AI Visibility: the Claude skill that tests whether AI tools recommend you
AI Visibility is a free Claude skill that tests whether ChatGPT, Gemini, Perplexity and Claude recommend your business by name. It generates the eight queries your customers actually ask, scores the real responses, and tells you exactly who AI recommends instead. One curl command installs it.
Install it
One command, then tell Claude who you are and where you operate.
curl -o ~/.claude/skills/ai-visibility-check.md https://raw.githubusercontent.com/lhitches/claude-seo-skills/main/skills/ai-visibility-check.md
Then start a conversation like: "Run the AI visibility check for [your business], [your industry], [your city]." No install? Paste the skill file contents into any Claude chat and the same process runs.
What it does, and what you get
Rankings tell you about Google. This tells you what happens when a customer asks an AI who to hire.
- Eight buyer-shaped queries generated from your actual business and location, phrased the way customers ask, never the way SEOs search.
- A real-response score. Claude gets tested live in the conversation. You run the same queries in ChatGPT, Gemini and Perplexity and paste them back. The skill refuses to simulate answers it did not see.
- Who AI recommends instead: a named list of the competitors taking your spot, with appearance counts.
- What AI is actually reading: whether answers come from training data, directories, review sites or live citations of your pages.
- Five actions for the next 30 days, ranked, plus a written baseline so the re-test in 30 days proves whether anything moved.
Real sample output
I ran this skill on myself. It scored me 0 from 8 and told me my own agency is more visible than I am. Publishing that is the point: the skill reports what it finds, even when the person it embarrasses wrote it.
AI VISIBILITY RESULTS: Lawrence Hitches
Visibility score: 0/8 responses mentioned the name. Rating: INVISIBLE (preliminary, Claude tested live, 24 responses across ChatGPT, Gemini and Perplexity still to run).
- Who AI recommends instead: StudioHawk (his own agency, 3 appearances, never linked to him by name), Online Marketing Gurus (2), Prosperity Media (1), and generic no-name advice (4 responses).
- What AI is actually reading: general training-data knowledge of the Australian SEO agency market. No directories, no review sites, no citations of lawrencehitches.com.
- Why the invisibility: the agency brand is in the AI tools' vocabulary, the personal brand is not. Third-party mentions have not yet made the name the answer.
- Action 2 of 5: bind the name to the agency's visibility. Claude named StudioHawk three times but never its Chief of Staff. A byline and bio on the agency site, linked both ways, closes that gap.
Excerpt from a live run, 9 July 2026, seeded from a real fetch of lawrencehitches.com/answer-engine-optimization/. The skill scored only the 8 genuine Claude responses it saw and marked the other 24 as pending, exactly as designed.
When to use it, and when to use a sibling skill
Use AI Visibility for the blunt first question: does AI recommend us at all?
| You want | Use | Why |
|---|---|---|
| To know whether AI recommends your business today | AI Visibility (this skill) | Real queries, real responses, a scored baseline and 5 actions |
| Ongoing tracking of AI citations for target queries | AI Citation Monitor | Built for repeat monitoring rather than a one-off score |
| Your share of AI answers from Bing citation data | Citation Share Auditor | Reads your actual Bing Citation Share export |
| To fix why AI does not recognise your brand as an entity | Entity SEO Auditor | Audits the entity layer: schema, sameAs, entity home, consistency |
FAQ
Which AI tools does it test?
Four: ChatGPT, Gemini, Perplexity and Claude. Claude gets tested directly in the conversation. For the other three you run the generated queries yourself and paste the responses back, because the skill never simulates another tool's answers. What it scores is real output, never a guess.
What does the visibility score mean?
It counts how many of the 32 possible responses (8 queries by 4 tools) mention your business by name, then rates you from Invisible to Dominant. Partial runs get a preliminary score. The written report becomes your baseline, so a re-test in 30 days shows movement.
My company shows up in AI answers but I do not. Is that normal?
Very. In the sample run on this page, Claude named StudioHawk three times but never named its own Chief of Staff. AI tools learn brands from training data and third-party mentions, so a personal brand needs bylines, speaker pages and press that bind the person to the company by name.
How often should I re-run the test?
Every 30 days, with the same eight queries. AI visibility shifts slowly, and model updates arrive without notice. The skill's report is written to be a baseline: keep each one, compare the per-query breakdown, and you will know whether the actions actually moved anything.
Want the theory behind it? The Hawk Academy deep dive covers how StudioHawk runs AI visibility work for clients.
Created by Lawrence Hitches
Chief of Staff at StudioHawk. I built this library from the processes our 120-person team runs across 500+ client campaigns. Every skill is free and MIT licensed. Yes, this one told me I am invisible. I am working on it.