What ChatGPT actually says about your client's competitors — and how to check

Every local-SEO agency already does a version of this for Google: check the map pack, note who's above your client, work out why. The AI-search version of that exercise is just as doable, and most agencies haven't started doing it yet — which is exactly why it's worth being early.

The check itself

Open ChatGPT, Gemini, or Perplexity and ask a plain question a customer would actually ask: "who's a good [category] in [town]", or something closer to how people in that trade get searched for. Note exactly who gets named, in what order, and what's said about each. Do it again with a slightly different phrasing — the answer isn't always identical, which is itself useful information about how stable a business's visibility is.

RankSignal's free scan tool (ranksignal.co.uk/scan) runs this same check for a specific business name, one model, no login, and shows you the literal prompt and answer — a faster way to do the first pass than manually round-tripping through a chat window, especially when you're doing it for a prospect you haven't onboarded yet.

What to do with what you find

• If a competitor is consistently named and your client isn't, that competitor likely has stronger, more specific web content the model is drawing on — reviews, directory listings, a well-structured site. Worth an audit.
• If nobody in the category gets named reliably, that's not a dead end — it's an opening. Being first to build genuine AI-visible presence in an uncontested category is a real pitch.
• If your client is named but described inaccurately (wrong services, wrong area), that's a content-and-citation problem worth fixing regardless of ranking.

Why this doesn't replace ongoing tracking

A manual check like this is a snapshot, and AI answers can shift as models update or as a business's own web presence changes. The value of doing it once is the pitch; the value of doing it monthly is the retainer.