AI, Live

Learn. Unlearn. Relearn.

A running series on artificial intelligence in hotel operations, pricing, and distribution — written from an owner’s seat rather than a vendor’s.

It proceeds from a single claim. This is not an efficiency story. Artificial intelligence is altering what a hotel costs to operate — structurally, and in a way effort cannot replicate. When the market recognizes that, it will not simply pay for the additional income. It will pay a different multiple for it.

Hotels trade at wider capitalization rates than any other institutional real estate class, because a hotel is not a building so much as an operating business that reprices its inventory nightly and carries a payroll to do it. That spread is the market’s standing charge for operational intensity, and labor is its largest component. The market has already shown, once, that it will pay for less of it — select-service commands strong pricing precisely because it removed the complexity full-service could not shed. A re-rating of the asset class is a larger event than any operating saving inside it — and it accrues to whoever is positioned before the market marks it.

That is the paradigm we are tracking. The technology is the mechanism. The multiple is the argument.

The market notes are conclusions. These are readings — taken while the instrument is still moving, published in the knowledge that a development next quarter may make them obsolete. When that happens we will say so here rather than quietly revise.

No. 01 · August 2026
The Second Re-Rating

The industry is discussing artificial intelligence as a cost programme. That is the smaller half. The larger half is that an asset class carrying a permanent discount for operational intensity is a candidate to be repriced — and the market has already done this once, when it paid up for select-service.
Working Model
The Labor Model

The argument in this series is that a structural reduction in labor does not merely improve a statement — it changes what the asset is worth on a multiple basis. This model lets you run that against a property you actually own.

InputsMarket tier, service level, keys, rooms revenue, labor ratio, technology cost, exit cap
ReturnsNet NOI impact, payback, value created, value per key, and where in the operation the reduction actually sits
Open the model →
Learn, unlearn, relearn — after Alvin Toffler, Future Shock, 1970, crediting Herbert Gerjuoy.
Entries are dated. Positions taken here are current as of publication and subject to revision as conditions change.
Both Series

One list. Market notes and AI, Live.
No cadence. No noise. Unsubscribe by reply.
Responses

This series is written to be argued with. If you are deploying any of this — or have tried and found it wanting — we would like to hear it. Responses are read by the principal and, with permission, occasionally published here. Write to us →