This is the question that decides whether your product exists in the next purchase: does an AI engine name you when someone asks about your category?
More buying decisions start inside an answer engine every quarter. "Best CRM for agencies," "top running shoes under $150," "which analytics tool handles LLM tracking" — the answer isn't a list of blue links anymore. It's a synthesized recommendation that names specific products and says why they're worth it. If your product isn't named, it doesn't exist in that moment of intent. That's generative engine optimization, and it runs parallel to classic SEO whether you measure it or not.
AI mention tracking is how you stop guessing about it.
The four signals, per prompt per engine
A real tracker doesn't test one prompt once. It tests a batch of prompts across multiple engines on a recurring schedule and records four things about every mention:
- Mention rate. What share of your tested prompts name your product at all. This is the closest thing to an "AI share of voice" — the headline number.
- Position. When you're named, are you first, mid-list, or a footnote? Presence isn't the whole story; status inside the recommendation is.
- Sentiment. Are you argued for ("top choice for…"), neutral ("one option is…"), or against ("lacks…")? A mention isn't automatically a win.
- Citation. Is your own site or profiles linked as a source? Cited brands reinforce entity and authority the way backlinks do in classic SEO — see entity SEO for AI search.
Our AI Search Tracker seeds 25 buyer prompts per product across awareness, consideration, and decision, then tests them weekly against ChatGPT, Perplexity, Gemini, and Copilot. The prompt set is the same shape on every engine so "mention rate on ChatGPT" versus "mention rate on Perplexity" is apples-to-apples — a real distribution difference, not a metric artifact.
Roll it into one number
The four signals combine into a 0-100 visibility score: mention rate weighted by position and tempered by sentiment. It's the framework explained in AI visibility score explained, and its value is that it degrades exactly when the AI stops recommending you — and improves when your fixes land.
A score isn't a vanity metric if it's actionable. Weak prompts map to specific fixes:
- Not mentioned, no citation → add an llms.txt entry + FAQPage schema for the exact category.
- Mentioned but negative sentiment → strengthen trust signals: backing, comparisons, testimonials.
- Mentioned third but competitor is first → publish a direct comparison page and earn a third-party citation.
- Cited with stale details → enforce consistent entity data across site and listings.
Our AI visibility audit guide walks through reading those gap reports in order of impact.
Why weekly beats monthly
AI answers shift as models update and content recirculates. A monthly check tells you what happened. A weekly check tells you what changed before it costs you a quarter of pipeline — and it lines up with the fix loop in the 30-day AI visibility plan.
Weekly tracking also surfaces the honest overlap between AI visibility and Google rankings — covered in AI visibility vs Google rankings. Your page-one keywords don't tell you what ChatGPT names in an answer; the mention tracker does.
Close the loop automatically
Tracking is the measurement half. The fix half is where teams stall, so the same scan infrastructure can feed a robot that closes the loop: snapshots of RankBot watch scan output for weak queries, draft the fix, and open a pull request. You merge, and the citation is earned back automatically.
One scan infrastructure, two views: a product scorecard here, a full marketing-site audit over here. For teams just starting, read what a free AI visibility scan reveals before paying for anything.
Start with the baseline
You can't move a number you haven't measured. Add your products — name and URL — and the tracker runs a baseline scan so you know where you start on day one.
Open the AI Search dashboard, add your first product, and get a baseline mention rate across all four engines this week. It's the fastest way to find out whether ChatGPT is quietly recommending your competitor while you spend the quarter focused on keywords.