Find out if ChatGPT recommends your product — or your competitor
Stop flying blind. See which answers name a competitor instead of you — and exactly what to change on your site.
- HubSpot: Best for integrated marketing and sales.
- Pipedrive: Excellent for visual pipeline management.
- Monday.com: Great for high-customization teams.
Your brand was not mentioned in this response.
One answer. A single check reads twenty of them and counts the ones that name you.
“This isn't an SEO audit. This is how you win in the age of LLM-first discovery.”
The run, exactly as you will see it
This is the screen inside your account, not a picture of it: the same six steps, in the same order, drawn by the same component.
- Queuing the runComplete
- Fetching your siteComplete
- Analyzing your pagesComplete
- Asking ChatGPT, Claude, Gemini, PerplexityComplete
- Checking Wikipedia, Wikidata and directoriesComplete
- Comparing with competitorsComplete
- Writing your fixesComplete
- Building the report and the verdictComplete
report ready — 20 answers, 10 pages, 3 mentions
Report ready
20 answers read, 10 pages checked. The brand was named in 3 answers out of 20.
The demo is sped up. A real run takes a few minutes: every answer is asked of a live model, and every page is fetched one at a time.
What comes back from one run
Five blocks of a real report, filled in for a made-up CRM at acme.com. This is the layout and the level of detail you get — none of it is a preview of something planned.
Which answers named you, and who was named instead
| Buyer question | ChatGPT | Claude | Gemini | Perplexity |
|---|---|---|---|---|
| “What's the best CRM for a small business looking to scale?” | HubSpot, Pipedrive | HubSpot | Monday.com | Pipedrive |
| “CRM with a free tier for a team of five” | HubSpot | You | Zoho | HubSpot |
| “Cheapest CRM that syncs with Shopify” | Pipedrive | Nobody | You | Pipedrive |
| “CRM alternatives to Salesforce for a startup” | Monday.com | Pipedrive | You | Monday.com |
| “Which CRM has the best pipeline view” | Pipedrive | Monday.com | Pipedrive | Nobody |
The denominator stays on the page: 5 buyer questions × 4 models = 20 answers, and 3 of them named the brand. Counted in answers you can go and reproduce, not as a percentage of a sample nobody shows you.
The answer itself, not our summary of it
For a small team that expects to grow, HubSpot is the usual starting point — the free tier is generous and the upgrade path is clear. If all you need is a pipeline you can actually see, Pipedrive costs less per seat and takes an afternoon to set up…
At five people the free tiers are close, so it comes down to what they cap. Acme CRM counts seats rather than records, which is the limit a small sales team hits last…
Every cell keeps the fragment around the name it found. You do not have to take our word for the count — the wording that produced it is right here.
The domains the models cited while answering
14 of 20 answers showed their sources.
- g2.com9 answers · 64%
- reddit.com7 answers · 50%
- capterra.com5 answers · 36%
- hubspot.comcompetitor4 answers · 29%
- acme.comyours1 answers · 7%
The one recommendation that is not about your own site: these are the pages you have to appear on. Shares are counted from the answers that cited anything at all — the rest answered from memory, and no edit to any page changes those.
Five requests to your homepage, seconds apart
- GPTBot403 blocked
- OAI-SearchBot403 blocked
- ClaudeBot200 OK
- PerplexityBot200 OK
- tobecited-audit/1.0neutral control200 OK
The control got through and GPTBot did not — the rule is aimed at AI crawlers, not at automation in general.
Identical requests that differ only in the user-agent they carry. robots.txt states your policy; this records what your server and your CDN actually do about it.
The same page, before and after your scripts run
/pricing
Crawler gets
84 words
Browser draws
1,180 words
Missing from the HTML
Plans and pricing340 words
Start free and upgrade when your pipeline outgrows it. Every plan includes two-way email sync and unlimited contacts…
What's included210 words
Custom fields, a shared inbox your whole team can answer from, and reporting you can hand to an investor…
Price drawn by script: $29 per seat / month
Crawlers do not run JavaScript, so the score is counted on the first number. The second one is the evidence: these are your own paragraphs, and the assistant never sees them.
What one run consists of
- 4
- assistants are asked every question — ChatGPT, Claude, Gemini and Perplexity. Live answers, asked at run time, not a stored index of what they said once. If one of them will not answer during your run, the report says so and leaves it out of the maths.
- 12
- answers in the free check when all four assistants reply: 3 buyer questions across the panel. You write the questions — they are the ones your buyers actually type.
- 20
- answers a week on a paid plan: 5 questions, repeated every week, so one snapshot turns into a line you can act on.
- 10
- pages read per run, without JavaScript — only what your server returned, which is all a crawler ever gets to read.
- 5
- user-agents knock on your homepage: GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and a neutral control that tells a rule against AI crawlers apart from a shield against everything.
- 0
- changes to your site are needed to run a check. No script to embed, no DNS record, no call with anyone.
Every problem comes with the correction written out: which file, which tag, which paragraph. The report opens as a page, exports to PDF, and can be handed on as a read-only link that needs no account. Ready-to-paste fix files are a paid add-on.
What tobecited is not
- It is not a guarantee that ChatGPT will start recommending you. We measure what the models say today and show you exactly what to change.
- We do not speak to the models on your behalf and we do not insert your brand into their answers. No tool can, ours included.
- It is not a Google rank tracker. Search positions and AI answers are different surfaces; we report on the second one.
- It is not a content factory. You get the specific files and paragraphs to correct, and you decide what actually gets published.
Questions people ask before the first check
- What is an AI visibility check?
- An AI visibility check is a measurement of how often AI assistants name your brand when a buyer asks them for a recommendation. The check puts your own buyer questions to ChatGPT, Claude, Gemini and Perplexity, reads every answer, and records three facts: whether you were named, which competitors were named instead, and whether your site was cited as a source. A free check asks 3 questions across all four models — 12 answers — and reads up to 10 pages the way a crawler reads them: no JavaScript, only what your server returned. Separately it knocks on your homepage as GPTBot, OAI-SearchBot, ClaudeBot and PerplexityBot and records who was let in. A paid plan repeats 5 questions every week, so one snapshot becomes a line you can act on. The result is a rate, not a grade: "named in 3 answers out of 20" is a fact you can verify by asking the same question yourself.
- How do you know what ChatGPT says about my site?
- We ask ChatGPT, Claude, Gemini and Perplexity the questions your buyers actually type, then read the answers and record whether you were named, who was named instead, and in what tone. Nothing is simulated: these are real answers from the live models.
- Do I need to install anything on my site?
- No. You enter your URL and the check runs from our side, reading your pages the way a crawler reads them: without JavaScript, only what your server returns. What comes back is a report plus files and text you paste in whenever you are ready.
- How long does the free check take?
- Minutes, not seconds: you create an account, write the three buyer questions we should ask, and we put them to all four models live while we read your pages. The free check answers one question — whether AI names you or a competitor — and names every problem it finds on your pages. It does not write the fixes: the ready-to-paste llms.txt, Schema and rewritten copy are paid, as are the competitor breakdown and the weekly re-checks. One free check per account.
- Why do the results change between checks?
- AI answers are not deterministic: the same question can name different brands on different days. That is why a single check is a snapshot and monitoring is the real signal — you watch the line move over weeks instead of trusting one number.
- Will these fixes help my Google rankings too?
- Often, but that is not what we measure. Clean server-rendered HTML, honest Schema markup and a readable llms.txt help both surfaces. We only report on what the AI assistants say about you.
Claim your spot in the AI answers
Don't let your competitors win the conversation. See where you stand free — the ready-made fixes are paid.
No credit card required.