IHG is rebuilding its hotel content into structured, machine-readable form so that AI tools surface its hotels, and the independent hotel has no such project. That pushes agencies towards the part of the portfolio machines read badly, which is a good corner only if the agency's own record of those properties is better than the industry's.
Key facts
- One new platform, due later this year, to change how IHG structures and manages hotel content — Heather Balsley, IHG's Chief Commercial & Marketing Officer, to Skift.
- Two aims for it: the right hotel information on the right channels, and stronger signals that help AI tools surface and recommend IHG hotels — the same Skift interview.
- 25 different ways: how one hotel is spelled by the agencies booking it — the author's operations archive.
The news. IHG says it will introduce a new platform later this year that changes how it structures and manages hotel content. The aim, as Heather Balsley, its Chief Commercial & Marketing Officer, puts it to Skift, is to get the right hotel information onto the right channels and to strengthen the signals that help AI tools surface and recommend its hotels. Skift's own read is blunter: IHG is rebuilding its content into structured, machine-readable form, and it is stating the bet before it has the proof.
In the same interview Balsley draws the line IHG intends to hold: "new discovery channels don’t replace direct relationships with travelers." Read from an operator's desk, that means IHG will do the work the machines ask for and still expects to keep the guest. IHG can afford to. Balsley calls this an area of significant investment, and a group that size can run the rebuild as a funded project.
Why can't the independent hotel follow?
The independent hotel sitting in an agency's portfolio has no such project. Nobody there will normalise room names, restate the cancellation policy in structured fields, or tag amenities so a model can parse them. That hotel will keep being described the way it always has been: by whoever last typed it into a system. Machine-legibility is an organisational problem before it is a formatting one, and small properties do not have the organisation.
Then there is the part no system has ever recorded. What makes an independent hotel sellable is usually something a person knows: which rooms face the service yard, which manager will hold rooms on a promise, which property takes a large Gulf family without an argument. That knowledge lives in a reservations agent's head and, at best, scattered through old email threads. No structured record carries it.
Where does the mess start?
The difficulty shows up long before you reach the small properties. From my own archive: one hotel is spelled 25 different ways by the agencies booking it. Same property, written by professionals whose job is booking it, in a trade that has had decades to agree on a name. Put that next to a chain's clean content feed and you can see which one a model matches first.
Why does it matter for the agency?
What this changes at the desk is narrower than the headline suggests. Chains will get more visible in AI answers, and they will get there first. That pushes agencies towards the part of the portfolio that machines read badly: the independents and small groups, whose value is judgement rather than metadata. That is not a bad corner to end up in. A recommendation engine cannot copy what it cannot read. The condition is that your own record of those properties is better than the industry's. If the knowledge exists only in one person's head, it leaves when they do.
Source: Skift — IHG's Chief Commercial & Marketing Officer on making hotels legible to AI, by Kanchi Jain.