If you were at ITB Berlin last week, you probably heard the word “AI” more times than you ever wanted to - so we’d love to cut through the noise and hear your personal conclusion: What’s the one AI takeaway from ITB 2026 that still sticks with you, and why does it matter?
Hotels are better positioned for the AI era than most people think. They own something no LLM can replicate: years of proprietary guest data and a direct relationship with the customer. That's a genuine competitive moat, and agentic AI is finally what turns it into action. A lactose intolerance note in a reservation becomes a housekeeping task before the guest arrives. A return visitor's dining history becomes a conversation at check-in. The data was always there; AI removes the manual effort and does it every time.
Capturing this opportunity requires a genuine shift in mindset, and that can be frustrating. During a panel at ITB, I asked a room of 400 people how many had actually built an AI agent themselves. Just a few hands went up. That has to change. You don't need to be a technical expert, you just need to be curious. Spend a Sunday with Claude Cowork or a similar tool. Expect the first few hours to be annoying and frustrating. Push through it. Build something small, throw it away if you have to, and do it again.
The best time to get your hands dirty was six months ago. The second best time is now.
The takeaway that stays with me isn't about any single technology. It's about what AI can actually do to change the way we experience travel.
An example of this was the live demo I gave on stage to change my own flight home - not by calling a helpdesk, not by navigating an app - but by having a natural conversation with an AI agent that handled everything live: rebooking, seat, baggage, transfer all in one simple phone call.
The audience went quiet for a moment. Not because it was magic but because it felt normal.
After years of AI being framed as a replacement for human roles in travel, I demonstrated live with Amadeus technology something more interesting: AI that makes the traveler feel more looked after, not less.
This agentic AI innovation means that a traveler can quickly and easily make changes to an existing booking. It also means that travel professionals can focus on where their skills are best deployed – in solving complex cases building greater loyalty.
My key takeaway is that the real promise of AI in travel is about making the experience better for everyone, everywhere. At Amadeus, we're building it now - together with our industry partners.
What stood out to me most at ITB Berlin this year was how much the conversation around AI has shifted from theory to real, everyday use. Hoteliers and operators are no longer asking whether AI matters, but where it can make the biggest difference for guests and teams right now.
The businesses in the strongest position will be the ones with open, flexible infrastructure as their foundation, because that makes it so much easier to put AI tools to work in practical ways, turning new ideas into actionable solutions that can improve the guest journey, simplify operations and create value much faster.
There is still a big gap between the AI marketing story and the reality of delivering true use cases, very few tech companies are able to prove that they can make AI work within their customer base.
My ITB takeaway is that AI is rapidly becoming the control plane of travel distribution, not just another feature layered onto existing systems. For years, we’ve debated who owns the customer, but AI reframes the question entirely: whoever controls the structured supply that AI agents can reliably transact against will increasingly shape demand.
What struck me while walking the halls was how much of the industry is still focused on AI at the presentation layer. Lots of chat interfaces, trip planners, and concierge bots. Interesting, but largely cosmetic. The real shift is happening deeper in the stack: the ability for machines to access clean, real-time, permissioned commerce data, rates, availability, policies, and ancillaries. AI doesn’t just change how travellers search; it changes who gets surfaced at all.
The industry spent a week talking about AI, but the real conversation should have been about plumbing. Because in an AI-mediated world, the most valuable thing in travel may turn out to be who does the best pipe work.
What I sensed at ITB is that many companies are adopting a fairly pragmatic wait-and-see approach, which, IMHO, is perfectly rational.
We are in the midst of a deep infrastructural shift: new interaction models are emerging, AI agents are moving from sandbox experiments to operational territory, and the first attempts to define protocols for the AI economy are emerging. In that environment, making long-term commitments to a specific stack is risky, so hesitation is becoming the new normal.
But the most interesting signal I saw actually comes before AI adoption: AI due diligence.
Before deploying models, agents, or copilots, organizations need to understand their data, their tech stack, and the operational processes underneath them.
If CRM, PMS, CDP, and booking systems contain fragmented or poorly structured data, no amount of AI layered on top will fix that. The same applies when systems do not interact properly, and when teams operate in silos with inconsistent SOPs. We talk a lot about tech silos and data silos, but not nearly enough about people silos.
So my ITB AI takeaway is simple: before we orchestrate agents, we need to orchestrate our data and our systems.
And, of course, our people.
At ITB Berlin 2026, one thing became unmistakably clear: the AI conversation in hospitality has matured. What was still driven by curiosity and hype last year has now turned into a much more strategic discussion. The industry is no longer asking whether AI matters, but what it can actually deliver in practice. And that shift is revealing a new reality: the true value of AI depends less on the technology itself and far more on the quality of the data behind it.
