Home / Blog / How to check if AI search engines name a brand
AI visibility · 7 November 2026 · Louise Sylvester
How can a brand check its AI search visibility?
To check whether AI search engines name a brand, run a fixed set of real buying prompts across ChatGPT, Gemini, Google AI Overviews, AI Mode, Perplexity and Copilot, then record whether the brand is absent, used as a source, named or recommended.
- Build 30 to 60 prompts from Search Console demand
- Test in clean sessions set to the target market
- Check AI crawler access and Bing indexation
- Fix pages that are used but not named first, then track monthly
Buyers now ask AI tools for shortlists before they visit a single website. A question such as "best route optimisation software for a 20-van fleet" goes to ChatGPT, Gemini or Google's AI Mode, and the answer names three or four brands. Brands that are missing never get the click, and often never know they were in the running.
A common finding in AI visibility audits is that a brand's content is already in the answer. The engine retrieved the page and used its facts, but credited a competitor or named nobody. SALTO Systems is a clear example: its content shaped smart-access answers that recommended SALTO in only 1 of 44 prompts, as the SALTO case study shows. The seven steps below are the method we use to find out where a brand stands.
1. Build a fixed prompt set
Start from real demand rather than guesswork. Export 90 days of queries from Google Search Console and group them into definitions, comparisons, how-to questions, regional variants and brand questions. Then write 30 to 60 prompts the way a buyer would type them into an AI assistant: longer, more conversational and more specific than a Google search.
Weight the set towards buying intent. For a SaaS brand that means category shortlists, comparisons and alternatives. For an ecommerce brand it means "best for" questions, compatibility questions and gift questions. Keep the set fixed once it is agreed, so results from one month can be compared with the next.
2. Test each prompt across the engines
Run every prompt in ChatGPT, Google AI Mode, Google AI Overviews, Gemini, Perplexity and Copilot. Use clean sessions with no personal history, logged out where possible, and set to the target country and language. Record the full answer, the brands named, the order they appear in and every URL cited.
Each platform finds sources differently, so results vary more than most teams expect. ChatGPT and Copilot lean on Bing's index, Perplexity runs its own crawler and Gemini draws on Google's systems. A brand can be well placed in one and missing from another. The platform guides for ChatGPT, Gemini and Perplexity explain the differences in detail.
3. Classify each answer
Give every answer a visibility level:
- Absent: the brand isn't mentioned and none of its pages are cited.
- Source: a brand page is cited or clearly used, but the brand isn't named.
- Named: the brand appears in the answer alongside competitors.
- Recommended: the brand is presented as a top choice for the prompt.
The pattern across the whole set matters more than any single answer. A brand that is often a source but rarely named has a content problem. A brand that is absent everywhere usually has an access or authority problem.
4. Check AI crawler access
Review robots.txt and CDN or firewall bot rules for OAI-SearchBot, PerplexityBot, Googlebot and Bingbot. Security tools often block unfamiliar crawlers by default. Training controls such as GPTBot and Google-Extended are a separate decision: blocking them doesn't remove a site from AI search answers, but blocking a search crawler does. The glossary entry on AI crawlers lists which bot does what.
Then check indexation. Verify the site in Bing Webmaster Tools and confirm the important pages are indexed there, because a page Google ranks well can still be missing from Bing, and so from ChatGPT and Copilot.
5. Fix pages that are used but not named
These are the quickest wins, because the engine already trusts the page. In most cases the passage it lifts doesn't contain the brand name. The engine quotes a definition or a how-to step, then credits whichever brand its other sources mention.
Open each important section with a direct answer of two or three sentences, name the product inside it, and keep naming consistent across the site. Back the page with Organization and Product schema so the connection between the content and the brand is explicit. For AI Overviews specifically, the guide to AI Overview optimisation covers how Google selects passages.
6. Look beyond the website
AI engines lean heavily on review platforms, directories, marketplaces, comparison articles and community threads. For many commercial prompts, those sources are cited more often than any brand website. List the sources cited for each prompt and find where the brand is missing, thin or described inaccurately.
Consistency matters as much as presence. If the founding date, headquarters, product names or pricing differ between the website, LinkedIn, Wikidata and review sites, AI platforms become less confident about the brand. Entity SEO aligns those facts. Paylists used this kind of off-site work to build AI citations across seven European markets, described in the Paylists case study.
7. Track monthly
Re-run the same prompt set each month and record the same fields. Track citation share, mention share and recommendation share, plus AI referral sessions and conversions in GA4. Log every page change and outreach win with a date, so it's possible to see which changes moved which prompts.
Movement is rarely instant. Access and indexation fixes can show within weeks. Content and third-party changes usually take two to three months to shift a set of prompts, which is why a fixed baseline matters so much.
Share the results with sales and product teams as well as marketing. The prompts where competitors are recommended often line up with the objections sales hears on calls.
What the results usually show
Across SaaS and ecommerce audits, the same patterns come up. Brands are often strong on definitions and weak on comparisons, because competitors and review sites own the comparison content. Regional prompts are frequently answered by local directories the brand has never heard of. And product facts in AI answers are often out of date, pulled from old pages or third-party listings.
Each pattern points to a specific fix. For software brands, the AI visibility for SaaS programme focuses on shortlist and comparison prompts. For online stores, AI shopping optimisation focuses on product and collection prompts.
The AI Visibility Audit covers all seven steps, with a prioritised fix list and a 90-day plan.