two directions.

The cost of adoption nobody puts on a slide

Every conversation about AI adoption eventually gets to cost, and almost every one of those conversations means the license fee. That is the smallest number in the equation. The larger cost is what it takes to make an organization’s own data usable by an agent in the first place: cleaning it, structuring it, deciding who owns which piece of it, and building the audit trail that lets someone explain a decision six months later.

I have made a version of this argument before about the corporate travel ROI problem: the tools that look cheapest on a vendor invoice are often the most expensive once you count what it costs internally to make them work well. Governance is the same math wearing a different hat. A company that skips the unglamorous work of data governance is not saving money. It is deferring the cost to the moment something goes wrong and nobody can explain why.

What this means for a region still writing its own rules

Most of the public conversation about AI governance in travel, including the Stanford conversations I sat in on, is anchored to US and European data and US and European regulatory expectations. Latin America keeps showing up in the data I work with at VOLL as a region that adopts fast and governs formally less often, which is a different risk profile than the one most governance frameworks assume.

That gap will not close by importing a governance framework built somewhere else and hoping it fits. It closes the same way trust closes in Konwiser’s framing: in plain sight, one explainable decision at a time, using the region’s own data rather than a global average that was never measuring it in the first place.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

A survey with an odd gap

Amadeus published a report this year called “Tailored Horizons,” and one finding in it does not behave the way you would expect. Seventy six percent of business travelers say they trust AI to put together a travel option for them. Only forty four percent trust AI to spend within their company’s budget while doing it.

That is a thirty two point drop between two tasks that look, on the surface, like the same task. Choosing a flight and paying for a flight both involve a decision. If trust in AI were a single, general quantity, it would not fall this hard just because the second decision touches a corporate card instead of a preference.

I think that gap is the most useful number in the whole report, more useful than either percentage on its own, because of what it implies about how trust actually works.

Why spending money is different from choosing a flight

Evan Konwiser, chief product and strategy officer at Amex GBT, made a similar point in a Forbes interview this month with Jeff Fromm, and it lines up closely with what Amadeus found. Konwiser argued that the internet spent two decades building trust by giving travelers more visibility: more options, more reviews, more control over the final choice. AI proposes the opposite trade. It processes in the background and hands you a recommendation, effectively asking you to trust that it already considered the alternatives.

That trade is easier to accept when the cost of being wrong is small. If an AI suggests a mediocre restaurant, you shrug and pick somewhere else next time. If it books a flight that arrives after your meeting starts, or approves a hotel that blows through a client’s budget cap, the cost is not just financial. It is the meeting you missed, or the conversation with finance you now have to have.

Konwiser’s phrase for the fix was “show your work.” If a flight gets excluded because of a policy rule, the traveler should be able to see why. If another option gets recommended because it saves two hundred dollars, that logic should be visible, and contestable. Amadeus’s forty four percent reads like the current state of that visibility problem. Travelers are comfortable letting AI narrow the field. They are far less comfortable letting it spend without a way to check its reasoning.

The market already moved this direction

While that trust gap sits unresolved, the market has already started moving toward less visibility, not more. Google confirmed a test, currently limited to the United States, in which its AI assistant chooses and books a hotel directly from a plain language request. No comparing ten tabs of options, no scrolling through reviews. You describe what you want, and the agent decides.

Google Maps gained a similar feature the same week. One example Google used to demonstrate it: find a well rated hotel, with good value, near a gym and a restaurant, for a conference next weekend. That is close to word for word what a corporate travel booker types into a search box every day.

This explains a number Skift reported not long ago: only about six percent of hotels currently show up in searches conducted by AI agents. The other ninety four percent simply do not make the list the agent presents. A shelf that used to hold thirty options now holds three, and whichever hotels are not among those three effectively do not exist for anyone searching that way.

Put next to the Amadeus numbers, this is not a future scenario to plan for. It is a live gap between where the technology already is and where traveler trust currently sits.

A stack you cannot see

The same Amadeus report found that sixty two percent of business travelers already use their own AI tools outside the channel their company provides. Not instead of the sanctioned tool, alongside it.

This is worth sitting with, because it means the trust gap is not preventing AI adoption. It is just pushing that adoption outside the system a travel manager can see. A trust gap in the sanctioned tool does not stop a traveler from asking a general purpose AI assistant to help plan a trip on the side. It just means that conversation, and whatever advice comes out of it, never touches the platform that is supposed to enforce policy.

A policy you cannot see being worked around is a policy that is already partly obsolete, no matter how well it is written. The sixty two percent is not a warning about the future. It describes something already happening, today, inside companies that believe their travel program is fully contained.

When the agent gets fooled

There is a second reason caution around AI spending money makes sense, and it has nothing to do with policy design. It has to do with whether the agent can be tricked in the first place.

Researchers from Stanford and Georgia Tech published a study, presented at ACL in 2025, showing that AI agents built to browse the web and complete tasks like booking travel can be manipulated by fake pop-ups designed specifically to confuse a machine rather than a human. A human ignores that kind of pop-up almost without noticing. An AI agent, trained to interpret and act on what it sees on a page, does not carry the same decades of learned suspicion.

The paper’s authors flagged travel booking specifically as one of the more exposed tasks, since it already involves navigating unfamiliar sites and clicking through several steps to complete a purchase. Read next to the Amadeus numbers, this stops being a hypothetical. It is a concrete reason the forty four percent should probably not rise to seventy six percent just because the technology gets a little better at picking flights. Spending safely requires a layer of verification that planning does not.

The travel manager becomes something else

Only four percent of the travel managers Amadeus surveyed think their role will disappear because of AI. Given the other findings, that number makes more sense than it might look at first glance.

If travelers trust AI to plan but not to spend, if the market is already pushing toward less visible, single recommendation booking, and if a meaningful share of travelers are running their own AI tools outside the approved channel, someone still has to decide where the boundary sits between what AI can commit to on its own and what needs a human check first. That job does not vanish when the technology improves. It changes shape. The travel manager stops being the person who approves each booking and starts becoming something closer to a governor of automation: the person who decides which categories of spend AI can clear on its own, which ones still route to a human, and how the company finds out when a traveler has gone around the system entirely.

That is a harder job to describe on an org chart, and possibly a more important one than the booking approver it is replacing.

What this looks like from Latin America

One caveat is worth stating plainly. This survey samples business travelers in the United States, and it comes from a vendor with a commercial interest in AI powered travel tools. Neither fact disqualifies the finding, but both mean the forty four percent and the sixty two percent should travel with that context attached, not as universal truths about business travel everywhere.

That caveat matters more here than it would in most regions. Every data set I have worked with at VOLL, the corporate travel and expense management platform I cofounded, shows adoption patterns in Latin America that diverge from US behavior more often than they mirror it. Payment habits, booking lead times, even which day of the week concentrates the most travel look meaningfully different once you look at transactional data from inside the region rather than a global survey sampled elsewhere.

