AI Share of Voice - Brand Visibility in LLM Answers
Pricing
from $4.30 / 1,000 ai answer analyzeds
Measure how often AI models recommend your brand: asks GPT, Gemini, Claude, Perplexity and more your buyers' questions, then reports mention rate, list position, share of voice vs competitors, AI-discovered rivals, cited sources and fixes. No API keys.
What buyers search for, in plain words (e.g. "note-taking apps", "CRM software for startups"). Used to generate realistic buyer prompts (discovery, recommendation, alternatives, comparison, budget, perception).
Competitor names to benchmark against (e.g. "Evernote", "Obsidian"). The report also lists competitors the AI recommends that you didn't name.
Other names the models may use for you (product names, old names, abbreviations).
e.g. "notion.com" - reports whether AI answers cite your site and (with 'Audit my site') tests whether AI bots can read it.
With a domain set, also test your site the way AI bots see it (robots.txt per AI bot, live bot fetch vs browser, JS-only content, llms.txt, Organization schema) and link the findings to your visibility numbers as root causes. Included in the report price.
Use your own buyer questions instead of the generated set (one per line).
Which AI engines to ask. API engines (cheap, billed at model cost): 'memory' answer from training data, 'web' search the live web. 'app' engines capture the REAL consumer products (ChatGPT search, Perplexity, Gemini, Copilot, Google AI Mode / AI Overviews) via Apify's own scrapers, billed to your account per prompt (~$0.20; Google AI Overviews ~$0.003). Mixing API + app engines adds an API-vs-app calibration. The form starts with one engine so a first try costs a fraction of a cent.
[ "openai-gpt", "google-gemini", "anthropic-claude", "perplexity-sonar"]How many buyer prompts to ask each engine.
Ask each prompt more than once per engine to average out answer randomness (costs more).
Country or region to localise the prompts (e.g. "Germany").
Two-letter country code passed to the consumer-app scrapers (e.g. "us", "gb", "de").
Also return the full text of every model answer (bigger output).
Pick (or create) a key-value store to keep a snapshot per brand+category. Each run then reports what changed since the last run: visibility, mention rate, share of voice, per-engine moves, rising/new competitors, new/dropped cited domains - plus alerts. Schedule the Actor weekly for GEO tracking.