AI demand signals for revenue management: two clocks and one new channel
Revenue decisions run on demand you can see. This page covers the two demand clocks a hotel already owns, and what the AI channel is worth today measured rather than predicted.
AI sessions measured across the sites we track, last 30 days.
Quick answer
AI demand signals for hotel revenue management are the measurable inputs an assistant driven guest leaves behind: the searches on the booking engine with their stay dates and lead time, and the sessions arriving from AI assistants. Across the sites we track, in the 30 days to 4 September 2026, ChatGPT was around 90 percent of the AI channel and more than twenty times the next assistant, and the panel at the top of this page carries the live counts. The AI channel is small next to organic search for most properties today and it is already large enough to be measured instead of assumed.
Two clocks, and mixing them is the classic error
A revenue manager reads demand on two different clocks and they are never the same picture. The first is when people search, which is the window on your date filter. The second is which dates they search for, which is a calendar of stay nights. A quiet week of searching for a busy August is not the same signal as a busy week of searching for a quiet November, and a dashboard that shows both with the same four tiles invites exactly that confusion.
reporte.ai keeps them apart on purpose. Every block states its scope in its own title, the calendar block updates as you move through months, and a bridge line says explicitly that one is a slice of the other. The metrics that appear on both clocks are named identically so nobody has to guess whether two similar tiles mean the same thing. The searches themselves, with their lead time and stay dates, are in demand.
The AI channel, sized rather than predicted
Here is the order the new channel took in those 30 days to 4 September 2026: ChatGPT, then Copilot, Perplexity, Gemini and Claude. One assistant is more than twenty times the next and around 90 percent of the total, which has a direct operational consequence. A revenue manager does not need five monitoring programmes, they need one channel watched closely and four watched cheaply, so that the week the distribution shifts is a week they notice.
Two disciplines keep this honest. A channel reading zero is a measurement suspect before it is a business conclusion: our own Gemini number sat at zero for a day in early September 2026 because of a bug in how we read its citations. And a count without its share cannot be read at all, so every AI session appears next to its percentage of the total visits of the same site. What that traffic then does, page by page and down to the booking step, is in the AI visitor journey.
Putting a price on the signal
The forecast side of this is cheap to keep running, which is unusual for anything a revenue manager buys. Measuring 57 questions across ChatGPT, Claude, Gemini, Perplexity and Google AI Mode cost 10.46 USD in the pass of 3 and 4 September 2026, so a weekly cadence is roughly 544 USD a year at 52 passes.
The per question cost decides where the budget goes: 0.0040 USD on Google AI Mode against 0.0796 USD on ChatGPT, a factor of almost 20, and in that pass the cheap surface produced appearances in 14 of 57 questions against 10 for the expensive one, across 7 domains of our own travel network rather than one hotel, and one pass alone proves nothing. That is a rate of measurement a property can commit to without a business case, which matters because the value of a demand signal comes from having it every week and not from having it once. The commercial framing of the whole channel is in AI search marketing and pricing is on the pricing page.
FAQ
It is big enough to measure and too small to price rooms on. Across the sites we track, ChatGPT was around 90 percent of the AI sessions in the 30 days to 4 September 2026 and the panel at the top of that page carries the live counts. Treat it as a channel you are sizing and watching, not as an input to tonight rate.
When people search, which follows the date filter you set, and which nights they are searching for, which follows the stay calendar. A block that does not state its scope is not a cosmetic problem: one of our own screens announced that no searches had been recorded while 2,377 measured searches sat in the database. Every block now carries its scope in its own title.
No. It measures the demand signals that reach your own website and booking engine, with their stay dates and lead time, and it measures the AI channel alongside them, which the panel at the top of that page shows live for the last 30 days across the sites we track. It is not a forecasting model and it does not connect to a rate shopper or a PMS to move prices.
A full pass of 57 questions across five surfaces cost 10.46 USD on 3 and 4 September 2026, so a weekly cadence is about 544 USD a year at 52 passes. The per question spread runs from 0.0040 USD on Google AI Mode to 0.0796 USD on ChatGPT, so a tighter budget cuts frequency on the expensive surface first.
Check the measurement before you draw a conclusion. Our own Gemini figure read zero for a day in early September 2026 because of a bug in how we parsed its citations, not because Gemini ignored anyone. A closed zero on a single channel is a suspect, and a channel that is genuinely dead stays dead across several passes.
Because the same number means opposite things at different scales. Forty AI sessions is a real channel on a property with 500 monthly visits and rounding error on one with 50,000, so every count in the panel appears with its share of the visits of that same site.