Whether the trust gap in Sao Paulo or Bogota is wider, narrower, or shaped differently than the one Amadeus measured in the United States is an open question. It is one somebody should measure directly, with regional data, rather than assume it simply travels along with the rest of the report.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

In August I flew JetBlue’s Mint cabin for the first time, Fort Lauderdale to San Francisco, five and a half hours across the country. I had flown premium internationally for years. I had never had that experience on a domestic flight, and I spent a good part of the flight thinking about why it surprised me. On the return leg a few days later, the same route in reverse, it held up. It was not a fluke or a good crew on a good day. It was a product designed deliberately for someone with exactly my reason for being on that plane.

I mention it because the number I want to write about explains why that product exists, and why more of them are coming.

Airline economics had a rough year. Fuel costs spiked, and industry profits were nearly cut in half as a result, a shock big enough to reshape how every major carrier thinks about its network and its margins. In the middle of that pressure, one line item quietly became the story: premium cabin revenue.

A recent study from Skift, a leading voice in global travel research, reproduces data from a major carrier’s own earnings that puts a precise number on a trend most of us could already feel from the front of the plane getting noticeably fuller.

The number itself

Premium cabin revenue now accounts for roughly half of this carrier’s total passenger revenue, up from 45% just two years earlier. It grew at more than double the rate of the main cabin over the same period. Put plainly, the fastest-growing part of a major airline’s business is the part with the highest margin, at exactly the moment fuel costs are compressing margin everywhere else.

That is not a coincidence, and it is worth being precise about the mechanism, because the easy explanation, that rich people are simply flying more, only accounts for part of it.

Why airlines lean harder into premium when costs rise

An airline facing a fuel-driven cost shock has a limited number of levers. It can cut capacity, which sacrifices revenue and market share. It can raise fares broadly, which risks demand elasticity across the whole customer base. Or it can reallocate the same physical aircraft toward the segment of demand that is least price-sensitive and most willing to pay for comfort, reliability, and time saved.

Premium is the lever that lets an airline extract more revenue from the same flight without needing more flights, more fuel, or more risk on the price-sensitive segment of its customer base. A widebody with more premium seats and fewer economy seats can generate meaningfully more revenue on the identical route, using the identical amount of fuel, which is precisely the trade an airline wants to make when fuel is the input squeezing every other part of the business.

This is the same dynamic I described in a different context, the way a fuel shock does not distribute evenly across a cost base, it concentrates wherever the airline has the most room to pass it through. Premium demand is exactly that room.

The corporate traveler is quietly part of this story

It would be a mistake to read this purely as a leisure luxury trend, wealthy travelers upgrading for comfort on a big trip. A meaningful share of premium demand growth is coming from business travel, where the calculus has shifted: with global business travel spending rising 7.2% this year while the number of trips grows barely 1%, as I have written about before, companies are increasingly paying more per trip rather than taking more trips.

If a company is going to send someone on fewer, more consequential trips this year, the case for putting that person in a seat where they arrive rested and productive gets easier to make, not harder, even as the overall travel budget tightens. Premium is not necessarily the discretionary indulgence it looks like from the outside. In a year defined by fewer, higher-stakes trips, it can be the more defensible spend, not the more extravagant one.

I recognize this from my own routine rather than from a spreadsheet. I fly LATAM’s Premium Economy regularly across Latin America, and the reason has almost nothing to do with luxury. It is that a four-hour flight that ends with me able to walk into a meeting and think clearly is a different economic product from one that ends with me needing an hour to recover. The airline is not selling me comfort. It is selling me the arrival, and the arrival is the only part of the trip my company is actually paying for.

What this means for anyone planning a travel budget

For a corporate travel manager, the premium pivot is worth watching as a pricing signal, not just a service tier. When airlines reallocate capacity toward premium, economy inventory on the same routes tends to tighten, which can push economy fares up even without economy demand itself increasing. A program that only tracks its own booking class may miss the reason its own economy fares are climbing: it is not that more people want economy, it is that airlines are quietly making less of it available.

For anyone advocating for premium travel within a company, the fuel-cost context gives a sharper argument than comfort alone ever did. The airlines’ own economics say the most efficient way to protect a margin under pressure is to concentrate value on fewer, better trips, and a corporate travel program facing the same pressure, more cost per trip in a year of tighter budgets, is making a structurally similar decision when it protects premium access for its highest-stakes travel.

Half, and rising

So the number to hold onto is the half. Not because it is dramatic on its own, but because of what it reveals about how an entire industry responds to a cost shock it cannot avoid.

Airlines did not respond to a fuel crisis by cutting the plane in half. They responded by reshaping what is inside it, and the direction of that reshaping, more premium, less standard economy, growing at more than double the rate, is a preview of where every travel budget touching that inventory is going to feel the pressure next.

Which brings me back to that flight in August. What surprised me was not that the product was good. It was that someone had decided a domestic route was worth building that product for, on an aircraft that would previously have been configured almost entirely for the cheapest seat available. That decision is what half of passenger revenue looks like from the inside of the cabin, and it is going to keep showing up on more routes, in more markets, including this one, for exactly the reasons the fuel numbers explain.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Ask people what worries them about companies using AI to handle customer interactions, and you would expect the usual list: privacy, job loss, bad information, the machine getting it wrong. All of those concerns are real, all of them show up in the data, and, according to a recent study from Skift, a leading voice in global travel research, nearly all of them are declining.

Poor quality interactions, down. Extra effort required to get understood, down. Distrust in the information provided, down slightly. The one concern that is rising, and the one that is now the single largest worry people report about AI-driven service, is simpler and harder to fix than any of the others: the absence of a human being to connect with.

What the numbers actually show

Lack of human connection is now cited by roughly half of travelers as their top concern about AI automating service interactions, ahead of job loss, ahead of data misuse, ahead of every other item on the list. And it is the only one of these concerns that increased year over year while every other concern on the same list fell.

That divergence is the finding. People are simultaneously growing more comfortable with what AI can competently do, and more aware of what it does not provide no matter how competently it does the rest. Those are not contradictory reactions. They are two halves of the same recalibration, happening at the same time, in the same population.

Why competence and connection are different axes

I think the natural mistake is to assume these concerns move together, that as AI gets better at the mechanics of service, the human-connection worry should fall too, the same way the accuracy and effort concerns are falling. The data says otherwise, and I think the reason is structural rather than temporary.

Accuracy, effort, and trust in information are properties of execution. A system either gets the task right or it does not, and AI systems have been getting measurably better at execution, which is exactly why those specific worries are declining. Connection is not a property of execution at all. It is not something a system does better or worse as its capability improves. It is either present, because a person is genuinely on the other end, or it is structurally absent, no matter how well the system performs everything else.

That is why fixing accuracy does not touch the loneliness number. The two are not on the same axis, and improving one has no logical reason to move the other. A perfectly executed interaction with no human anywhere in it is still an interaction with no human in it.

This is the argument I keep making, from a different angle

I have written before about a moment in Chicago when a business traveler’s flight was cancelled and an airline’s chatbot handled the disruption competently, right up until the options ran out and a human had to step in to actually solve the problem. I used that story to argue that the future of this industry is not automated, it is orchestrated, and that orchestration is still fundamentally human work.

