AI Visibility
Know exactly when AI recommends you — and when it doesn't
Buyers ask assistants for recommendations every day, and most brands have no idea what gets answered. The AI Visibility board tracks your real buyer prompts across engines: mentioned or absent, in what sentiment, against which competitor.
Product frame — real features, illustrative values
What it measures
Per-prompt presence
Every tracked buyer prompt shows its visibility percentage — how often answers mention your brand — with the engines that produced each run, not a blended average that hides where you're losing.
Sentiment, not just mentions
A mention that says "avoid it" is not visibility. Each stored answer carries sentiment, so the board separates being recommended from merely being named.
The leaderboard you're actually on
Most-visible brand per prompt and a workspace-wide leaderboard built from every recorded citation — you see who owns each answer today, by name.
The Answer Gap
The gap score rolls up the prompts where you're absent — split into competitor-owned versus unclaimed — and feeds "what to plant next" straight into the content planner. Measurement that ends in a to-do list, not a chart.
How it works
- 1Track the prompts that matterAdd the questions your buyers actually ask — unlimited prompts, no per-prompt pricing tier. Seeded from your funnel, editable any time.
- 2Answers get recordedEngine runs land as stored answers — the raw text, the engine, the timestamp — so every visibility number traces to a real answer you can read.
- 3Citations get extractedEach answer is decomposed into citations: brand mentioned or not, competitor named, domain cited, sentiment, category. Eight categories of evidence, one row each.
- 4The rollup tells you where to actVisibility percentages, engine splits, most-visible brands, top cited sources, and the Answer Gap — computed live from the stored rows, linked back to the planner.
The honest part: Honest gate: visibility numbers appear as engine runs accumulate — the board computes from stored answers, never estimates. A brand-new workspace sees its prompt list immediately and real percentages as runs land. No demo data, ever.
Asked plainly
How do I track my brand's visibility in ChatGPT?
Define the buyer prompts that matter ("best pet trust service", "how do I get a compliant lease"), record what engines answer, extract mentions and sentiment, and watch the trend per prompt. GrowthBerry's AI Visibility board does this end-to-end, and every number links to the stored answer it came from.
What is AI share of voice?
The portion of AI answers to a prompt set that mention your brand versus competitors. We compute it from recorded citations — brand mentioned, competitor named, domain cited — so the number is auditable, not modeled.
Which engines does it cover?
Runs are recorded per engine — ChatGPT, Claude, Perplexity, Gemini and peers — and the board filters by engine, because visibility in one assistant tells you nothing reliable about another.
What do I do about prompts where I'm absent?
That's the Answer Gap: absent prompts are split into competitor-owned versus unclaimed, ranked, and linked to "what to plant next" in the planner — the measurement loop closes into content production, which is the whole point.
Related: Citation tracking · the GEO guides · pricing
Start with the free audit
17 checks in about 60 seconds — see how AI engines read your site today, then track what they answer tomorrow.