See if AI recommends your business — and why it does or doesn't
Found by AI is a free tool that checks whether ChatGPT, Claude, Perplexity and Google's AI recommend your business, and scores how ready your website is for AI to read, trust and cite it. Here's exactly how it works — a 90-second demo, then step by step.
Run a free scan →Watch the demo (2½ minutes)
Captions are on-screen — sound optional.
How the free scan works
- Paste your website. ARGOS reads it the way the AI crawlers do — content, schema, robots.txt, llms.txt — and works out what you do and who your customers are. If your site can't be read, the scan says so instead of scoring a guess.
- NAVIGATOR writes your buyer questions, SOUNDER asks the engines. The free scan sounds ChatGPT and Google's Gemini, each question asked more than once. A deep measure sounds all seven: Claude, Perplexity, Grok and DeepSeek join, and Google's AI Overviews gets its own column.
- ASSESSOR scores 30+ signals into two numbers: AI Visibility (are you named right now?) and AI Readiness (how ready are you to be found?), each out of 100 — and LOG keeps what each engine actually said, word for word.
- PILOT hands you the fixes. The six signals AI checks before recommending you, your tech stack and agent-readiness, and the highest-impact things to fix first. Serve specific areas? Add your own locations for a local read.
The pipeline, end to end
One scan, five stages, every stage measured, nothing estimated.
What's in the complete report
- The whole board — 12 buyer questions across every engine, up to 228 sampled answers: a stable map, not a two-question corner.
- A suburb-by-suburb visibility heat map — where AI names you across the areas you serve.
- The competitors AI recommends instead of you — and exactly why they win.
- A citation game plan — the specific directories, registers and communities to get listed and mentioned on.
- Copy-paste fixes — your Organization / LocalBusiness schema and llms.txt, generated and ready to publish.
- Real Core Web Vitals and agent-readiness with a target range to aim for.
What is GEO, and why does AI visibility matter?
GEO — generative engine optimisation — is SEO for the AI era. Traditional SEO gets you ranked in a list of blue links; GEO gets you named inside the AI's answer. AI weighs different things: whether you're a recognised entity, your third-party mentions and reviews, and citations — far more than on-page keywords. Around Most of the brand mentions in AI answers sit on third-party pages rather than your own site. Found by AI measures all of it, so you know exactly where you stand and what to fix. Read the full framework →
Why the numbers hold still — when AI itself doesn't
Ask an AI engine the same question twice and you can get two different answers — that's how these systems work. A tool that asks once and prints a score is measuring noise. We built for stability instead:
- SOUNDER asks every question more than once. The free scan samples each question twice per engine (8 live answers); the full report asks up to three times per engine and takes the majority vote — up to 228 live answers per measure.
- Monitoring re-measures the same pinned questions week after week, so your trend line is real movement — not freshly re-generated questions pretending to be a trend.
- BOSUN halts degraded scans instead of scoring them. If we couldn't read your site or reach an engine, the report says so instead of printing a confident wrong number.
- LOG freezes reports as measured — they never silently re-score. Methodology upgrades are labelled, so a scoring change never masquerades as your movement.
See where you stand — free
Run your site through the scanner — about 60 seconds, no signup. Then drop your email and reply with what you found. I read every one.
Run a free scan →Video transcript
Your customers used to Google you. Now they ask AI — and the demo shows exactly what that looks like from both sides.
You drop a website in and hit Scan. The engine reads the site, works out what the business does and who buys from it, then asks all seven answer engines — ChatGPT, Claude, Gemini, Perplexity, Grok, DeepSeek and Google's AI Overviews — the questions real buyers type. Every answer is kept in full, so you can read exactly who each engine named and why.
Then the demo walks the results: your AI Visibility and AI Readiness scores, the rivals AI recommends instead of you, the per-suburb map for local businesses, and the prioritised fix plan — including the llms.txt and schema generated for your site.
The last part shows monitoring: the same measurement re-run every week, mention alerts, the live console, and the one-tap handoff that loads your numbers straight into ChatGPT or Claude so you can ask what to do next.
Common questions
How do I check if AI recommends my business?
Paste your website into the free scanner. It asks the engines the questions your customers ask — ChatGPT and Google's Gemini on the free scan, all seven on a deep measure — and tells you whether you're named, or which competitor is named instead.
How does the AI visibility scan work?
It reads your website, works out who your customers are, then asks the engines your buyers' questions, each sampled more than once. It scores you across 30+ signals into two numbers: AI Visibility (are you named now) and AI Readiness (is your site built to be found).
Which AI engines does it check?
The free scan sounds ChatGPT and Google's Gemini. Deep measures cover seven: ChatGPT, Claude, Perplexity, Gemini, Grok, DeepSeek — and Google's AI Overviews gets its own column.
Is it free?
Yes — the scan is free, no signup. The US$29 Snapshot is the one-off entry; the complete human-reviewed report is included in month one of Stay Found monitoring.
What's in the complete human-reviewed report?
Twelve buyer questions across every engine, a suburb-by-suburb visibility heat map, the competitors AI recommends instead of you and why they win, a citation game plan, copy-paste fixes including your schema and llms.txt, and your AI-traffic beacon. See a sample →
Why do AI visibility scores change between scans?
AI engines are probabilistic — the same question can get different answers on different runs. That's why we sample every question multiple times (twice per engine on the free scan, three times with a majority vote in the full report) and why monitoring re-measures the same pinned questions each week. Genuine movement shows as a trend; single-run wobble doesn't.
Does it work outside Australia?
Yes — the scan localises to your market automatically. It reads your site, works out where you operate, and asks the questions your local buyers ask: a Texas plumber gets Texas questions and Texan rivals, a Cardiff dog-walker gets Cardiff's. The public benchmark index is Australian because that's where we started measuring; more countries are being measured now.