This data is the same argument, arriving from consumer psychology instead of from a single anecdote. The rising loneliness concern is what it looks like, at scale, when an entire population starts to feel the gap between a system that works and a system that connects. It is the statistical shadow of exactly the boundary I described in that story, showing up not in one traveler’s account of one bad morning, but in how an entire market is starting to talk about its relationship with automated service.

What this means for anyone building service around AI

The lesson is not to build less capable AI. Capability is not the problem, and the data is explicit that capability concerns are the ones falling.

The lesson is that removing friction and providing connection are two different products, and a service strategy that only optimizes the first while assuming the second follows automatically is optimizing against a trend the data says is moving the other way. A company that automates everything it can and treats the remaining human contact as a cost center to minimize is building directly against the one concern that is getting worse, not better, as AI adoption rises.

The companies that will feel this first are the ones in high-stakes, high-stress categories, travel very much among them, where the moment someone reaches for a human is rarely a moment of convenience. It is a moment of disruption, urgency, or fear, precisely the moment where the absence of a person to connect with is felt most acutely, and precisely the moment a rising concern like this one becomes a competitive difference rather than a footnote in a survey.

The gap that gets more valuable, not less

So here is the pattern underneath the two Skift findings, held together. AI is getting better at the parts of service that can be measured as correct or incorrect, fast or slow, and people are responding to that improvement exactly as you would expect, with rising comfort.

AI is not getting better, and structurally cannot get better, at being a person, and people are responding to that with rising concern, even as everything around it improves. That gap is not a temporary side effect of AI still being new. It is the permanent shape of what AI is and is not, and the businesses that plan around its permanence, rather than assuming it closes with the next model update, are the ones that will still have a human connection to offer when the traveler needs one most.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Rising prices usually produce a simple prediction: demand falls. Fewer people travel when travel costs more. It is intuitive, it shows up in plenty of other categories of spending, and it is largely wrong for how people are actually responding to this year’s higher travel costs.

A recent study from Skift, a leading voice in global travel research, shows what travelers are doing instead, and the pattern is more specific and more interesting than a simple pullback.

What people are actually cutting

Facing higher prices this year, the overwhelming majority of travelers report they will still travel. What changes is not whether, but how.

Roughly 37% are searching for cheaper flights. About 30% are choosing less expensive destinations. Another 30% are shortening trips. Around 20% are accepting lower-quality accommodation, and a similar share are simply traveling less often. Look at that list again. Every single item on it is a way of preserving the trip while reducing something inside it. Almost nobody’s first move is canceling the trip outright.

This is the opposite of how price sensitivity usually works in retail, where a category under cost pressure typically sees volume fall first and mix shift second. Travel is behaving like a category where the trip itself sits closer to a fixed cost of living than a discretionary purchase that gets cut when money tightens.

The income split that complicates the story

The picture gets more textured once you separate travelers by income, and it resists an easy morality tale about the wealthy being untouched and everyone else suffering.

Higher earners are not trading down at anywhere near the rate of lower earners, but they are not indifferent to cost either. They are optimizing: watching prices more closely, adjusting timing, choosing smarter rather than choosing less. Lower earners are the ones actually trading down, in the more visible sense of shorter trips, cheaper destinations, lower-tier accommodation. The luxury end of the market has proven durable through exactly the same price pressure that is reshaping the middle and lower end of it.

That split matters for anyone selling into this market, because it means the demand curve is not one curve. It is at least two, moving somewhat independently, and treating the whole traveler base as a single price-sensitive population will misread both segments.

Why the trip survives even when the budget does not

I think the explanation sits in what a trip actually represents to the person taking it, which is rarely just the transportation and the room.

A trip usually anchors something else: a wedding, a reunion, a long-planned milestone, a piece of identity the person has built around being someone who travels. Cutting the trip entirely means cutting that anchor, and the data suggests people will absorb a great deal of discomfort in the logistics before they give up the anchor itself. A shorter trip to a cheaper destination in a lower-tier hotel is still the trip. A cancelled trip is not a smaller version of anything. It is a different outcome entirely, and travelers appear to treat it that way, reserving cancellation for genuine last resort rather than a first response to a higher fare.

I have written before about the industry-wide version of this pattern, where global spending grows faster than the number of trips taken because price is absorbing most of this year’s pressure. This consumer-level data is the same story from the other side of the transaction: the traveler is absorbing the price increase inside the trip, category by category, rather than walking away from the trip altogether.

What this means for anyone selling or managing travel

If you sell travel, the mistake is assuming a price increase will cost you the customer. The data suggests it is far more likely to cost you a line item inside their itinerary than the itinerary itself. The traveler who cannot afford the destination they wanted last year is not gone. They are looking at your cheaper inventory, your shorter package, your off-peak dates.

If you manage a corporate travel program, the same logic applies with a different frame. Employees are not going to stop traveling because costs rose. They are going to look for permission to cut somewhere, and a program that gives them a structured way to do that, a slightly less expensive hotel tier, an earlier booking window, a nonstop instead of a connection with a lounge, will capture the saving the traveler was already looking for. A program that offers no structured way to trade down simply pushes that same instinct toward whatever workaround the traveler invents on their own.

The trip is the last thing to go

So the headline prediction, that higher prices reduce travel, turns out to describe the wrong variable. Higher prices are reshaping travel, category by category, trip by trip, while leaving the decision to travel at all remarkably intact.

That resilience is not evidence that price does not matter. It is evidence of what travelers value most inside a budget under pressure, and it is the trip, not any particular version of it. Everything else is negotiable first.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Every few years, someone writes the obituary for the travel agent. Online booking would kill the profession in the 2000s. Mobile apps would finish the job in the 2010s. Now it is AI’s turn to be cast as the final blow.

A recent study from Skift, a leading voice in global travel research, tells a different story, and it comes with a number that undercuts the obituary better than any counterargument I could construct myself.

What actually happened to the profession

US travel agency employment did collapse, and the collapse was real. It peaked in 1999 and fell hard through the 2000s as the industry restructured for the online era. That part of the story everyone already knows.

What almost nobody mentions is what happened after the collapse: employment has been steady for more than a decade. Not growing dramatically, but not continuing to erode either. The profession found a floor and has been standing on it, largely unnoticed, while three successive waves of disruptive technology were each supposed to be the one that finally finished the job.

A quarter of all travel is still booked through a traditional agent today. That share concentrates heavily in the trips that are hardest to plan alone: 63% of cruise travelers used an advisor in the past twelve months, and the channel as a whole has found its footing in complex, high-value trips where trust and expertise matter more than price comparison.

The capital is voting too

If the labor data is the quiet signal, the capital markets just sent a loud one.

Fora, an AI-powered travel agency, raised 60 million dollars in July at a valuation above 1 billion, on the back of 2.3 billion dollars in sales over five years and a network of more than 15,000 advisors. Read that sentence slowly, because the detail that matters is easy to skip past. Investors did not put a billion-dollar valuation on an app that removes the advisor. They put it on a company that gave the advisor better tools.

This is not an isolated data point. It is consistent with everything else in this year’s research: AI is not eliminating the human layer of this industry, it is capitalizing it.

Why the obituary keeps being wrong

I think the recurring mistake is treating “booking a trip” as a single task that technology either automates or does not. It is not one task. It is dozens of different tasks bundled together, and they have wildly different tolerances for automation.

