But Hotels Still Have a Chance to Write a Different Ending.
Last week’s news about OpenAI scaling back its plans for direct in-chat purchases was interpreted quickly — and loudly.
Booking Holdings’ share price jumped.
The market declared the OTA model safe again.
And many in hospitality quietly exhaled.
For a moment, it looked as if the biggest disruption to hotel distribution in twenty years had just been postponed.
The idea had been simple, and potentially explosive:
A traveler asks an AI assistant for a hotel.
The assistant recommends a few options.
The traveler completes the booking directly inside the AI interface.
No redirect.
No OTA.
No Booking.com sitting between hotel and guest.
That scenario has now been shelved — at least for now.
Instead, AI platforms like ChatGPT will recommend and refer.
And someone else will close the transaction.
Right now, that someone is still Booking.com.
The Real Reason OTAs Still Matter
The reason is not brand power.
It is infrastructure.
Payments across hundreds of currencies.
Inventory synchronization across millions of properties.
Cancellation rules, dispute handling, merchant agreements, fraud prevention.
These are not product features.
They are operational systems that took Booking.com more than two decades — and billions of dollars — to build and maintain.
When OpenAI stepped back from building its own checkout layer, it effectively acknowledged something the hospitality industry already knows:
Owning discovery is easy compared to owning transactions.
And right now, OTAs still own the transaction layer.
AI Changes Discovery — Not the Need for Infrastructure
What AI is changing is not booking mechanics.
It is the moment before them.
Discovery.
Travelers increasingly start their journey by describing what they want:
- “A quiet design-led hotel near the sea.”
- “A lively boutique with social energy.”
- “A calm place to disconnect for a few days.”
AI interprets these requests and recommends hotels that fit the meaning behind them.
This is a fundamental shift.
Discovery is moving:
from search → interpretation
from keywords → meaning
from catalogues → understanding NOW HOTELS 1-page master
But interpretation still needs somewhere to send the guest.
And if hotels do not build the infrastructure to receive that guest directly, the referral will simply route through the systems that already exist.
Which brings us back to Booking.com.
The Metasearch Lesson Hotels Forgot
This dynamic should feel familiar.
When metasearch platforms like Trivago and Google Hotel Ads appeared, the industry believed the OTA model would weaken.
Hotels would bid directly for traffic.
Guests would compare options neutrally.
OTAs would lose control of discovery.
Instead, the opposite happened.
Booking.com and Expedia outspent everyone.
Metasearch became another distribution layer largely funded by hotel commissions.
The discovery layer changed.
The transaction layer did not.
And the companies that already owned the infrastructure absorbed the new channel.
AI could easily follow the same pattern.
The Referral Economy Is the New Battleground
If conversational AI becomes the primary discovery interface, the economics of travel distribution shift upstream.
The key question becomes:
Where does the AI send the traveler when they are ready to book?
Possible destinations include:
- OTAs
- hotel direct websites
- regional travel platforms
- channel manager booking pages
- new affiliate-driven networks
For the first time in years, hotels have a theoretical path back to the guest.
But theory alone will not change the outcome.
Because a referral engine only works if there is a destination ready to receive it.
The Real Risk for Hotels
The danger is not that OTAs remain strong.
They will.
The danger is that hotels repeat the same strategic mistake they made in previous distribution shifts:
Waiting for someone else to solve the infrastructure problem.
Historically, hotels have relied on intermediaries to build the systems that connect them to travelers.
Search engines.
OTAs.
Metasearch platforms.
Brand loyalty ecosystems.
Each time, the intermediary that built the infrastructure captured the economic leverage.
AI will not change that dynamic.
Unless hotels decide to participate in building the next layer themselves.
A Different Path Is Emerging
What is different today is that the tools now exist for hotels to reclaim more control — if they choose to.
AI is dramatically lowering the cost of interpretation.
Platforms like NOW Hotels focus on the upstream problem:
making hotels understandable to AI systems and to travelers.
Because most hotels do not have a marketing problem.
They have a clarity problem.
They are hard to describe, hard to distinguish, and therefore hard to recommend. What can NOW Hotels do for you …
If AI cannot confidently interpret what a hotel is — quiet or lively, social or contemplative, urban or retreat-like — it will not recommend it.
Clarity becomes the first layer of infrastructure.
The Second Layer: Distribution Without Dependency
But clarity alone is not enough.
Hotels also need pathways that allow AI referrals to land somewhere other than an OTA checkout page.
This is where emerging models — including affiliate-driven networks such as Affiliotel — begin to matter.
These systems create a transaction layer that still allows:
- hotels to own the guest relationship
- bookings to land on the hotel’s own website
- data to remain with the property
- commissions to remain lower and more aligned with hotel economics
They are not replacing OTAs.
But they are building alternatives.
And alternatives are the only thing that ever changes distribution economics.
Why This Moment Still Matters
OpenAI stepping back from in-chat transactions did not save Booking.com.
It simply bought the incumbents more time.
AI discovery will continue to expand.
Recommendation engines will become the first step in travel planning.
And eventually, one or more platforms will succeed in owning both discovery and transaction inside the same interface.
When that moment arrives, the outcome will depend on the same question that has shaped every distribution shift in hospitality:
Did hotels build their own infrastructure?
Or did they wait again?
The Ending Is Not Written Yet
Every previous chapter in hotel distribution ended the same way.
A new discovery layer emerged.
Hotels celebrated new access to guests.
Intermediaries built the infrastructure.
Intermediaries captured the margin.
But this time the tools — and the knowledge — are different.
AI can interpret identity.
Platforms like NOW Hotels can make that identity legible.
Networks like Affiliotel can route bookings back to the hotel.
For the first time in decades, hotels have the ingredients needed to compete for the transaction layer.
Whether they use them is another question.
The Companies Worth Watching
The companies that matter over the next five years will not necessarily be the biggest travel brands.
They will be the ones building the missing infrastructure between:
AI discovery
and
hotel direct bookings.
Because the moment an AI assistant recommends a hotel, someone still has to close the deal.
The companies that control that moment will shape the next era of hospitality distribution.
And if hotels choose to participate in building it — instead of outsourcing it again — they may finally write a different ending.
DOUBLEDOWN Hotels 
Leave a comment