See which competitor ChatGPT names instead of you, and the sources deciding it.
8 live ChatGPT questions · about a minute · no card · email opens the sources
Snapshot · 8 buyer questions
Lever in 3 and Ashby in 2. Raw counts, not a score.
What ChatGPT cited, ranked
“What's the best applicant tracking system for a company of around 200 people?”
For a company of about 200 people, Greenhouse is the most frequent recommendation — it has the deepest structured-hiring workflows and connects to most HR stacks. Lever and Ashby are the usual alternatives, with Ashby often named for its reporting.
01 · How it works
01 · Free snapshot
See who wins
8 real buyer questions, asked to ChatGPT. You see who it named, in which answers, and the sources it cited.
02 · Monitor
Work the ranked queue
Sources are ranked by how often they decided an answer you lost. Start at the top: a review profile, a roundup, a comparison page.
03 · Monitor
Check that it moved
Every Monday the same questions run again, beside a control question that should not move. If your count rises by more than the control, you see it with its denominator. If it does not, we say so.
02 · Run it
8 live ChatGPT questions · about a minute · no card · email opens the sources
03 · Pricing
Snapshot
Free
Monitor · recommended
$99 a month
The 14-day trial is free. Thirty days to change your mind after your first payment, refunded without a call. No annual lock-in.
04 · Method
We ask the logged-out ChatGPT web interface the same thing your buyer sees, not the developer API, which answers differently. The free snapshot asks 8 questions once. In each weekly benchmark, Monitor asks 60 core questions up to 5× on ChatGPT (extra samples wherever answers disagree) and 5× on Perplexity — up to 600 observations. That exposes variation and shows whether both engines agree; it is not enough to estimate a stable probability for one question.
After you act, Prescy re-tests the same question against a held-out control question and says plainly when a move is indistinguishable from platform drift — not a causal claim. We show the raw denominator and do not invent a confidence interval. If another tool will not say which of these they do, ask them. See the full methodology →
New to this? What AI search actually is → — why reviews, roundups and comparison pages decide the answer, and llms.txt and schema do not.
The free snapshot returns
05 · Questions
Prescy is an AI visibility tool for B2B marketing teams. It shows who is beating you on the buyer questions that matter in AI search — which rival ChatGPT and Perplexity recommend — and, on Monitor, what to do next and whether a later check moved.
No. You can run a live ChatGPT snapshot of buyer questions and see who was named, without creating an account. Monitor is the paid plan: the ranked source queue, weekly rechecks, and a re-test of what changed.
AI assistants are more likely to surface brands when they can find specific, citable information that directly answers the buyer's question. Depending on the measured gap, useful actions can include improving product and category detail, publishing a relevant comparison or buying guide, answering a specific FAQ, adding substantiated customer proof, or clarifying structured data. No individual tactic guarantees an AI mention.
The steps most teams skip are writing the fix and measuring it. Prescy repeatedly asks the questions your customers actually ask on ChatGPT and Perplexity, shows which rival keeps winning, and ranks the reasons you're losing — each marked Confirmed or Emerging with the raw fraction, and a source that comes through on both engines counts for more than one that only repeats on one — so you know which are real before you act.
Prescy deliberately doesn't use a 0–100 visibility score — every vendor calculates one differently, none are comparable, and they mostly exist to make a dashboard look finished. Instead it reports how often ChatGPT and Perplexity recommend a rival over you across repeated samples, with the raw fraction behind every finding ("recommended in 12 of 600 conversations") so you can see what's real rather than trusting one number.
Usually because the AI has more specific, quotable material about them: comparison pages, reviews, documentation, press coverage and community discussion. AI answers favor brands whose content directly matches the question asked. The fix is rarely 'more marketing' in general — it's changing the page ChatGPT already used, or getting onto it. Prescy shows the buyer question, who it recommended instead, the cited URL, and the excerpt from the answer — then re-checks whether AI started recommending you instead.
Yes — AI assistants that browse (Perplexity, Google AI Overviews, increasingly others) draw heavily on pages that already rank well, so strong SEO feeds AI visibility. But it's not sufficient: chat models also answer from training knowledge where rankings don't exist, and one synthesized answer replaces ten blue links, so being #4 can mean being invisible. Treat AI visibility as its own channel with its own measurement, built on the same content foundations.
Yes — Prescy's free snapshot reads your website, asks ChatGPT a set of live buyer questions, and shows how often you were named plus which domains were cited. A typical snapshot takes about a minute. You don't need an account to see the counts; the finding stays private by default. Monitor unlocks the ranked source queue and weekly measurement.