Booking a known flight on a known date is close to a commodity transaction, and software has handled it well for two decades. Planning a multi-stop honeymoon, a large group incentive trip, or a cruise itinerary with dozens of interdependent decisions is a different kind of task entirely, one where the cost of a mistake is high and the value of judgment is correspondingly high. Every wave of travel technology has automated more of the first category. None of them has meaningfully touched the second, because the second category was never really about information retrieval. It was about judgment under complexity, which is precisely the thing AI tools are best at augmenting and worst at replacing outright.

The agents who survived, and the ones now raising capital, are not the ones who resisted technology. They are the ones who let technology take the commodity work off their plate so they could spend more time on the part only a person does well.

What this means beyond travel agencies specifically

I write about this industry, but I think the pattern generalizes, and it is worth stating plainly because I see the same mistake made about corporate travel management, about customer service, about any profession built on judgment applied to complexity.

The mistake is assuming that because AI can now do part of a job, it will eventually do all of it. What the travel agent data actually shows is a job splitting into two parts: a commodity part that keeps getting automated further, and a judgment part that keeps getting more valuable as the commodity part shrinks. The total number of tasks a person handles may fall. The value of the tasks that remain rises to compensate, sometimes more than compensate, which is exactly what a billion-dollar valuation on a company built around advisors, not instead of them, is telling you.

The obituary that never runs

I do not think this means the travel agent profession is safe from further change. It clearly is not. The mix of what an agent does today will look different again in five years, the way it looked different five years ago from what it was in 1999.

But safe from further change and headed for extinction are not the same prediction, and the data keeps confirming the first one while commentators keep writing the second. A profession that found a floor after its worst technological disruption, and is now attracting billion-dollar valuations for versions of itself equipped with the newest disruptive technology, is not a profession on its way out. It is a profession that has already answered the question everyone keeps asking it.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Two datasets, published within weeks of each other by two different research organizations, describe the same region in ways that should not fit together and somehow do.

A recent study from Skift, a leading voice in global travel research, measured regional adoption of AI for trip planning and found Latin America at the top of the list, ahead of North America, Asia, and Europe. Meanwhile, the GBTA’s own index shows the same region sitting at the bottom of nearly every measure of formal travel governance: the lowest adoption of expense management systems in the world, and roughly half the rate of managed travel programs compared to the global average.

I want to hold both facts at once, because I think the temptation is to pick the flattering one and ignore the other, and the real story only shows up when you refuse to do that.

The adoption number

Start with what Skift found. Across every region measured, business travelers use AI to plan trips more than leisure travelers do, and Latin America shows the highest rate of AI-familiar travelers who have used or intend to use AI for trip planning, ahead of every other region in the study.

This is not a small gap. It is the region leading, not trailing, on a technology adoption curve that most global narratives assume moves from developed markets outward. I have written before about why this pattern keeps showing up here: a region that never fully built out the infrastructure a previous technology required is often the one that adopts the next technology fastest, because it has less legacy behavior to unlearn.

The governance number

Now the other half of the picture, and it is a genuinely different kind of fact.

Only about a third of companies in Latin America use a formal expense management system, the lowest adoption rate of any region in the GBTA’s data. Roughly half require or encourage the use of a managed travel program or booking tool, compared to close to two-thirds globally. By almost every formal governance metric this industry tracks, the region sits at or near the bottom.

There is a detail inside that same data that keeps this from reading as simple underdevelopment. Among companies in the region that do have access to a formal expense system, 82% submit expenses through a mobile app, one of the highest rates measured anywhere. The adoption is not absent. It is unevenly distributed, concentrated wherever a good enough tool actually exists.

Why these are not the same finding

It would be easy to collapse these two datasets into one sentence: this region loves technology but lacks discipline. I think that sentence is wrong, or at least imprecise enough to mislead anyone building a strategy on top of it.

AI adoption for trip planning and formal expense governance are not two points on the same curve. They are two different kinds of behavior, responding to two different kinds of incentive. A traveler deciding whether to ask an AI assistant for a hotel recommendation is making an individual choice, in the moment, with no organizational friction standing between the impulse and the action. A company deciding whether to roll out a formal expense system is making an institutional choice, one that requires procurement, training, and a supplier willing to build for the specific regulatory and behavioral texture of the region.

The first kind of adoption happens the moment a good tool exists, because nothing organizational has to move for an individual to try it. The second kind of adoption has historically required someone to build the specific infrastructure this market needs, and that infrastructure, until recently, mostly didn’t exist here in a form that fit.

The gap is not a verdict on the region

I think the natural reading of a region ranking first in one measure and last in another is that something is wrong with the region. I want to argue for the opposite reading.

A market that adopts individual technology fastest while lacking institutional infrastructure is not a broken market. It is a market waiting for the institutional layer to catch up to behavior that already exists. The demand signal is already there, visible in the AI adoption number. What has been missing is not appetite. It is a formal system built specifically for how this region actually operates, rather than one imported from a market with different payment rails, different regulatory defaults, and different behavioral starting points.

That is a different, and much more investable, story than the lazy version. A region resistant to good tools is a hard market to enter. A region that adopts good tools instantly, once they exist, and has simply been underserved by formal infrastructure until now, is close to the best market there is to build for.

What this means for anyone building or buying here

For a company building travel technology for this region, the lesson is specific: the AI adoption number is your permission to move fast, and the governance number is your instruction manual for what to build. Travelers here will use an AI-driven tool the moment it exists and works. The thing that has to be earned, separately, is the institutional trust that gets a company to mandate the tool formally, and that trust is built through governance features, not through AI features. Audit trails, policy enforcement, data visibility. The unglamorous half of the product.

For a company buying travel management here, the lesson runs the other way. Do not read your own low adoption of formal systems as evidence that your travelers resist structure. The data suggests the opposite: your travelers are already comfortable with far more sophisticated tools than the systems your company has offered them. The resistance, if there is any, is more likely to be institutional inertia than traveler preference.

Two numbers, one honest read

So here is where I land, holding both datasets at once instead of picking the one that tells a cleaner story.

This region is not behind on technology and it is not undisciplined about governance. It is a market where individual behavior has outrun institutional infrastructure, which is exactly the condition that precedes the fastest kind of catch-up growth, once someone builds the infrastructure that behavior has already been waiting for.

The AI adoption number and the governance gap are not contradicting each other. They are describing the same opportunity from two different vantage points, and the opportunity is larger, not smaller, for the distance between them.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

The first is a product launch. The second is a research paper. Neither one, on its own, is the whole story.

The launch

Google confirmed to Skift that it has begun a limited test in the United States in which its own AI assistant chooses and books a hotel on the user’s behalf, inside Search’s AI Mode. No comparing ten tabs. No scrolling a results page. You describe what you want, the agent decides.

The company also introduced something called the Universal Commerce Protocol, infrastructure that could let a booking be completed inside the chat itself, without redirecting the user anywhere else. In the same week, Google Maps rolled out similar agentic features, including the ability to book a hotel or order food directly from the app. The example Google itself used to demonstrate the new function is worth sitting with: find a well-reviewed, reasonably priced hotel near a gym and restaurants for a conference next weekend. That is not a hypothetical leisure query. It is, almost verbatim, what a business traveler asks every week.

