--- name: ai-search-visibility description: Check how a business appears when buyers ask AI assistants and search engines for a recommendation (ChatGPT, Gemini, Perplexity, Claude, Copilot, Google AI features and Maps), find why competitors are named instead, and list fixes. Use when someone asks whether AI tools recommend them, how to show up in AI answers, or about answer engine or generative engine optimisation. --- # AI search visibility check Buyers now ask AI assistants questions like "best family lawyer in Dubai" or "which clinic in Riyadh does veneers". The assistants answer from what they can find and trust about each business. This check finds out what they say today, why, and what the business can change. ## Build the question set Write 10 to 20 questions a real buyer would ask, with the person's help: - Category and place: "best [service] in [city or area]". - Problem led: "how do I [solve the problem the business solves] in [country]". - Comparison: "[business] or [competitor]", "alternatives to [competitor]". - Brand: "what is [business] known for", "is [business] good for [need]". - Include Arabic versions if the business serves Arabic speakers. Keep the same questions each time so results can be compared month by month. Save the list in a file the team can reuse. ## Run the check For each question and each assistant available to you or the person: - Record whether the business is named, its position among businesses named, which competitors appear, and what sources are cited. - Note anything wrong in what is said about the business: old address, wrong services, closed branches, incorrect prices. - Answers vary between runs and between users. Run important questions more than once, and treat the results as a sample, not a ranking. If you cannot query the assistants yourself, give the person the question list and a simple table to fill in. Ask them to use a private or signed-out window, so their own history does not shape the answers. ## Find the causes Compare the business with the competitors that are named. Check: - **Website clarity**: is there a page that answers each buyer question plainly, with the service, the place and who it is for? Assistants quote clear pages more readily than vague ones. - **Crawl access**: does robots.txt block AI crawlers the business wants to be read by (for example GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended)? Blocking is a valid choice, but it should be a deliberate one. - **Structured data**: Organization or LocalBusiness, Service, FAQPage and Article markup that matches the visible page. - **llms.txt**: a plain summary of the site's main pages. It is a proposed convention; it helps some tools and costs little. - **Google Business Profile and Maps**: correct categories, hours, photos, services and recent reviews. These feed local answers. - **Third-party mentions**: directories, industry associations, press, review sites and lists that the assistants cite for competitors. Being absent from the sources they cite is the most common reason for not being named. - **Consistency**: name, address and phone the same everywhere. ## Output 1. **Scorecard**: the business was named in X of Y answers, with the position when named, by assistant. Use only the results you recorded. 2. **Who gets named instead** and the sources cited for them. 3. **Errors** in what assistants say about the business, with where the wrong information probably comes from. 4. **Fixes**, ranked: content pages to write or improve, technical changes, profiles to claim or correct, and third-party sources worth being listed in. 5. **Re-check plan**: the same questions, monthly. Never promise a ranking or a guaranteed mention. Assistants change often, and no one controls their answers. Never suggest fake reviews, paid fake articles or spam listings; they damage trust and may break platform rules.