AI Shelf Report

AI visibility audits ChatGPT, Claude, Gemini

AI search is recommending your competitors.

When buyers ask ChatGPT, Claude, or Gemini who to pick, what do they hear about you? We find out.

Get the free 5-question check

We run five buyer questions in your category and send you what the AI says. Free, no call needed. The free 5-question check is the quick version. The AI Shelf Test below is the full one.

Grok (xAI) and Perplexity audits coming soon.

The free check

Not another instant score.

The free graders give you a number in 30 seconds. We give you the actual answers. A human writes five questions your buyers would really ask, runs them through ChatGPT, Claude, and Gemini, and emails you what the AIs said about your business, quoted, with one honest finding.

Why 24 hours and not 30 seconds? Because a human does it. No automated grader can tell you what the AI said about YOUR business in YOUR buyers' words.

  1. You give us your website.

    That is the only homework. Your site tells us your business, your market, and who you sell to.

  2. We write five buyer questions and run them through the three AIs.

    Questions a real buyer in your category would ask, put to ChatGPT, Claude, and Gemini, word for word.

  3. You get a personal email within 24 hours with the answers and our read.

    What the AIs said about you, quoted, plus the one finding that matters most. Not a score. The actual words.

What a free check looks like

Example

Illustrative example. The company below is made up, and so are these answers. They show the format of a real free check, which uses your business, your buyers' questions, and the AIs' actual words.

Q1“Who is the best roofing company in Atlanta for storm damage repair?”

  • ChatGPT Listed three national chains and one local rival. Peachtree Ridge Roofing was not mentioned at all.
  • Claude Recommended two local roofers with strong review profiles. Peachtree Ridge was not one of them.
  • Gemini Named Peachtree Ridge third, with nothing about the company beyond its name.

Q2“My roof is leaking after a storm. Which Atlanta roofer should I call first?”

  • ChatGPT Pointed to a competitor's 24-hour emergency line. Peachtree Ridge never came up.
  • Claude Suggested checking review sites first, then named two companies. Neither was Peachtree Ridge.
  • Gemini Mentioned Peachtree Ridge for emergency repairs, but put a competitor first.

Our readThis company is nearly invisible where it counts. Two of the three AIs skip it entirely, and the third names a rival first. That gap is fixable, and the full audit maps exactly what to publish to close it.

Get your free 5-question check

Tell us where to look and where to send it. A human takes it from there, and your answers land in your inbox within 24 hours.

The method

The AI Shelf Test

Buyers don't Google you anymore. They ask ChatGPT, Claude, and Gemini who to pick. The AI Shelf Test finds out what the AIs say.

  1. We ask like your buyers ask

    40 questions phrased the way real buyers talk, not keyword lists. Conversations.

  2. We ask three AIs, four times each

    One answer is an anecdote. 480 captures is a pattern.

    AI Shelf Report is not affiliated with or endorsed by OpenAI, Anthropic, or Google. Their logos identify the models we test.

  3. We score what matters

    How often you're recommended. Whether your name comes first. Where competitors beat you, and exactly what to publish to change it.

No dashboard you'll never open. A report that tells you what to do, in plain English.

71 of 72
A real finding from a recent audit

In 71 out of 72 urgent-buyer questions across three AIs, not a single brand was named. The buyer was ready to act, and nobody was on the shelf. That's an open shelf.

Sample Audit

What an audit finds: the project management test

To show what this looks like, we ran a pilot of our audit on ClickUp, up against Asana, Monday.com, and Notion. This is a demonstration we ran ourselves, not a client result. The findings below are the real numbers it produced, and they are the kind most brands never see about themselves.

Questions

40 buyer questions

Format

Pilot run, single assistant

Brands tracked

ClickUp, Asana, Monday.com, Notion

Measured

Recommendations, mentions, sentiment

Finding 01

The most recommended brand still has a problem

ClickUp came out on top: it was the recommendation in 10 of the 40 questions, a 25% recommendation rate, more than any other brand in the set.

But ClickUp was also the only brand that picked up negative mentions: 4 of them, and every single one was about complexity. The AI likes the product and warns people about it in the same breath.

Why it mattersWinning the recommendation is only half the story. A warning attached to your name follows buyers into the trial.

Negative mentions, by brand

25% recommendation rate for ClickUp, the best in the set, with a warning attached
ClickUp4
Asana0
Monday.com0
Notion0

Every negative mention in the audit was about one thing: complexity. One clear problem to fix, not a long list of small ones.

Finding 02

Mentioned everywhere, recommended nowhere

Notion showed up in 23 of the 40 answers. It clearly registers: it gets reached for whenever options are being listed.

Then it was recommended zero times. Notion is in the conversation and never the conclusion.

Why it mattersMost visibility tools count mentions and would call this a win. Mentions do not close deals. Recommendations do.

Notion, out of 40 answers

0 times recommended, despite appearing in more than half of all answers
Times mentioned23 of 40
Times recommended0 of 40

Visible in the answers, never the pick. If this is your brand, being famous is not the same as being chosen.