This is the moment the shelf actually shrank, not the moment someone predicted it would. A recent study from Skift found that only 6% of hotels currently appear in AI-powered search, meaning 94% simply do not surface when a traveler asks an assistant for a recommendation. That statistic explains why Google is moving from suggesting options to making the choice directly: if the discovery layer already compresses thirty options into three, the company that owns the layer has an obvious incentive to own the decision too.

The paper

Now the second thing, and it complicates the first in a way that deserves more attention than it has gotten.

Researchers published a study demonstrating that vision-language agents, the same kind of system now being tested to book your hotel, can be reliably distracted by adversarial pop-ups: fake elements deliberately designed to mislead the machine rather than the person. The paper explicitly names travel booking as one of the task categories most exposed to this kind of attack, and it measured high success rates for the deception across the testing environments used. The work was presented last year at the Association for Computational Linguistics, one of the field’s most rigorous venues, which means this is peer-reviewed research, not speculation from a blog.

The mechanism is almost unsettling in how simple it is. A human ignores a fake pop-up instinctively, a skill built from years of exposure to bad ads and phishing attempts. An AI agent parsing a screenshot or an accessibility tree has no equivalent instinct yet. It processes what is in front of it, and a sufficiently well-designed fake element can sit right in the path of the action the agent is about to take.

Why both facts belong in the same sentence

I do not think the right reaction to the research is alarm, and I do not think the right reaction to Google’s test is uncritical enthusiasm. The two findings are not in tension. They are describing the same transition from two different angles.

The launch tells you that agentic booking has moved from roadmap to shipped product, tested by a company whose search engine is most people’s starting point for travel discovery. The paper tells you that the exact category of task Google just automated is one where independent researchers have already demonstrated a real, reproducible vulnerability. Put together, they describe a technology that is simultaneously more capable and less battle-tested than the pace of its rollout suggests.

That combination is not new in technology generally. It is relatively new in a category where the failure mode is not a wrong answer in a chat window, but a real charge on a real card, for a hotel a real person is going to sleep in.

What this changes for anyone managing travel

If I ran a corporate travel program, this week would change three things about how I think about agentic booking, not in the direction of avoiding it, but in the direction of asking better questions about it.

First, I would ask any vendor pitching agentic booking exactly one question: what happens between the agent’s decision and the money moving. Not whether the agent is smart. Whether there is a verification layer between intent and transaction that a manipulated screen cannot bypass. The pop-up research is specific about where the vulnerability lives, at the exact moment an agent parses what is on the screen and decides where to click. That is precisely the moment a corporate program needs a checkpoint, not an assumption.

Second, I would treat visibility and trust as two separate problems, because the industry is currently conflating them. Being one of the 6% of hotels an AI assistant actually surfaces solves a discovery problem. It says nothing about whether the transaction that follows is secure. A supplier optimizing purely for AI visibility without also hardening the booking flow against exactly this kind of manipulation is solving half the problem and calling it done.

Third, I would stop treating “the AI decided” as a satisfying answer to “why this hotel.” The whole appeal of agentic booking is removing friction, and removing friction is real value. But friction and scrutiny are not the same thing, and a program that automates away scrutiny along with friction is trading a slow process for a fast one with a new, largely unaudited failure mode.

The pattern underneath

I keep returning to a version of this argument, because I think it is the actual shape of what is happening across this industry right now, not just in hotel booking.

Capability is advancing on one curve. Trust infrastructure is advancing on a slower one. Google shipping agentic hotel booking is capability. The pop-up study is a reminder that trust infrastructure has not caught up, and that the gap between the two curves is exactly where the next round of fraud, error, and bad press in this industry is going to come from.

None of this means slow down. It means build the checkpoint before you need it, not after the first story about an agent that booked the wrong hotel, at the wrong price, because it clicked the wrong thing.

The shelf really did shrink to three options this week. The research just reminded everyone that someone can still rearrange which three you see.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Every year the industry’s data tells the same story about my region, and every year almost everyone draws the same conclusion from it, and I think the conclusion is a mistake.

Here is the story. Across the world, roughly two thirds of companies require or encourage the use of a managed travel program, a travel management company or a corporate booking tool. In Latin America, that figure is about half. The region also has the lowest adoption of formal expense systems anywhere measured, used by only about a third of companies. On nearly every measure of how professionally managed corporate travel is, Latin America sits at the bottom of the global table.

The conclusion almost everyone draws is that the region is behind, and that catching up is a matter of time and adoption. I want to argue for a completely different reading of the exact same numbers.

What “unmanaged” actually means

Start with what those numbers describe on the ground, because the word unmanaged hides more than it reveals.

A company without a managed travel program is not a company that does not travel. It is a company whose travel is happening anyway, booked on consumer sites, paid on personal cards, reimbursed through spreadsheets, approved over messaging apps, and reconstructed painfully at the end of every month from a pile of receipts. The travel is real. The spend is real. What is missing is not the activity. It is the visibility.

So when half of Latin American companies run travel without a managed program, that is not an absence of a market. It is an enormous volume of business travel happening entirely outside anyone’s ability to see it, negotiate it, protect it, or improve it. The demand is already there. It is simply invisible.

That distinction changes everything about how you read the number.

Why a mature market is a worse opportunity, not a better one

Here is the part that global players consistently get backwards.

If you are a travel platform, a hotel group, an airline, or an investor looking at where to build, a fully managed market is not the attractive one. It is the saturated one. A region where ninety percent of companies already run a professional program is a region where growth means taking share from an incumbent, one painful contract at a time, competing mostly on price because the category is already understood and already sold.

An under-managed market is the opposite. The growth there does not require convincing a company to switch from a competitor. It requires converting travel that is currently informal, invisible, and unmanaged into travel that is finally organized. No incumbent has to lose for the new provider to win. The market itself expands as the informal becomes formal.

That is the single most attractive growth dynamic there is, and it is precisely the dynamic the data describes in Latin America. The low management rate that reads as backwardness is, read correctly, the largest greenfield in the global travel industry.

The Uber lesson

I keep coming back to a comparison everyone in this region understands in their body, because they lived it.

Most people reading this remember what it was like to get a taxi before the app. You called a number. You waited without knowing if the car was coming. You had no idea what it would cost until the end. You had no record of the trip and no recourse if something went wrong. It was a bad experience that everyone accepted because nobody had shown them there was another way.

Then the technology arrived, and adoption was not slow. It was instant, because the old experience had been quietly painful the whole time, and people only understood how painful once they felt the alternative. The market did not need to be educated into wanting it. It needed to be shown it existed.

That is exactly the shape of the corporate travel and expense opportunity in Latin America. The pain of the unmanaged status quo, the personal card, the sixty-day reimbursement, the month-end receipt hunt, the total lack of visibility, is enormous and universal and mostly unspoken, precisely because most companies have never felt the alternative. A region does not adopt slowly when the new thing is genuinely better. It adopts all at once, the moment it sees the other side of the story.

