Know when AI stops citing you before revenue feels it.

What gets checked, and how often

Two clocks run against one domain. Every 24 hours the infrastructure check fetches llms.txt, robots.txt, the sitemap, the canonical tag, the structured data, and the page itself, and records whether an AI crawler could still get in. A robots.txt edit that quietly blocks GPTBot, a deploy that drops the schema, or an outage that returns 503 to ClaudeBot shows up on the next check, not in next quarter's traffic.

Every 31 days the citation scan asks OpenAI, Claude, Perplexity, and Gemini each of your buyer questions and records a verdict per engine per question. The result is a delta against last month: which questions you gained, which you lost, and who the engines recommended in your place. The methodology is versioned, so a change in how we score never silently rewrites your history.

Questions worth tracking

Onboarding suggests these ten shapes. Keep the ones a real buyer would ask and add two that name a real competitor.

1. What are the best providers for [your category]?
2. Which [your category] platform is best for a small team?
3. What should I look for when choosing [your category]?
4. Compare the leading [your category] providers.
5. What is the best alternative to [known competitor]?
6. How much does [your category] cost?
7. Which [your category] provider is most trustworthy?
8. How do I solve [customer problem] with [your category]?
9. Which [your category] tools work for [customer segment]?
10. Who are the experts in [your category]?

What you receive

A dashboard at this URL once you log in, showing the latest infrastructure status with its age against the promised cadence, the latest citation baseline, and the last delivered report. A monthly report with the full verdict grid, the cited paths, and the competitor list. If a collector stops, the dashboard says stale instead of staying green.

Monitoring measures. It does not fix. When the report shows a gap you want closed, the Citation Gap Diagnostic is the next rung, and the free citability scan stays available for spot checks.

Before you subscribe

What is a buyer question?

A question a prospect would type into ChatGPT or Perplexity when shopping for what you sell. You pick up to ten. The list on this page is the starting set; replace the brackets with your category and your named competitor.

What counts as cited?

Cited means the engine linked to or named your domain in its answer. Mentioned means it named your brand without sending the buyer to you. Absent means neither. Every cell in the monthly report carries one of those three verdicts, and the report lists the paths that were cited and the competitors that were recommended instead.

What if an engine is down or refuses the query?

The report is delayed until the manifest is complete. A failed call is never recorded as absent, because a scoring gap would look like a citation loss you did not have.

Do I need Google Search Console?

No. GSC is optional enrichment for the infrastructure side. The citation measurement runs against the engines directly.

How do I cancel?

In the Stripe customer portal, any time. There is no free trial.