AI visibility (brand lookup)
How brand lookup measures mentions, citations, and share of voice in ChatGPT and Google AI Overviews, including hard limits, coverage, and costs.
Brand lookup is how you measure whether a brand or domain shows up in AI answers on ChatGPT and Google AI Overviews. It is one measurement surface inside the Visibility Engine, not the whole GEO product. The full loop (Do Next, Site Changes, Proof, access, citation losses, referrals) is mapped on GEO (/docs/geo). Marketing overview of the engine: AI visibility (/features/ai-visibility).
What one lookup shows
One lookup is meant to answer "are we even in the room?" without pretending it is every GEO workflow.
- Total mentions, plus a platform split between ChatGPT and Google AI Overviews
- AI search volume overall and per platform
- A 12-month monthly trend
- Share of voice versus up to 5 competitors
Cited pages and prompts have their own sections and hard caps in the table.
Honest share of voice
Share of voice here is competitive context, not a false sense of precision. Sample-size gates and honesty badges flag thin samples. When the badge says the sample is weak, treat SoV as directional.
Which pages are getting cited?
You can see up to 40 cited pages, each with URL, domain, citation count, and example prompts. That is how you move from "we get mentioned" to "which URLs the models actually lean on." Those pages are the natural starting points for Site Changes and citation readiness work later in the engine.
Which prompts do you appear in?
You can see up to 50 prompts, with volume, platform, first and last seen, sources, and co-mentioned brands. That is the "what questions put us in the answer" view. It pairs with prompt explorer for deeper digging (/docs/prompt-explorer) and with Opportunities when you care about gaps (/docs/opportunities).
| What you get | Limit |
|---|---|
| Competitors for SoV | Up to 5 |
| Cited pages | Up to 40 |
| Prompts | Up to 50 |
| Trend history | 12 months |
Beyond brand lookup (engine bridges)
Lookup measures. The rest of the Visibility Engine turns measurement into work:
- Do Next (prioritized queue across rank, gaps, CTR, AI, technical)
- Site Changes (create/improve recommendations with question targets)
- Citation loss diagnosis (where you faded from answers)
- AI access audit (can AI systems reach you?)
- Hallucinated 404s (AI-referred dead URLs)
- Proof / remeasure (did citations or referrals move?)
- Weekly snapshots (ongoing measurement rhythm)
- AI referral traffic in analytics (/docs/analytics)
How the agent uses it
Ask whether you are visible in AI answers for a brand or domain. The agent should call brand lookup (and related tools), show cost on paid runs, and answer only from what those tools return. External MCP clients can call the same tools (/docs/mcp). Judgment and approvals: the agent (/docs/the-agent).
Coverage caveat
ChatGPT mention data is US English only. Do not read a lookup as global ChatGPT coverage. Google AI Overview behavior follows the country and language settings of the query context you are measuring. Coverage limits are product reality, not fine print to skip.
Costs and caching
Brand lookup is paid. You see the cost before it runs. A free re-run is available from cache for 24 hours. Availability and packaging are still being finalized during private testing. Credits overview: credits (/docs/credits).