We have proof of this in the same dataset. The region with the lowest management rate also shows some of the highest mobile and app-based behavior in the world. Where a good tool exists, people here do not resist it. They take to it faster than almost anyone. The barrier was never culture. It was coverage.

What this asks of the people who serve this region

If I am right that this is opportunity rather than deficiency, then it places a real obligation on the companies that want to serve it, and I include my own.

It means the job here is not to sell a managed program the way you would in a saturated market, as a marginally better version of something the buyer already has. It is to show a company that has never had visibility what visibility feels like, and to make the alternative to the painful status quo so obviously better that adoption becomes the easy choice, the way hailing a car on a phone became the easy choice.

It means building for the reality of this region rather than importing a program designed for New York or London and assuming it fits. The instant-payment behavior, the mobile-first habits, the specific frictions of operating across many currencies and tax regimes, these are not obstacles to work around. They are the design brief.

And it means treating the low management rate not as evidence that the region does not care about professional travel management, but as evidence that nobody has yet made it good enough, local enough, and easy enough to adopt at the scale the region is clearly ready for.

Not behind. Early.

So I will keep pushing back on the reading that says Latin America is behind on managed travel.

The region is not behind. It is early. And early, in a market this size, with demand this real and visibility this low, is the most valuable place a market can possibly be, for the companies serving it and for the companies inside it that are about to discover what they have been missing.

The lowest management rate in the world is not the bottom of a table. Read correctly, it is the front of a line.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

Over three days in Chicago, the chief executives of three major airlines took the stage at the GBTA Convention, one after another across the week, and talked about where their businesses are going. They run very different companies with very different strategies. And if you listened past the strategy, they were all circling the same handful of truths, one of which the industry has been reluctant to say out loud.

I want to walk through what each of them said, and then the thing none of them said directly but all of them implied.

United: the honest number

Scott Kirby, who runs United Airlines, offered the line I have not been able to stop thinking about since, and it was notable precisely because it cut against the optimism of the week.

Demand is strong, he said, but business travel has not fully returned to where it was before the pandemic. There are still fewer corporate trips than there used to be. In a week when the industry was celebrating a record nominal spending forecast, the head of one of the largest airlines in the world stood up and gently reminded the room that the record is in dollars, not in trips.

He said something else that has stayed with me. Airlines, in his framing, are one of the best real-time economic indicators that exist, because business travel is the first thing a company cuts when conditions tighten and the first thing it restores when confidence returns. It is a beautiful way to describe the industry: a live feed of corporate confidence, read in seats sold.

Put those two ideas together and you get a quietly sobering message. The strength everyone can feel is real, but it sits on top of a base that has not fully recovered, and the people best positioned to read the economy are watching it carefully.

Southwest: the culture under transformation

Bob Jordan, the chief executive of Southwest, described a company in the middle of the most significant commercial transformation in its history.

The list of changes he walked through would have been unthinkable at Southwest a few years ago. Assigned seating. A redesigned approach to how the airline handles its cabin. New international partnerships. Connectivity being installed across the fleet. For an airline whose entire brand was built on doing things its own singular way, this is a substantial reinvention.

And yet the heart of what he said was about continuity, not change. His argument was that culture is not a product feature and not a policy, and that a company can transform what it sells and how it operates while keeping intact the thing that actually made it what it is. He reached back to the airline’s founding philosophy to make the point, the old idea that if you do not change you die, which is itself part of the culture he is trying to protect.

What struck me was the tension he was holding: transforming the commercial model precisely in order to preserve the company underneath it. The changes are the means. The culture is the end.

American: the strategy built from listening

The leader speaking for American Airlines described a commercial plan organized around a few clear pillars, a leading network, a premium customer experience, revenue built on being easy to do business with, and loyalty earned as a result of the first three rather than pursued on its own.

But the detail that stayed with me was not the pillars. It was where he said the strategy came from. Not from headquarters, not from a consulting deck, but from listening to customers, to travel agencies, and to corporate travel managers. A global airline’s commercial strategy, shaped in significant part by the people who manage corporate travel programs.

I want to sit on that for a moment, because it says something about my own audience that my own audience rarely hears. The travel managers who so often feel like price-takers, like the department that gets handed fares and told to enforce policy, are in fact a voice that helps shape the commercial strategy of one of the largest airlines on earth. The bargaining power of this profession is greater than the profession tends to believe.

The thing all three implied

Here is what none of them said in exactly these words, and all of them said in effect.

Kirby framed airlines as a read on human confidence. Jordan insisted the technology and commercial changes are in service of a culture made of people. The American leader built an entire strategy out of listening to people. Underneath three different messages was a single shared conviction: in a year when this whole industry was arguing about how much to hand over to artificial intelligence, the people who actually run the airlines spent their time on stage talking about human judgment, human confidence, and human relationships.

None of them was anti-technology. All of these companies are investing heavily in it. But not one of them stood up and said the future is automated. Each of them, in his own way, described a future in which technology is in service of something human, the confidence to travel, the culture that holds a company together, the relationships that shape a strategy.

Why this matters beyond the keynote

I think this matters because it is easy, in a week saturated with sessions about agentic AI, to conclude that the direction of this industry is toward removing people from it.

The three people with the most operational skin in the game described the opposite. They described technology as a tool in the service of human confidence, human culture, and human relationships, which is very close to the argument I have been making about where AI actually belongs in this business: amplifying people rather than replacing them.

When the buyers debating autonomy and the CEOs running the airlines independently arrive at the same conclusion, that the human layer is the point and the technology is the tool, it stops being one commentator’s opinion and starts looking like the actual shape of the industry.

The record, and the base beneath it

So I left those three keynotes holding Kirby’s honest number in one hand and the two culture-and-relationship arguments in the other.

The market is at a nominal record and has not fully recovered in real terms. The airlines are transforming aggressively and insisting the transformation is in service of something human. And the profession I write for turns out to have more influence on all of it than it usually claims.

Three CEOs, three strategies, one quiet agreement. The future they described is not automated. It is human work, supported by better tools. It is worth noticing that the people betting billions on that future said so themselves, plainly, on the biggest stage our industry has.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

I am writing this from Stanford, where I am spending the week in an immersion program about a single premise: that the organizations of the future will, inevitably, be AI organizations. The room is full of leaders from industries that look nothing like mine, which is exactly why the most useful question I have been asked about the future of travel came from one of them.

A classmate, the chief executive of a bank in Poland, turned to me between sessions and asked something deceptively simple. In a world where AI handles more and more, where does the human travel agent actually fit when a trip falls apart?

I had just come from Chicago, where I watched an entire industry wrestle with a version of that exact question. So I told him the story that, for me, answers it better than any framework.

The story I told him

At the GBTA Convention, a corporate travel manager described how she got to Chicago, and without intending to, she resolved the whole debate.

Her flight from New York was cancelled. It was raining torrentially and the airline was dropping flights across the city. Her travel agency’s chatbot worked well at first: it flagged the disruption, offered options, let her rebook, let her manage standby. Everything with a structured answer, the automation handled cleanly and fast.

Then they cancelled everything. The next flight the chatbot could offer was the following day, too late for the event she had crossed the country to attend. At that point the automation had nothing left to give, because it was still trying to solve the problem it had been given: find this person another flight on this airline.

