AI Brand Visibility Monitor
Find out what AI assistants actually say about your brand
Run structured mystery-shopping checks across ChatGPT, Claude, Gemini and others: ask a fixed question set about your brand, record mentions, citations, recommendations and sentiment with quoted evidence, then re-run monthly to track how your AI visibility changes. Honest by design — no fake APIs, no prompt tricks, no cherry-picking.
Install it in 3 steps
No sign-up. No CLI tools. Pick your AI, copy one line, paste it over.
Prefer doing it yourself? Buy the skill ($4.99), download the ZIP from your checkout confirmation, unzip it into the directory shown above, restart your AI session — same result. Already bought it? Your AI assistant can download it for you.
What's inside
One SKILL.md (the playbook) + one starter template. A preview:
AI Brand Visibility Monitor
The facts this skill runs on (verified 2026-11-03)
- There is no API for "what does ChatGPT say about my brand." AI assistants answer from training data plus retrieval, and the same question asked twice can get different answers. The only honest way to know what they say is to ask — the way a mystery shopper tests a store. This skill is a disciplined mystery-shopping routine, not a dashboard with a live feed.
- Mentions come in four flavors, and they mean different things. A *named mention* ("Acme makes widgets") is basic awareness. A *citation* (link or named source) means the assistant trusts you enough to source you. A *recommendation* ("best widget for small teams: Acme") is the money answer. A *factual claim* ("Acme was founded in 2019") can be right or wrong — wrong ones need fixing at the source. Track each flavor separately or the numbers lie.
- Sentiment here is a human judgment, not a score. Each mention gets one of three labels — positive, neutral, negative — plus the exact quoted snippet that justifies it. Any tool that gives you "sentiment: 0.73" without showing the quote is asking you to trust a black box. This skill shows the quote.
- Variance is the signal, not noise. Ask the same question three times in fresh sessions. If the assistant recommends you twice and a competitor once, that's your real position: "usually recommended, sometimes not." A single answer proves nothing.
- AI answers change when the web changes. New reviews, new competitor pages, new forum threads — all of it feeds the next answer. That's why this is a *monitor* (re-run monthly) and not a one-time report.
When to use
- You want to know whether AI assistants mention your brand at all when people ask about your niche.
- A prospect said "ChatGPT recommended you" (or recommended a competitor) and you want the full picture.
- You're investing in content/SEO/GEO and need a baseline to measure against next quarter.
- You suspect an AI assistant is stating wrong facts about your company (founding year, pricing, features) and want documented evidence before you fix the sources.
Safety rules (non-negotiable)
Template preview
AI Brand Visibility Monitor — Job Report
One report per monthly run. Fill every field — "n/a" is an answer, a blank is not.
Brand sheet
- Brand / products:
- Canonical facts (founding, HQ, pricing, differentiators):
- Assistants checked:
- Run date:
Question set
Mention table
Scored table
Monthly diff (vs. previous run)
- New mentions:
- Lost mentions:
- Sentiment shifts:
- Facts fixed / still wrong:
- Net assessment:
FAQ
Does this skill connect to ChatGPT or Claude via API to pull brand mentions?
No — and any skill claiming that would be lying. There is no public API for 'what does an AI assistant say about my brand.' This skill structures honest manual checks: you (or your assistant) ask a fixed question set in fresh sessions and log the answers with quotes. It's mystery shopping, not a dashboard.
How is this different from Google Alerts or brand monitoring tools?
Those track the web. This tracks what AI assistants *say* — which is synthesized from training data plus retrieval and varies by phrasing. A brand can rank #1 on Google and still be invisible (or misdescribed) in AI answers. This skill measures the AI layer directly.
How often should I re-run the checks?
Monthly. AI answers shift as the web changes — new reviews, competitor pages, and forum threads all feed future answers. One run gives you a baseline; the month-over-month diff is where the value is.
What do I do when an assistant states wrong facts about my company?
The skill flags each wrong fact with a correction and traces the likely source (outdated profile, competitor comparison page, old press). Fix it at the source — your site, your profiles, the third-party pages assistants cite — then watch the next monthly run.