Walking through ITB this year, I kept asking myself one uncomfortable question: are we about to make the same mistake again — just with AI this time?
Everywhere you look there are agents, copilots, and automation. The halls are full of carefully scripted demo theatre and big promises. And to be fair, many people even recognize the real issue: the intelligence layer behind it all.
They say it out loud. But then the conversation stops. Because building that layer is the hard part nobody dares to touch.
Anyone can build a bot. The real work is organizing, governing, and evolving the data foundation behind it. It means understanding your own data and working with partners who strengthen that foundation — not vendors who simply drop another shiny agent on top.
Only then will AI agents actually have purpose.
I do hope that I’m wrong.
If everyone sees the problem, why does the discussion stop there? Is it fear? Complexity? Is the industry simply overwhelmed?
Because talking about the intelligence layer isn’t enough anymore. Someone actually has to build it.
We’ve seen this movie before.
Last time hospitality outsourced distribution. This time it might outsource its brain. And once that happens, good luck buying it back.
At ITB Berlin, the shift was clear: last year AI was mostly a topic of discussion; this year it is being adopted in real application stacks, with different levels of integration. AI is no longer a concept for the future, but something companies are actively putting to work.
The most memorable examples for me were native booking apps inside ChatGPT, which I saw at Apaleo (presenting its vision of the 'Autonomous Hotel), as well as Lighthouse. More broadly, the conversation around agentic AI has moved from theory to early reality. I learned that 15 to 20 hotel groups are already deploying agents in live environments, with adoption moving faster than anything previously seen in hotel technology. Smaller, more innovative operators are leading, while large chains remain more cautious.
The real takeaway from ITB 2026 for me is that the industry is talking about AI mostly as a technology story, not an economic one.
Most conversations focused on tools: chatbots, copilots, automation, agents, MCP. Much less attention was paid to how AI will change the economics of hotel distribution, or who will ultimately pay for the tokens these systems consume.
AI is already influencing how travelers discover and compare hotels. For now that influence is essentially “free”. But as AI agents increasingly sit between the traveler, OTAs, and hotels, they will shape decisions and eventually expect to capture value from the transaction.
At the same time, the final step remains unchanged. Checkout is still the domain of the booking engine. AI may guide the traveler, but the reservation must still be completed somewhere trusted, with the full commercial logic behind rates, inventory, policies, and payments.
This is where the hotel’s own commerce infrastructure becomes strategic. If AI connects travelers to OTA checkout, the OTA captures the booking. If it connects to the hotel’s website and booking engine, the hotel captures the value.
AI may influence the decision. The booking engine still closes the deal.
As more and more companies are adopting AI in their operations, it becomes clear to me where AI really shines, and that is primarily in the digital realm. That means that many processes, either before hotel stays, such as sales and marketing, or after hotel stays, such as feedback processing or gathering insights in loyalty programs are increasingly getting more streamlined with AI.
As a researcher focused on the physical work in hospitality, I still see very little that AI is doing for operational hospitality work. In the most physical lines of work, cleaning rooms and preparing food, AI applications focus on managerial work such as checking cleanliness or tracking food waste. I am keeping my eyes open for AI-based innovations that will help relieve those employees currently performing the heaviest tasks in our industry.
For me, the biggest takeaway from ITB Berlin was that while everyone is talking about AI, there are still only a few examples in hospitality where you can really see it working live and creating value. What stood out is how quickly innovation is moving, and that it is increasingly being driven by operators who want to solve real business problems, not just by vendors launching features. Use cases like McDreams and THE FLAG are strong examples of that, because they show AI moving beyond assistance and into autonomous action. That is the point where AI starts to have real meaning for hospitality, through faster execution, less friction, and more scalable operations.
AI at ITB 2026: Stop Watching the Demo. Start the Reset.
If ITB 2026 taught me anything, it’s that AI is no longer the headline: execution is. We’ve moved past the novelty of polished demos and into a season of deliberate, quiet action.
To navigate this, I believe organizations must embrace a fundamental triple-reset. First, a reset of mindset: AI isn't a bolt-on feature added to old workflows; it’s a structural transformation. Second, a reset of expectations: we must look past the shiny interface and focus on the weight of better decisions. Finally, a reset of deadlines: this shift is already outpacing traditional leadership cycles.
Why does this matter? Because hospitality doesn’t have an information problem, it has an execution bottleneck. We’re living in a world where data expires in seconds, and guest behavior shifts on a dime. For decades, we forced humans to learn complex software. Now, AI is finally forcing the technology to understand humans.
The future won't belong to the companies that talk the most about AI. It’ll belong to those who turn complexity into decisive action. In this cycle, waiting isn't "caution." Waiting is just drifting.
One AI takeaway from ITB Berlin that stayed with me came from a prominent panelist from Amadeus Ventures. They noted that AI will not replace SaaS platforms in hospitality but will increasingly be embedded within them. That perspective resonated with me.
