/ SEOJune 20267 min read
Will ChatGPT Recommend Your Business? GEO, Explained in Plain English
Somewhere today, a potential customer typed a question into ChatGPT that used to go into Google: 'best aesthetic clinic near Manchester', 'accountant for e-commerce businesses', 'who should build my company website'. They got back three names and a paragraph of reasoning. Either you were in that answer, or you didn't exist.
This is the shift the industry has started calling GEO — Generative Engine Optimisation — and it's moving from conference-talk novelty to measurable revenue channel. Research on B2B buying already shows purchasers using AI assistants to build vendor shortlists before a single sales conversation happens. By the time anyone contacts you, the AI has already decided who's in the running.
The strangest part? The answers aren't even consistent. Studies comparing Google's AI Overviews with ChatGPT found they disagree on brand recommendations roughly two times out of three. Visibility in one engine tells you nothing about the other — which means this is a discipline, not a lottery.
/ How AI engines actually pick answers
When an AI assistant recommends businesses, it's drawing on what it has read: your website, directories, reviews, press coverage, comparison articles, forum threads. It favours sources that state facts plainly — who you serve, where you operate, what you're best at — and brands that appear consistently across independent sources.
This is why GEO overlaps with traditional SEO but isn't identical to it. Google rewards pages engineered to rank; AI engines reward entities they can understand and cite. A site full of vague slogans ('we deliver synergistic growth solutions') gives a language model nothing to work with. A site that says 'we build patient-enquiry systems for private clinics in the UK' can be quoted, and quoted brands get recommended.
/ The practical GEO checklist
Say what you do in plain sentences, on pages an AI can read. Every service and industry you serve should have its own page stating clearly who it's for and what outcome it delivers. Answer real questions on your site — the FAQ sections buyers actually ask, phrased the way they ask them — because question-and-answer content maps directly onto how assistants compose replies.
Get mentioned where AI reads. Reviews, industry directories, local press, expert commentary — third-party corroboration is what turns 'a claim on your website' into 'a fact about your business'. And keep your basics ruthlessly consistent: name, location, services, positioning, everywhere the same.
Then measure it. Ask the major assistants the questions your customers would ask, monthly, and track whether you appear. It's crude, but so was checking your Google position in 2005 — and the businesses that did it early ate their markets.
/ Why early movers win disproportionately
AI answers have a compounding quality that search rankings never had. An assistant recommends the businesses that are well-documented; being recommended generates more coverage, citations and reviews; that documentation feeds the next answer. The gap between cited and invisible widens on its own.
Meanwhile, ads are already appearing inside ChatGPT — meaning organic AI visibility is about to become the thing money can't easily buy, exactly as organic search once was. The window where plain-spoken, well-structured content is enough to win these answers is open now. It will not stay open.