The Deal

On July 31, 2026, Expedia Group announced it had acquired Layla, a Berlin-based AI trip-planning startup founded in 2023 (Expedia Group investor announcement). Financial terms were not disclosed. Skift, which broke the story, reported that Layla had raised about €5 million total, employed roughly 25 people, and would continue to operate under its own brand while its team worked alongside Expedia’s Expedia.com, Hotels.com, and Vrbo brands (Skift exclusive).

By any obvious measure, this is a small deal. A small European team, modest funding, no disclosed price. The kind of transaction the travel-industry press usually notes and moves on from.

It is not a small deal. It is a signal.

Why Expedia paid for a 25-person chatbot

Layla is, in the words of its own founder, an “AI-native trip planning and booking platform.” Its five million users have exchanged more than thirty million messages with it. It has generated over two million trip itineraries. And it has processed over $1 billion in planned trip value (Layla milestone announcement, March 2026).

The single most interesting number in Layla’s own data isn’t any of those. It is this one: over 40% of Layla’s users start a trip without a destination. In 2023, that number was around 12%. In a little over two years, the share of travelers willing to say “I want to go somewhere, help me figure out where” nearly quadrupled.

That is not a small behavioral shift. It is a change in what travelers expect from the moment they start planning.

Traditional travel platforms like Expedia, Booking.com, Kayak, and Google Travel,  all assume the traveler starts with a destination in mind. Their search boxes ask “where?” first. Their entire information architecture is built on that assumption. What Layla’s data suggests is that a growing share of travelers no longer want to be asked that question at the start. They want to describe a mood, a season, a budget, a set of preferences, and let something figure out where.

That is what a conversational AI is good at, and it is what a search box is not. Expedia’s acquisition is a bet that the next five years of travel discovery are going to look more like a conversation and less like a filter panel.

Expedia is not alone

Expedia’s deal is the most public data point, but it is part of a broader pattern.

• Booking.com launched an AI trip planner in ChatGPT in October 2025, one of seven launch partners for OpenAI’s Apps SDK (OpenAI Apps announcement).
• Kayak rebuilt its front door around Kayak Ask AI, a conversational search layer that runs on top of its existing metasearch engine.
• TripAdvisor shipped AI-driven itinerary planning inside ChatGPT in the same October 2025 wave.
• Priceline launched Penny, its own conversational AI, in 2024, and has continued to expand its capabilities.
• Mindtrip, one of the most-cited AI travel planners of 2026, moved from planning-only into booking during the same year, launching Mindtrip Flights in May 2026 through a Sabre and PayPal partnership and Mindtrip Stays in July (monkeytravel comparison, 2026). What began as a chatbot became a chatbot that closes the transaction inside its own app.
• Hopper, Wonderplan, and a dozen smaller startups are all iterating on the same core idea: replace the search box with a conversation.

Alongside those strategic-AI products, a smaller set of AI travel tools is being built for a different moment in the traveler’s journey. SlickTrip, for one, is not trying to help travelers decide where to go. It is being built for the traveler who has already decided, and now needs to find the best fare, watch a better seat, and know when to act. That is a very different job from the one Layla and Mindtrip are optimized for, and it is the job the current wave of strategic-AI acquisitions is largely leaving alone.

What Skift’s reporting on the Expedia-Layla deal makes clear is that the strategic direction inside Expedia is not just “add a chatbot to the homepage.” It is a shift toward what Expedia now calls a “multi-agentic” architecture a set of specialized AI agents that talk to each other and to the traveler at different points in the journey. Layla’s team and technology are being folded into a broader move away from the earlier Romie concept, an all-in-one concierge that Expedia has since deprioritized (Skift).

The lesson Expedia appears to have taken from Romie is that a single AI that tries to do everything for a traveler doesn’t work. What might work is a network of specialized AIs, each good at one thing, coordinated by an orchestration layer.

That reframing matters. It means the OTA arms race is not really about who has the best chatbot. It is about who has the best portfolio of AI agents for different jobs.

What the money is chasing

Look at who was on Layla’s cap table before Expedia showed up. The March 2026 investment round included United Airlines Ventures, Baidu Capital, and INCE Capital, joining existing investors M13, Firstminute Capital, and SparkLabs (Layla, March 2026).

An airline venture arm, a Chinese search-engine venture arm, and a Silicon Valley growth investor were all making the same bet on the same category, at the same time. The bet is that the discovery-and-planning layer of travel, the earliest, most emotional, least-monetized moment in the traveler’s journey, is about to be worth an enormous amount of money to whoever owns it.

That bet is almost certainly right. The discovery layer is where preferences are formed, where destinations are chosen, where the eventual booking decision is shaped. If you own that layer, you have a claim on every downstream transaction. It is why Booking.com built its own AI. It is why Kayak did. It is why Expedia is now buying its way in.

But there is a second layer, further down the funnel, that gets almost none of the same attention.

The moment nobody is talking about

Strategic AI helps a traveler figure out where to go and what to do. Booking.com in ChatGPT can find hotels. Layla, inside Expedia, will help a traveler brainstorm the shape of a trip. Kayak Ask AI can compare options. All of these products serve the moment when a traveler is early in the journey, still exploring.

Then there is a different moment. The traveler already knows they are going to Lisbon. They already know the week. They already have the trip roughed in. What they need now is a specific answer to a specific question: is this flight a good price, and should I book it now?

That question does not need a conversation about their preferences. It does not need an itinerary. It does not need a mood board. It needs a fast, confident answer and a way to act on it.

Almost none of the AI travel products being built today are optimized for that moment. They are optimized for the exploration end of the funnel, because that is where the money is chasing. A small handful of products, SlickTrip among them, are being built specifically for the transactional end, but they are dwarfed in coverage and investment by the strategic-AI wave. 

The transactional end, the moment when the traveler is ready to buy, and just needs to know whether now is the right time, is a different problem, and remains under-invested and under-covered relative to the exploration end.

There is a related gap worth naming. Even for a trip that has already been booked, almost no consumer AI product will change the booking on your behalf: rebook a cancelled flight, adjust a seat and pay the fare difference, request a fare-drop refund. The reasons those actions are still manual are structural, including authentication, fare-rule complexity, payment consent, and liability, and worth their own piece. We wrote one: Why AI Can Book Your Flight but Not Rebook It.

The Expedia-Layla deal is a bet on the exploration end. It is a good bet. The industry is going to look very different in three years because of these acquisitions.

But if strategic AI, the kind Layla and Mindtrip built, and the kind Expedia just bought,  is now table stakes, the tactical moment is where the next round of interesting AI travel products will get built. SlickTrip is one of the products being built specifically for that moment, with a forthcoming ChatGPT app for seat alerts, fare alerts, and travel bucket list already in OpenAI’s app review process (SlickTrip for ChatGPT).

What “tactical AI in travel” actually means, and why the distinction matters, is worth a piece of its own, and it is coming next.