AI Visibility
Measure how visible a brand actually is inside AI answers. Configure a client, run real prompts against the engines you choose, and see where the brand appears, where competitors do, and which sources the engines cite.
System preflight admin only · live canaries spend a few paise of API credit
Smoke-tests the tools end to end: one minimal call per engine, mentions-index balance, database integrity, disk space and pricing. Run it before demos or onboarding a customer.
Your clients
| Client | Brand | Domain |
|---|
No clients yet — add the first one below.
Add a client
Share of voice — latest completed run, measured brand mentions
Run trend — deltas compare each run to the previous completed run
| Run | Score | Mention rate | Citation rate | Competitor-only | Avg brand rank | Valid rows | Errors |
|---|
Average brand rank shows where the brand appears relative to tracked competitors in answers that mention it (1 = named first). Runs measured before this metric existed show "—".
Edit client details
Changes apply to FUTURE runs — past runs keep the detection they ran with.
Prompts
Tick the prompts to include in the next run — exactly 10, 20, or 30. Unticked prompts stay active for future runs; nothing needs to be retired.
| Run | Prompt | Topic |
|---|
One question, a few phrasings. They use the same run slots as ordinary prompts and report as a range, so you can see how much the wording changes the answer. Phrasings already in this client's library are skipped, not duplicated.
One prompt per line — copying a column in Excel and pasting it here works as-is. Duplicates are skipped automatically.
Suggested prompts — untick or edit any, then add
Prompt categories
A category is a name + an explicit head keyword (2–5 words, how a buyer would search it) + an optional audience. Each category can generate a 30-prompt slate built from indexed AI questions, measured category demand, and curated patterns — every prompt labeled with how it was derived. The top 2 categories also feed the AI Mentions add-on's keywords.
| Category | Head keyword | Audience |
|---|
No categories yet — add the first one below.
Add a category
One head keyword per category — to cover several keywords (e.g. AI visibility AND GEO optimization), add a separate category for each. The audience is free text; commas are fine there.
Generated slate
Evidence-based Prompt Foundations
Real Google search demand (with volumes) + measured buyer-intent funnel stages + a four-engine consensus signal. Every suggestion carries its evidence badge. Requires the client's industry/category to be set.
Evidence-based suggestions — grouped by funnel stage; untick or edit, then add
▸ Archived prompts
Restoring a prompt returns it to the active (priced) set. Re-adding an existing prompt is blocked — restore it here instead.
| Prompt | Topic |
|---|
Competitors
Named competitors are detected in every answer — by name or any brand name/alias you list — so you can see who wins the prompts the brand loses. Up to 5 per client.
| Name | Domain |
|---|
Run center
What do these engines and the add-on measure?
Google AI Overviews — The AI answer box at the top of Google Search results. We query Google through an indexing provider and record whether an Overview appeared, what it said, and which pages it cited. Some queries show no Overview at all — we report that honestly rather than counting it as a miss.
Google AI Mode — Google's conversational AI search (a separate, chat-style experience from Overviews). Same measurement: the answer text and its cited sources, recorded as returned.
AI Mentions & Trends (add-on) — A different lens: instead of asking the engines your prompts, this reads a third-party index of thousands of real AI answers (Google AI Overviews + ChatGPT, US) to measure how often your domain and tracked competitors are mentioned across the whole category, which sources are cited alongside those mentions, and the category's AI-native demand trend. Counts are of indexed AI answers — never verified facts about any company. Requires at least one engine selected: it's an add-on to a run, not a run by itself.
Credits are charged when the run starts. If any answers can't be retrieved, the credits for those answers are automatically returned to your balance.
Runs
| # | Started | Setup | Status | Credits |
|---|
No runs yet.
Companies AI engines mention
Reports
Generated on demand from this run's stored evidence. The CSV opens in Excel; the two documents are client-ready. Newest reports are listed first.
| Report |
|---|
Answers
| Prompt | Engine | Pass | Status | Brand mentioned | Site cited | Competitors seen | Score |
|---|