AI visibility is how often, and how favourably, your business, brand or web pages appear in answers from AI tools such as ChatGPT, Gemini, Perplexity, Microsoft Copilot and Google’s AI features. It covers being named, being recommended and being cited with a link, and is usually reported as a share across a fixed set of questions rather than as a single position.
How AI visibility works
Unlike a ranking, there is no single list to check. AI answers change from one run to the next, and between users, locations and tools. So AI visibility is measured by sampling: you choose a set of prompts your customers might type, run them repeatedly across the tools that matter, and record what happens each time.
For each prompt the useful things to note are whether your brand is mentioned, whether your site is cited as a source, where in the answer you appear, how you are described, and which competitors appear instead. Spread over dozens of prompts and many runs, that gives a citation share, sometimes called AI share of voice: the proportion of answers in which you appear compared with competitors. At the time of writing (October 2026), several SEO platforms and specialist trackers automate this sampling, though each uses its own method, so figures from different tools are not directly comparable.
Two indirect signals help. GA4 records referral visits from sites such as chatgpt.com, perplexity.ai and copilot.microsoft.com, which shows whether AI answers send you traffic. And growth in branded searches can suggest people heard of you somewhere, including in AI answers.
Why it matters
People increasingly ask an AI tool to explain a problem or shortlist suppliers before they search in the usual way, particularly for considered purchases and B2B services. If your business never appears in those answers, you can miss the shortlist without ever knowing it existed.
What shapes AI visibility is mostly familiar. The tools that search the web need to reach your pages, so crawlers such as OAI-SearchBot must not be blocked. They draw on what the web says about you, so reviews, directory listings, press coverage and forum mentions all count. And they favour pages that state facts clearly. Being a well-defined entity, with consistent details everywhere, helps tools connect those mentions to you.
For many UK small businesses, AI referral traffic is still small next to Google. That is my judgement, not a published figure, and it varies by sector. It is a reason to measure before spending heavily, not a reason to ignore it.
Common mistakes
- Treating one screenshot as evidence. A single answer, good or bad, tells you very little because answers vary from run to run.
- Testing only branded prompts. Tools usually know your name when asked directly. The real test is whether you appear when someone asks for a type of business in your area.
- Blocking AI crawlers by accident. A firewall or robots.txt rule set years ago may be keeping search-facing AI crawlers out.
- Buying a score without knowing the method. Visibility scores mean little unless you know which prompts, tools and runs produced them.
- Ignoring what is said. Appearing with wrong opening hours, old prices or a closed branch is visibility that hurts.
How to act on it
Write 20 to 50 prompts that cover the buying journey: problem questions, comparison questions and “who should I use in Manchester for…” questions. Run them across the tools your customers use and record a baseline. Check robots.txt and firewall settings so search-facing AI crawlers can reach your site, and set up a GA4 channel group for AI referrals.
Then work on the causes: clear, specific pages that answer the prompts; consistent business details; reviews and genuine mentions on sites that AI tools read. Re-run the same prompt set each quarter, and use brand mention monitoring to see where you are being discussed. My generative engine optimisation service is built around this cycle.