AI Brand Visibility Monitor — ChatGPT, Perplexity, Gemini & Google AI Overviews

Find out whether AI assistants mention your brand — and who they cite instead.

When buyers ask ChatGPT, Perplexity, Gemini, Claude, or Google's AI Overviews "what's the best X", someone gets recommended. This actor tells you how often it's you, how often it's your competitors, and which websites the AI engines cite — so you know exactly where to earn coverage. This is GEO/AEO (Generative Engine Optimization) monitoring without a $300–500/month SaaS subscription.

Quick Start

1. Enter your brand (plus aliases like your domain) and your competitors.
2. Add the prompts your buyers actually ask ("best CRM for small business"), or list topics and let the actor generate ~10 buyer-intent prompts per topic.
3. Pick your engines and hit Start. No API keys needed — the actor brings its own.

Run it on a weekly schedule to track your visibility over time — enable trendMode to get deltas against the previous run.

What you get

One dataset record per prompt × engine × sample:

{
  "prompt": "best note taking app for students",
  "engine": "perplexity",
  "sampleIndex": 1,
  "brandMentioned": true,
  "mentionPosition": 2,
  "recommended": true,
  "sentiment": "positive",
  "competitorsMentioned": ["Obsidian", "Evernote"],
  "citations": [{ "url": "https://zapier.com/blog/best-note-taking-apps/", "domain": "zapier.com", "title": "The best note-taking apps" }],
  "responseText": "...",
  "model": "sonar",
  "runAt": "2026-07-12T15:00:00Z"
}

Plus an aggregated scorecard in the run's OUTPUT record (free — you only pay for result items): a blended 0–100 visibility score, a visibilityByEngine breakdown (your score on each AI engine at a glance), plus per-engine average mention position, recommendation rate, share of voice vs competitors, and top 10 cited domains — with trend deltas when trendMode is on. Failed queries are listed there too and are never billed.

{
  "brand": "Notion",
  "visibilityScore": 87.5,
  "visibilityByEngine": { "perplexity": 100.0, "chatgpt": 50.0, "gemini": 100.0, "claude": 100.0 },
  "shareOfVoice": { "Notion": 38.0, "Obsidian": 41.0, "Evernote": 21.0 }
}

Why results vary — and why that's handled

LLM answers are non-deterministic: the same question can name different products run to run. That's why this actor supports statistical sampling — set samplesPerPrompt to 3–5 and you get mention rates, not one-shot anecdotes.

Pricing math

You pay per analyzed response (engine-query event). A full weekly audit of 25 prompts × 4 engines × 3 samples = 300 queries ≈ $6–12/run. The prefilled demo input costs well under $0.50.

Use cases

• Agencies — run white-label AI visibility audits per client brand; schedule weekly and export the scorecard into your reports.
• In-house SEO / content teams — the top-cited-domains list is your GEO backlog: those are the sites AI engines trust for your category.
• Founders — check whether AI assistants know your product exists in your category, and track the needle as you publish.

FAQ

Why do results vary between runs? LLMs are non-deterministic and their web results change. Use samplesPerPrompt: 3 or more for stable rates.

What is GEO/AEO? Generative Engine Optimization / Answer Engine Optimization — the practice of getting your brand mentioned and cited by AI answer engines, the way SEO targets classic search results.

Which models are queried? Perplexity sonar, OpenAI GPT-5 mini with web search, Gemini 2.5 Flash with Google Search grounding, Claude Haiku 4.5 with web search, and the live Google AI Overview block (via residential proxies). Not every Google query triggers an AI Overview — aiOverviewPresent: false records tell you that too.

Do I need my own API keys? No. The actor uses its own keys; the cost is built into the per-query price.