When travelers ask ChatGPT where to stay in Los Angeles, which hotels does it name?

It measures the AI, not the hotel. We ask ChatGPT the same questions a traveler would — 'best hotels in NYC', 'where to stay in Williamsburg' — and record which hotels it names and in what order. A hotel's placement rate is the share of those answers that name it. So the board is a readout of what ChatGPT tells travelers, not a quality score we invented.

ChatGPT today — the same web-connected assistant a traveler gets, not an offline model. Different engines answer differently, so we never blend them into one number; each will get its own board as we add it. More engines and more cities are on the way.

No. Google ranks pages; this measures whether an AI assistant names your hotel when it answers a traveler in plain language. The inputs overlap — a clear, accurate web presence helps both — but a strong Google position doesn't guarantee ChatGPT mentions you, and some hotels that rank modestly on Google still show up often in AI answers. It's a separate shelf, and it's the one more travelers are starting from.

The overall board comes from citywide questions ('best hotels in New York'). Each neighborhood board comes from questions about that area ('hotels in Williamsburg'). A hotel appears on a board because ChatGPT named it for those questions — not because of a location tag we assigned.

Because ChatGPT named it when answering questions about that area. A board reflects what the engine associates with a neighborhood, not our own map — so it can put a SoHo hotel on a Tribeca list, or a marquee name on several boards at once. That association is exactly what we're measuring, so we show it as the engine gave it rather than filtering to our own tags.

By the budget in the traveler's question. We ask ChatGPT things like 'budget-friendly hotels in NYC' or 'luxury hotels near Central Park' and record which hotels it names for each, then group them into the board that matches. A hotel sits on the Luxury board because ChatGPT surfaced it for luxury questions — the tier is the engine's read, straight from its answers. Because of that, the same hotel can appear on more than one board: name it for both mid-tier and luxury questions and it shows up on both. That overlap is the engine's own doing, and it's part of what we're measuring.

Each board only lists the hotels ChatGPT actually named for that board's questions, so the count reflects how much attention an area or budget gets. A busy hotel district turns up far more names than a small, specific neighborhood, and citywide questions surface the most of all. Boards with too few answers to be reliable aren't published at all. So a shorter board isn't missing data — it's the engine naming fewer places there.

Three ways to read the same answers. Placement rate is how often ChatGPT names the hotel at all. Recommended rate is how often it actively recommends it, not just mentions it in passing. Top-pick rate is how often the hotel is the first or #1 choice. Use the 'Rank by' menu to sort by any of them.

A placement rate — 'named in 41% of relevant answers' — is comparable across boards and honest about coverage; a raw count isn't. We publish the percentage and the position on the board, and keep the sample sizes and confidence ranges on the methodology page.

No. There's no paid placement — you can't buy a position or a higher percentage. The board is the engine's own output; the only way up is for ChatGPT to name your hotel more often.

They get no edge. The board is ChatGPT's own answers, with no paid or preferred placement, so a customer appears only when the engine names it — exactly like every other hotel. We note internally which hotels are customers for transparency, never to promote them. The one set we hold back is a group of hotels kept off the public board so we can check our own accuracy.

No — we're not affiliated with OpenAI. We ask ChatGPT the same questions travelers do and record its public answers. This is Sigtrip's own independent measurement, not OpenAI's ranking.

We re-run the census in waves and stamp every board with the run it came from. Ask an AI the same question twice and you can get different answers, so placement shifts from run to run — a board is a snapshot of what the engine said then, not a permanent verdict. Watch the trend across runs, not a single week.

ChatGPT didn't name your property for these questions in this run. That's the same thing a traveler experiences: ask the AI where to stay, and if your hotel isn't in the answer, it isn't on their shortlist. A free scan shows where you stand and what's driving it.

Because the board reports what ChatGPT says about you, not what's true at your property — it measures the engine, not the hotel. If it credits you with a rooftop bar you don't have, or misses the spa you do, that's how ChatGPT currently reads your web presence: your own site first, then the review and travel sites it cites. When that read is wrong, it's the same wrong answer a traveler gets. A free scan shows exactly what ChatGPT gets wrong about you and where it's reading it from, so you can fix it at the source.

Engines build their answers from what they can read about you — your own site first, then the review and travel sites they cite. Making that information clear, accurate, and easy for an AI to pull is what lifts placement over time. It moves as the engine does, so no one can promise you a spot; a scan shows what ChatGPT sees about you today and where the gaps are.

This ranking is a wide, shallow read of a whole market — many hotels, a fixed set of questions, public. A scan is the opposite: a deep read of one hotel, checking what AI knows and gets wrong about you across far more questions, with the specifics behind it. The ranking tells you whether you're in the conversation; the scan tells you why, and what to fix.

Those are directory details — the Google guest rating and the hotel's own website — shown only for context and to make competitor research quicker. They're not part of the AI measurement and never affect placement; the ranking itself is only ever ChatGPT's own answers.

We assembled the list of hotels in each market from public directory data, then measure how ChatGPT talks about them — we don't need a hotel's permission to include it, and being on the board doesn't make a hotel a Sigtrip customer. If a detail about your property looks wrong, tell us at hello@sigtrip.com and we'll fix it.