Visibility baseline
A structured test of how ChatGPT, Gemini, Perplexity, Copilot and Google AI Overviews currently answer the questions your buyers ask — including who they name instead of you.
ai/assistant
People increasingly ask an AI rather than scanning ten blue links. The answer names two or three suppliers. Everyone else is not ranked lower — they are simply absent.
An assistant answering a live question fetches pages, extracts what it can parse, and prefers claims it can corroborate somewhere else. That produces a fairly consistent set of advantages:
scope
A structured test of how ChatGPT, Gemini, Perplexity, Copilot and Google AI Overviews currently answer the questions your buyers ask — including who they name instead of you.
Auditing robots rules, firewall and CDN blocks so that retrieval crawlers can reach your pages, with deliberate decisions about which AI crawlers to allow.
Schema markup for the organisation, its services, products, FAQs and articles, plus consistent details wherever your business is listed elsewhere.
Rewriting key pages so answers come first: clear definitions, direct question-and-answer sections, specific figures and comparisons a model can lift cleanly.
Identifying the sources assistants keep quoting in your market, and working to get you accurately represented on them.
Repeat testing on a schedule, plus analytics configured to separate AI referral traffic so the commercial effect is visible rather than assumed.
honest framing
There is no submission form for AI assistants and no ranking dial to turn. Outputs vary between users and change without notice as models are updated. What can be done is to remove every reason a system would skip you, make your facts easy to lift and verify, and measure the result over time instead of relying on a single flattering screenshot. Anyone quoting guaranteed placement in AI answers is describing something that does not exist.
questions
It is the practice of making a website easy for AI assistants to crawl, interpret, trust and cite when answering a question. It also goes by answer engine optimisation (AEO) and generative engine optimisation (GEO). It shares foundations with SEO, but weights clear structure, unambiguous facts and credible third-party sources more heavily.
Two ways. Ask them systematically and repeatedly, using the questions your customers would use, and record what comes back over time. Then look at referral traffic from AI assistants in your analytics, which can be isolated once tracking is configured to identify it.
It depends on whether your content is the product. Publishers may want to restrict training crawlers while still allowing the retrieval crawlers that fetch pages to answer live questions — the two are separate and can be controlled separately. For most businesses selling goods or services, blocking is a way to become invisible.
It rests on the same foundations, and a site with poor technical SEO will not do well here either. The differences are real, though: no ten-result page to climb, far more weight on corroboration from other sources, and a much harder measurement problem.
related
The crawlability and content quality that AI visibility is built on.
seo →Isolating AI referrals so this work can be judged on enquiries, not anecdote.
analytics →Where organic and AI visibility are thin, paid search still buys the buying moment.
ppc →