Finding 03

The brand nobody invited owns the word “simple”

Trello was not one of the four brands we set out to track. It showed up anyway, in 11 of the 40 answers, volunteered without being asked about.

Ask for something simple and Trello is the name that comes back. That is why ClickUp's complexity warnings sting: every one of them hands the buyer to Trello.

Why it mattersYour real competitor set is whoever the AI names, not whoever is on your battlecards.

Uninvited appearances

11 answers mentioned Trello, a brand that was not part of the audit
Trello appearances11 of 40

Nobody asked about Trello. It came up on its own, mostly when the question mentioned “simple”.

Finding 04

On price, ClickUp wins outright

The audit included two questions about price, the kind buyers ask when the budget is the deciding factor. ClickUp won both of them.

So the picture is specific, not vague: strong on value, strong on features, held back by one repeated word, complicated.

Why it mattersAn audit that ends with “improve visibility” is useless. An audit that ends with one word to fix is a work plan.

Price questions won

2 of 2 price questions in the audit went to ClickUp, a clean sweep
ClickUp2 of 2
Every other brand0 of 2

Price is a strength worth protecting. Complexity is the tax being charged on it.

About this sample: we ran this pilot on our own to demonstrate the method, on a single assistant. A full audit runs the same kind of 40 buyer questions across ChatGPT, Claude, and Gemini, each question asked 4 times, for 480 captured answers in your category, against the competitors you name. Full sample report available on request, just email us.

Pricing

Fixed prices, fixed timelines

Pick the depth you need. Every tier includes the charts, the findings, and the plan, written so a busy founder can act on them the same day.

Each tier includes everything in the tier before it, plus more. Here is the ladder, feature by feature.

What is included in each one-time tier: Snapshot at $295, Gap Audit at $595, and Audit + Launch at $1,250. Each tier includes everything in the tier before it.
Feature Snapshot$295 Gap Audit$595Most popular Audit + Launch$1,250
Buyer questions tested 20 40 40
AI engines (ChatGPT, Claude, Gemini) Included Included Included
Runs per question 4 4 4
Plain-English executive summary Included Included Included
AI-readiness check (crawler access, schema, llms.txt) Included Included Included
Top-10 gap list Included Included Included
Per-engine competitor gap analysis Not included Included Included
Citation source map (where AI gets its answers) Not included Included Included
Prioritized 30/60/90 action plan (CREATE/FIX/EARN) Not included Included Included
Content brief per gap Not included Included Included
30-min walkthrough call Not included Included Included
Raw data export Not included Included Included
Top 3 briefs drafted into publish-ready outlines Not included Not included Included
Implementation QA before you publish Not included Not included Included

Snapshot

$295

A fast read on where you stand. 20 buyer questions across ChatGPT, Claude, and Gemini, your top 10 gaps, and a plain-English summary of what the AI says about you.

Start with Snapshot
Most popular

Gap Audit

$595

The full audit. All 40 buyer questions, a head-to-head comparison with your competitors, the complete action plan, and a content brief for every gap we find.

Start with Gap Audit

Audit + Launch

$1,250

Everything in the Gap Audit, plus we turn your top 3 content briefs into publish-ready outlines and review the finished pieces before they ship.

Start with Audit + Launch
Ongoing monitoring

GEO Watch

$295/month

Separate from the audit tiers above. GEO Watch keeps watching after your report lands.

  • Weekly prompt tracking
  • Monthly full re-run
  • 2-page monthly brief
  • Drop alerts within 48 hours
  • 2 new content briefs per month
  • Quarterly strategy call
  • Billed monthly, cancel anytime
Start GEO Watch

Questions

Asked often

Which AI assistants do you check?

ChatGPT, Claude, and Gemini on every audit. Those are the three most buyers use today. Grok (xAI) and Perplexity audits are coming soon, and GEO Watch customers get them added as soon as they launch.

How long does it take?

Seven days from the kickoff call to the report in your inbox. GEO Watch runs continuously after that, with a monthly brief.

What do I get?

A written report in plain English with charts: your recommendation rate, how you stack up against your competitors, who gets named first, what the AI says about each of you, and a prioritized action plan with a content brief for every gap. You also get your raw data and a 30-minute walkthrough call. Readable in about 20 minutes.

What if the AI already recommends us?

Then the audit shows you where that lead is thin. Most brands that get recommended still have gaps: question types where they disappear, a rival named first, or one repeated criticism, like the complexity warnings in the sample audit above. You also get a baseline, so if your visibility slips later, you will know it from numbers instead of lost deals.

Do you need access to anything?

No logins, no analytics, no code. Everything we measure comes from asking the assistants public questions, the same way your buyers do. All we need from you is a 30-minute kickoff call: your category, who you sell to, and the competitors you want tracked.

Start here

Find out what the AI says about you.

Email us and we will run a free 5-question check in your category: five real buyer questions, three AI assistants, and the honest answer to who gets recommended. If it is not you, the report shows what to publish to change it.

Get the free 5-question check