So she called a human. Her corporate travel consultant, a person she knows by name. She said: I need to get there, I do not care how. And the consultant put her on a flight on a different airline that was practically already leaving the gate.

Look at the division of labor in that story. The chatbot tried to solve the flight. The person understood that the problem was not the flight. It was the commitment on the other end of it.

Why the machine stopped being able to help

The reason that boundary matters is not that the AI was bad. It was genuinely good, right up until the moment it was not.

Automation is excellent at the structured problem, the one with defined options and clear rules. Detect the disruption, present the alternatives, execute the rebooking within the airline’s own inventory. That is real work and the AI did it faster and more calmly than a stressed human on a rainy morning could have.

It stopped being able to help at the exact moment the problem stopped being structured. When every conventional option disappeared and the solution required stepping outside the original airline, reinterpreting the goal, and improvising, the machine had no move. The automation optimized the transaction it was handed. The human re-defined the problem.

That is the difference, and it is not a small one. The chatbot was trying to rebook a flight. The consultant understood that her actual job was not to rebook a flight at all. It was to get a person to a place by a time. Same situation, completely different understanding of what the problem even was.

What I told the banker the future looks like

So my answer to his question, where does the human fit, is this.

The human does not fit as a backup for when the machine breaks. That framing is wrong, and it is the framing most people reach for. It imagines a sequence: the AI tries first, and when it fails, a person picks up the pieces. Plan A and Plan B.

The reality I described to him is not sequential. It is a division of labor by type of problem. The AI owns the structured and the predictable. The human owns the ambiguous and the exceptional. They are not each other’s fallback. They are each other’s complement, working the same case from the start. The automation was already working while the consultant was reachable the whole time. When the case crossed from structured to exceptional, the ownership shifted, but neither replaced the other.

He is a banker, so I put it in terms he would recognize. Nobody thinks the future of banking is a world with no bankers. They think it is a world where the routine is automated and the human attention is concentrated on the decisions and relationships that actually need judgment. Travel is the same. The question was never whether AI or people. It was always which problems go to which.

The word I keep coming back to

There is a word that captures this better than automation does, and I have been using it since Chicago.

Orchestration. An orchestra is not a machine that replaced musicians, and it is not a crowd of musicians with no coordination. It is many specialized capabilities, human and instrumental, brought into coherence by something that knows which one should be playing at each moment. That is what the good version of this future looks like. Not autonomous. Orchestrated.

In a travel program, orchestration means the AI monitors, compares, flags, and executes the structured work continuously, and the human specialist is brought in precisely when the situation exceeds the rules, when it needs interpretation, negotiation, or the judgment to redefine the goal. Neither is in charge all the time. Something has to know when to hand off, and that knowing is the actual design challenge of the next decade.

What the AI-native organization gets wrong, and right

Here at Stanford, the phrase in the air is the AI-native organization, and I believe in it. The organizations that win the next decade will be built around these tools from the ground up, not have them bolted on.

But the version of that idea worth wanting is not the one where AI does everything. It is the one where the organization is designed, from the start, around the intelligent division of labor between machine and human, with the handoffs engineered rather than accidental. The naive version automates everything it can and treats the remaining humans as a cost to be minimized. The sophisticated version automates everything it should and treats human judgment as the scarce resource to be concentrated where it matters most.

The travel manager stuck in the rain did not need less technology. She needed technology that did its job flawlessly and then knew, at the right moment, to get out of the way and let a person who knew her name solve the thing only a person could solve.

The answer, in one line

I do not know if my classmate got the answer he expected. But I told him the truest version I have.

The human travel agent does not disappear in a world of AI. The human travel agent becomes the person you reach the moment the problem stops being a problem a machine can define. The future of this industry is not automated. It is orchestrated, and orchestration, in the end, is still human work.

A chatbot tried to solve her flight. A person understood she needed to save the meeting. Until machines can tell the difference between those two things on their own, and they cannot yet, the most important seat in the AI-native travel organization still belongs to a human.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

There is a sentence buried in the methodology of the industry’s most important research that I think is more revealing than any headline number the research produces.

The GBTA Business Travel Index is the most comprehensive measure of business travel spending in the world. It runs on decades of data across dozens of country markets and industries, and when it speaks, the entire industry listens, including me. I have quoted it repeatedly. I trust it. And precisely because I trust it, I want to take seriously something it says about itself: that its data is richest for the United States and Western Europe, and thinner elsewhere. That supplier data does not always cleanly separate the business traveler from the leisure traveler. That in large parts of the world, the picture is estimated more than it is observed.

That is not a criticism of the study. It is an unusually honest disclosure by a serious research organization, and I respect it enormously. But it has a consequence that almost nobody discusses, and the consequence is the entire reason this newsletter exists.

What it means to be estimated rather than observed

Think about what that methodological note actually implies for a region like Latin America.

In a market with deep, clean data, the number the index reports is close to a measurement. It reflects what was actually spent, captured through well-developed reporting systems, corporate cards, managed programs, and mature supplier data. The forecast is an informed projection built on a solid floor of observed fact.

In a market where the data is thin, the number is closer to an estimate built on a model. And the model, necessarily, is calibrated on the markets where the data is rich, which means it carries the assumptions of those markets into a region that does not share them. A model trained on how the United States and Western Europe behave will describe Latin America as a version of those places, scaled down. But this region is not a smaller version of those places. It is a structurally different market, and the ways it differs are exactly the ways a model built elsewhere is least equipped to see.

When your data is thin, you do not see a region as it is. You see a region as your model assumes it should be. And the gap between those two things is largest precisely where the reality is most distinctive.

The specific things a thin dataset misses here

Let me be concrete about what falls into that blind spot, because vagueness helps no one.

It misses the enormous volume of unmanaged travel I have written about, the trips booked on personal cards and reimbursed through spreadsheets, which by definition do not flow through the managed channels and supplier feeds that global data relies on. A region that manages only half its travel is a region where a large share of activity is structurally invisible to the very instruments that measure it.

It misses the informal-to-formal conversion that is the real growth story here, because an instrument calibrated on mature markets is looking for the kind of growth mature markets have, incremental expansion of already-measured activity, not the sudden illumination of activity that was always happening but never counted.

It misses the structural texture. The specific weight of airfare in the cost of a trip. The particular way a fuel shock transmits through a market with no rail alternative. The currency dynamics that make a dollar-denominated number mean something completely different on the ground. These are not details a global model gets slightly wrong. They are the defining features of the market, and they are the first casualties of estimation.

The number that proves the point

Here is the piece of evidence I find most persuasive, and it comes from comparing two sources that should agree and do not.

The GBTA index estimates Brazil’s business travel market at 35.8 billion dollars. Meanwhile, Abracorp, the Brazilian corporate travel association, measures the actual intermediated volume flowing through its member agencies, and that figure, converted to dollars, is a small fraction of the GBTA number.

Now, these two numbers are measuring different things, and I want to be careful and fair about that, because the naive version of this comparison is wrong. The GBTA figure is total business travel spend, including everything unmanaged and direct. The Abracorp figure is only the volume intermediated through its members. They are not supposed to be equal. The gap between them is not an error.