Hospitality is a complex business built on both structured systems and deeply human interactions. Hotels rely on dozens of SaaS tools—from PMS and revenue management to guest messaging and operations—but those systems only work because people constantly feed them context, judgment, and real-world experience. AI can enhance these platforms by automating repetitive work, extracting insights from data, and accelerating decisions, but it cannot operate meaningfully in isolation.
What stood out at ITB is that much of the AI conversation is still driven by the technology community rather than operators. Many people building and promoting AI solutions have never worked inside a hotel—on a sold-out front desk shift or managing daily guest expectations.
The real opportunity lies in combining AI, SaaS platforms, and human expertise to improve operations and deliver better guest experiences.
At ITB Berlin, particularly across the vendor halls, it was evident that AI capabilities have reached a level where semi-structured interactions and limited reasoning are now reliably achievable. AI chatbots have become table stakes, while AI voice agents are emerging as the next frontier in customer interfaces. The immediate challenge for the industry will be moving beyond isolated point solutions and scaling these capabilities into fully operational, interoperable suites of AI tools that integrate seamlessly across the technology stack.
A parallel wave of innovation involves embedding large language model (LLM) assistants directly within existing enterprise systems—for example, chat interfaces within revenue management platforms that allow users to query the rationale behind automated decisions. While this represents a meaningful step forward in transparency and usability, I don't believe it should be viewed as the end state.
The more transformative opportunity lies in evolving these passive chat interfaces into proactive AI assistants. Rather than waiting for user prompts, hotel systems should actively guide hotel staff by surfacing relevant insights, highlighting areas requiring attention, and recommending data reviews or operational decisions. This shift from reactive tools to proactive decision support, has the potential to truly improve productivity and operational efficiency in hotel organizations.
The one thing I noticed is that I did not have so many discussions about AI specifically. The reason for that was because I was actually able to experience some genuine AI based tools within vendor technology vs. what has been up to this point, largely a discussion of potential.
I can think of three or four encounters at the show where I was looking at genuine AI based tools either as an introductory product or as an addition to an existing product. These were not first gen chatbots, or even chatbots in general. These were tools that could successfully automate or provide information useful to a customer or a staff member, produced through what we would identify as a real application of the AI capability that exists now. That's a great transgression from probably a couple of years of discussion.
Most important, was the introduction of dialogue with software applications. For me, this has been the greatest opportunity for effective interface between people and data. It's the one I've been waiting for and I believe is where the greatest promise lies in bringing software and people closer together.
One thing that really stayed with me after ITB this year is how much the conversation around AI has changed. Not long ago, most of the discussion was still about possibilities -what AI might do for the industry. This year felt different. The focus has clearly shifted to how companies are embedding AI into everyday operations.
Walking around the show floor, it was striking to see how many of those applications are already becoming tangible, whether it’s optimising pricing, managing supply more intelligently, or improving how companies support and connect with their partners. In a sector as complex as travel, where platforms link thousands of suppliers and distributors, the ability to process huge amounts of data and respond in real time is becoming incredibly valuable.
What matters now isn’t simply adopting AI tools. The real differentiator will be how well companies combine those technologies with strong data and a deep understanding of the travel ecosystem, whilst maintaining the uniqueness that their Brand represents in the eyes of their consumers. That’s what ultimately turns AI from an interesting capability into something that genuinely improves how the industry operates.
At ITB this year, AI was everywhere, I mean, literally everywhere. Mostly, though, as a label rather than a real capability. Many vendors claimed to “do AI,” yet after a few questions, it was revealed that most are simply layering interfaces on top of third-party tools like OpenAI rather than building systems that genuinely learn, adapt, and change decisions, or even better, using the OTA words: Secret Sauce. At the same time, the tech part of the industry is trying to squeeze AI into outdated SaaS pricing models, ignoring the reality of usage-based costs that can kill profitability. Hotels and tech vendors have not fully understood the new cost of doing business. Not that hotels should care tbh as they are buying a service.
But above all, we’re NOT solving new problems. We’re repackaging the same challenges in pricing, upselling, and forecasting with an AI sticker. Hotels were asking, “Do you use AI in your system?” and when I said, “For what exactly?” … the conversation stopped. Very little is asked: what the AI actually improves, where should I still leave the humans, or what happens if it fails and breaks (oh yes, one OpenAI outage, boom, most tech vendors no AI no more).
Meanwhile, a quiet tension is emerging. AI will reduce certain roles, but more critically, it will force a shift toward higher-value, decision-driven work. The risk is that without clearer thinking, the industry repeats its usual cycle by chasing buzzwords like big data, BI, and automation, all given to them. Honestly, I was disappointed….there was literally zero innovation. Same same but different.