Does AI Choose You?
Audit how ChatGPT, Claude, Gemini, and Perplexity perceive your brand
What it is.
Does AI Choose You? answers a question traditional SEO tools can't: when someone asks ChatGPT, Claude, Gemini, Perplexity, or DeepSeek about your product or industry, do you show up — and how are you described?
Submit a website URL (and optionally a brand name, competitors, and target market) and the Actor runs 40 buyer-intent questions against 8 AI models across the AIDA framework — Awareness, Interest, Desire, Action — in two passes: once with no brand context (inferred, the model's baseline opinion of the category) and once with your brand named (informed). The gap between the two (the delta) shows exactly how much lift your own content and positioning are giving you.
The report returns 13 sections: an overall score and verdict, per-AIDA-stage scoring, category/competitor detection, a per-model breakdown, detected proof signals, ranked risks, a prioritized action plan, and copy-paste JSON-LD schema markup templates. A liteMode option runs a single model on sampled questions for a fast directional check first.
What it checks and does.
Eight-model coverage
GPT, two Claude models, Gemini, Perplexity, DeepSeek, Kimi, and Gemma audited in the same run for a cross-model visibility score.Inferred vs. informed delta
Every question runs brand-agnostic and brand-informed, isolating exactly how much lift your own content gives versus the category default.7-dimension AIDA scoring
Brand Recall, Sentiment, Specificity, Evidence, Completeness, Category Clarity, and Position/Rank, scored per AIDA stage.Competitor detection & ranking
Auto-detects your top 5 competitors (or takes a manual list) and scores your prompted-awareness rank against them.Action plan + schema markup
Ranked risks, an immediate/short-term/long-term action plan, and copy-paste, priority-graded JSON-LD templates.Lite mode
A fast, single-model, sampled-question check in under a minute, versus the full 8-model/40-question audit.
Where it fits.
- Check whether your product gets recommended when someone asks an AI assistant for options in your category
- Isolate whether your own content actually moves AI perception, via the inferred-vs-informed delta
- Benchmark your AI visibility and prompted-awareness rank against named competitors
- Generate ready-to-ship JSON-LD schema markup prioritized by impact
Call it with this.
| Field | Type | Required | Description |
|---|---|---|---|
url | string | Yes | The website to audit. |
brand | string | No | Brand name, if different from what's inferable from the URL. |
geoMarket | string | No | Target geographic market for the audit questions. |
competitors | string[] | No | Named competitors to compare against in the same audit; auto-detected if omitted. |
liteMode | boolean | No | Run a fast, single-model, sampled-question audit instead of the full 8-model battery. |
Pricing:$15 per audit, pay-per-event — charged once the audit completes and a report is successfully fetched. ~640 LLM calls per run, typically 15–30 minutes.
Frequently asked questions.
What is an AI visibility audit?
It's whether AI models like ChatGPT, Claude, and Perplexity know your brand, describe it accurately, and recommend it when someone asks a buying question in your category — similar in spirit to an SEO audit but for AI answer engines instead of search results. This is often called Generative Engine Optimization, or GEO.
How is this different from tools that just track brand mentions?
Mention-tracking tools tell you that a model said your name. This audit runs the same 40 buyer-intent questions twice — once with no brand context (inferred) and once with it (informed) — so it isolates why you're winning or losing at each AIDA stage, then turns that into a ranked, prioritized action plan instead of a raw mention count.
Which AI models does it actually query?
Eight, in a single run: GPT, two Claude models, Gemini, Perplexity, DeepSeek, Kimi, and Gemma.
How long does a full audit take, and what does it cost?
$15 per audit, pay-per-event — 8 models × 40 questions × two passes (~640 LLM calls), typically 15–30 minutes. liteMode: true runs a single model on sampled questions in under a minute for a fast directional check first.
Do I need to give it my own OpenAI/Anthropic/etc. API keys?
No — the Actor calls all 8 models itself through its own backend. You only need your Apify token to run it.
What do I actually do with the JSON-LD schema markup it gives me?
Copy the templates from the schemaMarkup section straight into your site's <head> (or your CMS's structured-data field) — they're prioritized (p0/p1/...) so you know which to ship first.