But the gap is the point. That enormous space between total estimated spend and actually intermediated volume is the unmanaged, uncounted, informal market, rendered as a number for once. It is the size of the blind spot, made briefly visible by the disagreement between two honest sources measuring from opposite ends. Most regions do not have a gap that large, because most of their travel flows through channels that see it. Ours does, and the size of that gap is the truest single description of this market I know.

I want to add one honest caveat, because rigor demands it. Comparing these figures requires fixing an exchange rate and being explicit that the scopes differ, and anyone who quotes the comparison without those two caveats is misusing it. But done carefully, the comparison illuminates exactly what the global data admits it cannot fully see.

Why I write from here

This is, in the end, why The Moving Biz exists, and why it is written in English from São Paulo rather than in Portuguese for a local audience.

The best data in this industry is honest about its own blind spot. But a blind spot that everyone acknowledges and no one fills stays empty. The global reader gets a number for Latin America that is real, carefully produced, and structurally incomplete, and unless someone who lives inside the region translates what the number is missing, the incompleteness travels silently into every strategy, every investment decision, and every global program built on top of it.

I am not trying to correct the index. I am trying to annotate it. To stand at the edge of the blind spot and describe what I can see from inside it that the instrument, by its own admission, cannot. The data says here is our best estimate of Latin America, and it is genuinely the best. My job is to add the sentence that comes next: here is what that estimate cannot show you, and here is why it matters.

The honest number and the missing sentence

So I will keep quoting the GBTA index, because it is the best thing we have, and I will keep trusting it, because it earns that trust, including by being candid about its own limits.

But I will also keep insisting on the sentence the data cannot write about itself. A region that is estimated rather than observed is not a smaller version of the regions that are observed. It is a different market wearing a borrowed model, and the distance between the borrowed model and the real thing is the most interesting territory in this entire industry.

That distance is my beat. The blind spot is not a problem to me. It is the whole reason to write.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.

I started my career inside hotels, not travel programs, and there is a number in this year’s GBTA data that I read differently because of it.

Across every region the Business Travel Index measures, Latin America has the lowest average daily hotel rate in the world. Cheaper than Asia Pacific. Cheaper than Europe. Cheaper than North America. If you manage a global travel program and you look at that line, it reads like a win. A night in São Paulo or Bogotá costs your company less than a night almost anywhere else you send people.

I want to explain why that number, which looks like good news on a procurement dashboard, is actually a warning sign, and why the person who should be most worried about it is the traveler.

What a hotel rate actually has to cover

Let me start from the inside, because most people who buy hotel nights have never had to sell one.

A hotel rate is not the price of a room. It is the price of everything that has to happen for that room to exist and be worth staying in. It covers the mortgage or the rent on an enormous, expensive building in a central location. It covers electricity running around the clock, hot water, air conditioning, laundry at industrial scale. It covers the front desk at three in the morning, the cleaning staff, the maintenance team, the security, the breakfast, the elevators that have to be inspected, the fire systems that have to be certified. It covers renovation cycles, because a hotel that is not renovated every few years quietly dies.

A hotel is one of the most capital-intensive, labor-intensive businesses there is, running twenty four hours a day, every day, forever. And unlike most businesses, it cannot store its product. A room not sold tonight is gone. That revenue never comes back.

So when I see the lowest average rate in the world, I do not think cheap. I think about what had to be cut to get there, and who is absorbing the difference.

The three ways a rate gets that low

There are only a few ways a hotel market ends up with structurally low rates, and none of them is a free lunch.

The first is thin margins. The hotel is charging little because it can charge little, and it survives on a razor of profit that leaves nothing for reinvestment. That hotel slowly degrades, because every renovation postponed is a rate kept low this year and a problem deferred to the next.

The second is currency. Rates measured in dollars look low partly because the local currency is weak against the dollar, which flatters the number for a global buyer while telling you nothing about whether the hotel is actually cheap to operate locally. The building still has to be paid for in local reais or pesos, at local inflation, which in this region is rarely gentle.

The third is a market where supply has outrun demand, so hotels compete on price to fill rooms, which is good for the buyer this quarter and corrosive for the market over years, because a hotel that cannot make money cannot invest, and a market full of hotels that cannot invest is a market whose quality is quietly falling while the rate stays attractively low.

In Latin America, the honest answer is that all three are usually happening at once.

Why the cheap rate lands on the traveler

Here is the part that a procurement dashboard cannot show you, and the part I care about most.

When a hotel rate is structurally too low to sustain the property, the room does not stop being sold. It just gets worse. The air conditioning that is not replaced. The staff that is cut from four to two on the overnight shift. The breakfast that shrinks. The renovation that is postponed for the third year running. The safety investment that is deferred because it does not show up in a booking photo.

The buyer sees a rate. The traveler sleeps in the consequence.

And this matters more in business travel than in leisure, because the business traveler is not on vacation. They have a meeting at eight in the morning that could decide a quarter. They need to sleep, to shower with hot water, to have working internet, to feel safe walking back to the hotel at night in a city they do not know. A cheap room that fails any of those tests does not save the company money. It puts the entire purpose of an expensive trip at risk to protect the smallest line in its budget.

I have written before that the most expensive part of a business trip is never the trip. It is the outcome the trip exists to produce. A hotel rate optimized down to the point where the traveler sleeps badly before the most important meeting of their quarter is a false economy dressed up as a saving.

What the number is really telling a global buyer

So if you run a program from New York or London and Latin America shows you the lowest hotel rate on your global report, here is how I would read it.

Do not read it as the region being cheap. Read it as the region being under-priced, which is a different and more fragile thing. A cheap market is one where costs are genuinely lower. An under-priced market is one where rates have been pushed below what the product sustainably costs, and that gap always gets paid by someone, eventually, usually in quality and often in the experience of your own traveler.

Read it, also, as a reason to buy on more than price. In a market with this rate structure, the difference between two hotels at nearly the same rate is enormous, because one of them is investing to stay good and the other is coasting toward decline. Choosing well here requires knowing the property, not just the rate, and that knowledge is exactly what a serious travel program is supposed to provide.

The rate is a symptom, not a bargain

I left the hotel industry for the technology side of this business, but I never stopped reading a rate the way an operator reads it.

The lowest average hotel rate in the world is not a trophy for Latin America. It is a symptom of a market operating under structural pressure, with thin margins, weak currencies, and quality held together by people working very hard for very little. The buyer who celebrates it is reading the price. The operator who lived it, and the traveler who sleeps in it, know the price is not the same as the cost.

The number is real. It just does not mean what the dashboard thinks it means.

About me

I am an entrepreneur with over 20 years of experience at the intersection of tourism and technology. I am co-founder and Chief Business Officer of VOLL, the largest mobile-first corporate travel and expense management platform in Latin America, and a recognized reference in the development of the corporate travel industry.

A Marketing specialist from Fundação Dom Cabral, I serve on the Tourism Council of FecomércioSP and on the Executive Council of the Latin American Association of Corporate Events and Travel Management (Alagev). A frequent traveler and close observer of human behavior in motion, I write and speak about innovation, digital transformation, entrepreneurial leadership, and the future of corporate travel.