Trip.com, OTAs, AI agents, Booking Holdings, Expedia, Skift, APIs, MCP, GEO, SEO, trust

Ever since ChatGPT appeared, OTAs have repeated like a mantra that their great advantage over artificial intelligence is traveler trust. Booking Holdings and Expedia say it loud and clear: users won't trust a hallucinating LLM, they will turn to a known brand to complete the purchase.

But while they keep looking in the rearview mirror, Trip.com is driving in a different direction. Its executive chairman, James Liang, dropped it on the latest earnings call: our goal is not just to be the traveler's favorite app, but to become the trusted infrastructure for AI agents. And that changes the game.

As Skift reports, the real battle is not between OTAs and chatbots. It's between OTAs to be the invisible provider that feeds those chatbots. When an AI agent looks for a flight or a hotel, which OTA will it ask first? That's the prize.

Trip.com's pivot: from D2C to B2B for machines

What Liang proposes is a dual strategy: keep the consumer app running, but at the same time expose verified inventory, real-time pricing, and transaction rails through APIs, the MCP protocol, and agent frameworks. In other words, Trip.com wants to be the backend for LLMs.

This is not just a defensive move against disintermediation, which can come both from assistants and from suppliers that decide to bypass OTAs. It's an offensive move. While Booking and Expedia invest in brand campaigns to retain human users, Trip.com optimizes for the machine user.

And watch out, because the first-mover window is closing. OTAs that manage to integrate first into AI agent workflows will have a structural advantage that will be hard to overcome.

Daily Cartoon

notitur.com

From SEO to GEO: the new art of visibility

If SEO (search engine optimization) was the great obsession of the last decade, now comes generative engine optimization (GEO). It's no longer enough to appear in the top Google results. You have to be the source that the language model chooses to answer a query.

For small hotels and airlines, the challenge is bigger: without the inventory volume of a global OTA, how do you get an AI agent's attention? The answer lies in strategic partnerships and having clean, real-time, well-structured data.

The trust model is inverted

Historically, trust in the travel sector flowed from supplier to OTA and from OTA to traveler. Now there's a new link: the AI agent. And its trust is not won with pretty logos or a 24/7 call center. It's won with technology, reliable APIs, and the ability to prove that your inventory is the most complete and accurate.

As the Skift article rightly points out, the key questions that will define the future of distribution are: what is the Model Context Protocol and how does it change the landscape? can a boutique hotel compete for an AI agent's attention against Expedia's scale? will major LLMs strike exclusive deals with a single OTA?

My take: traveler trust is smoke, algorithm trust is real

Forgive my skepticism, but we've been hearing for years that "customer trust is our greatest asset." And sure, it's true... until the customer gets used to an AI agent handling the entire trip without opening a single app. Then, brand loyalty fades. What really matters is algorithm trust.

OTAs that survive this transition will not be the ones that talk best about themselves, but the ones that let themselves be read best by machines. And in that race, Trip.com has just overtaken on the right.