AI language models are changing how we interact with systems and technology. The point and click model is evolving for the first time in 30 years. This means the battle for the customer interface is back.
We went through this when the internet arrived. Before that, discovery happened through travel agents, brochures, hotel brands etc. OTAs mostly won in the point-and-click era of search, filter and bookings.
With conversational interfaces things are going to change. Instead of scrolling through 200 hotels, it is much easier to say: “Find me a quiet boutique hotel in Rome, walking distance from good restaurants, with a pool suitable for toddlers, and a room that actually works for a family of six”, (this is probably an impossible request, tell me if you found one).
But I’m not sure the infrastructure underneath can deal with the data needed. Today things are just noted as “Pool: Yes/No” “Connecting rooms: Yes/No” etc.
Current hotel distribution systems are designed around a simple concept: availability, rates and inventory (ARI). That works when the guest is doing all the searching and reading and clicking and filtering. It works much less when an AI is doing that as the data simply isn’t granular enough.
So even if OTAs become the plumbing behind AI, basic ARI will no longer be enough. An AI can only give a great answer if the underlying systems actually has the answer. This is where attribute-based selling might finally have its day of glory. For years, the idea has been discussed as a better way to sell rooms based on specific features rather than fixed room types. But it was going to become an interface problem, nobody wants to tick 8 boxes to build their room-rate. AI on the other-hands doesn’t mind, conversational search is basically attribute-based search on steroids.
However, almost none of the three letter systems (OTA, GDS, PMS, CRS) were built for that. Adding an AI layer will essentially just convert a language query into a simple search string (ie “Hotel in Rome with pool”, Pax:6, Date:03/27/2028) and that won’t really answer the question of the guest.
So it’s going to take a serious rethink and re-architecture of all the plumbing that connects hotels and distributors. And it isn’t going to generate any immediate revenue. Will OTAs have the patience to do it? More importantly will their investors?
