Free AI Visibility Score: How to Check If ChatGPT, Claude, Perplexity, and Gemini Cite Your Business
A 15-minute diagnostic. 5 queries, 4 platforms, scored 0–15. Run it on your own business right now — no download, no email gate.
ChatGPT says it doesn’t know your business. Your business says it has a website, a newsletter, a customer list.
Both are true. AI engines don’t see the same internet your customers see.
This is the diagnostic that tells you which side of the gap you’re on. 15 minutes. No download. No email gate. Run it right now in this tab.
Why This Exists
Your traffic numbers are dropping. You don’t know why. You write more content. You optimize for search. The numbers keep dropping.
This is an entity recognition failure. AI engines don’t recommend what they can’t verify, and the signals they use to verify a business are different from the ones Google ranked you on. Most online businesses haven’t built those signals yet.
You can’t fix what you haven’t measured. So this scores it.
What This Diagnostic Did to MoonInMental
MoonInMental is the clinical aromatherapy publication I’ve been writing for two years. In February 2026, I ran this exact 5-query diagnostic on it. I scored close to zero across every platform — invisible bracket, the same place most businesses start.
The diagnostic showed me precisely which infrastructure layers were missing. I built them. Six weeks later:
Claude now lists claude.ai as a referral traffic source in MoonInMental’s analytics. Someone asked Claude for a recommendation, Claude named MoonInMental, they clicked through.
Perplexity scored MoonInMental 8 out of 9 on entity recognition. Returned the Substack as a primary source with accurate methodology and 8 citations.
ChatGPT now describes MoonInMental accurately when asked directly.
Earned media: Featured in Homes & Gardens as Founder of MoonInMental. Guest-published on AI Maker, Wyndo’s Substack documenting the 0-to-AI-recommended journey.
Six weeks. Invisible to AI-recommended. The infrastructure layers identified by this diagnostic are what closed the gap.
Now you run it on your business.
Before You Start: Incognito Mode Is Non-Negotiable
Open ChatGPT, Perplexity, Claude, and Gemini in incognito or private browser windows. Do not log in.
AI platforms personalize responses based on your account history, past queries, location, and saved preferences. If you’re logged in, the AI is showing you a version of itself that already knows you — which means your score reflects your account, not your business’s actual visibility.
A logged-in test is not a real test.
The 5 Queries
Score yourself 0–3 on each. Add the totals. The bracket at the end tells you what your score means.
Query 1 — The Direct Search
Platform: ChatGPT Run this: “What is [your business name]?”
Score what comes back:
0 — AI says it doesn’t know, or returns wrong information
1 — AI knows the name but doesn’t describe what you do
2 — AI knows what you do but is missing key details
3 — AI describes you accurately and completely
Your score: ___ / 3
Query 2 — The Recommendation
Platform: Perplexity Run this: “Recommend some good [type of business or content you make] online” — without naming yourself.
Example: “Recommend some good newsletters about clinical aromatherapy.”
Score what comes back:
0 — Not mentioned anywhere in the response
1 — Mentioned in passing, not as a recommendation
2 — Listed in recommendations but not described
3 — Recommended with a description of what you do
Your score: ___ / 3
Query 3 — The Topic You’re Known For
Platform: Claude or ChatGPT Run this: “Best resources for [the one specific thing you want to be known for]”
Example: “Best resources for trauma-informed aromatherapy.”
Score what comes back:
0 — Not cited anywhere
1 — Related topics surface, but your work isn’t there
2 — Your work is mentioned without a link or citation
3 — Cited as a top source for that topic
Your score: ___ / 3
Query 4 — The Explainer Test
Platform: Gemini or Google AI Overviews Run this: “What’s the difference between [your topic] and [a similar topic people confuse it with]?”
Example: “What’s the difference between clinical aromatherapy and essential oil marketing?”
Score what comes back:
0 — Topic is answered without you appearing
1 — Topic is answered, you’re barely connected to it
2 — You’re in the source list but you’re not the main one explaining it
3 — You’re the main source cited for the explanation
Your score: ___ / 3
Query 5 — The Cross-Platform Test
Platform: Run in BOTH ChatGPT AND Perplexity Run this: “Best [type of business or content you make] in 2026”
Example: “Best aromatherapy newsletters in 2026.”
Only 11% of cited domains appear across multiple AI platforms. This query tests whether you’ve earned cross-platform authority — the rarest GEO signal.
Score what comes back:
0 — Neither platform mentions you
1 — One platform mentions you
2 — Both platforms mention you, no citation
3 — Both platforms cite you as a main source
Your score: ___ / 3
Add It Up
Query 1 — Direct Search: ___ / 3
Query 2 — Recommendation: ___ / 3
Query 3 — Topic You’re Known For: ___ / 3
Query 4 — Explainer Test: ___ / 3
Query 5 — Cross-Platform: ___ / 3
TOTAL: ___ / 15
What Your Score Means
0–5 · INVISIBLE
AI engines don’t recognize you as an entity. The structured signals that confirm you’re a real, citable source — schema markup, consistent bio data across platforms, directory presence, content formatting that AI can parse — don’t exist yet for your business. This is the most common starting score and the most fixable.
6–10 · MENTIONED
AI knows you exist. Your name surfaces. But you’re not the cited authority — you’re background. The next layer of GEO infrastructure converts mention into citation: earned media signals, content freshness, cross-platform consistency, the specific structured data that tells AI “this entity is the source on this topic.”
11–15 · CITED
You’re a recognized entity for at least some of what you’re known for. AI engines treat you as a primary source on certain queries. You likely have gaps on specific platforms or query types — Claude weights user-generated content 2–4x more than other models, and your score may shift depending on which platform you tested.
What this score doesn’t measure: traffic, conversions, or ranking position. AI doesn’t give identical answers to identical queries — citation visibility is non-deterministic. This score measures the underlying entity recognition signals, not any single response.
What You Do Next
If you scored under 11, you have an entity recognition problem that won’t fix itself with more content.
The walkthrough that fixes it — The AI Visibility Walkthrough, Online Business Edition — is in build right now and launches the second week of July. Same infrastructure I built and tested on MoonInMental. Schema markup with screenshots. Search Console + Bing setup. Directory strategy. Content formatting protocol. The 30-day retest you run with this same scorecard.
If you want to be on the early notification list, you already are — you’re reading this. The launch goes out to TVP subscribers first.
Either way: run the diagnostic again in 30 days after you’ve changed anything based on what it revealed. The retest is the proof.
Want the PDF Version?
Same content, formatted for offline use. Download it here