The first audit platform scoring your site on CRO + GEO + ASO.
Know if humans, AI search & AI agents can find and convert on your site.
Free audit in 60s ๐
We asked 4 AI assistants for a cosmetic surgery clinic in Toronto.
3 named 14 clinics.
1 refused:
"I don't have access to current, real-time information about specific cosmetic surgery clinics in Toronto."
It told the patient to check credentials instead.
@AgenticOperator Yep. And the bit that convinced me it's structural rather than just LLM randomness: Ask the same assistant twice and 51% of its sources come back. Ask two different ones and 7% do. Each one is ~7.5x more consistent with itself than with its rivals.
10 questions patients actually ask. 4 AI assistants.
214 different websites read.
75% of them read by exactly one assistant.
The other three never saw the page.
What AI actually reads when someone asks about a cosmetic clinic:
~13% directories (Practo, JustDial, RealSelf, WhatClinic, Reddit)
~87% clinics' own websites
The thing you control completely is the thing being read most.
This week we published what 4 AI assistants actually say when asked to recommend a lawyer. ยท 7% overlap between any two assistants ยท 1 of 10 questions had a shared source.https://t.co/4nY8d15GCo
#airesearch#orignaldata
Out of 10 legal questions, only ONE produced an agreement. "Best tax law firm in Singapore?" โ 3 of 4 assistants named Allen & Gledhill. "Best personal injury firm in NYC?" โ 34 different firms, no agreement. Difference isn't quality. https://t.co/4nY8d16erW
Ask ChatGPT the same question twice and only ~32% of its sources repeat. Perplexity: 72% Claude: 54% Gemini: 42% ChatGPT: 32% That screenshot of AI recommending you? Might not survive being taken again. We re-ran 10 questions 3ร on each assistant: https://t.co/4nY8d15GCo
When AI recommends a lawyer, it's mostly reading directories. Sources cited most across 10 legal questions: https://t.co/qtBc2UMAL8 -10 https://t.co/luWCC9B6hq -9 https://t.co/AT9mWoN6HG -8 https://t.co/vlR963dj3Z -6 Your own site barely gets a look in.https://t.co/4nY8d16erW
"Best personal injury law firm in NYC?" ChatGPT-Kramer Dillof, Rosenbaum & Rosenbaum; Claude -Gair Gair Conason, Edelman & Edelman; Gemini -Block O'Toole & Murphy, The Perecman Firm; Perplexity-Kramer Dillof, Richard M. Kenny 34 firms named. https://t.co/4nY8d15GCo
We asked 4 AI assistants: "Best personal injury law firm in NYC?" They named 34 different firms. Two assistants asked the same question and overlapped on ~7% of sources. It's four. 10 questions, full data: https://t.co/4nY8d16erW
#AISearch#LegalTech#SEO
Added custom prompts: write the exact question your buyers ask, we run it across ChatGPT, Perplexity, Gemini, and Claude, and show which one names you. Generic questions prove the concept. Your questions are the ones that actually keep you up at night.
#startup
Why AI recommends a competitor instead of you, in three lines that come up constantly: they're cited on 12 review sites, you're on 1. They have a product schema, you don't. They're in 3 directories you've never heard of. It's a checklist, not a moat. Which is good news.
#startup
We started showing people which source types AI cites in their category. Most common reaction: "I've been optimizing the wrong page." Turns out the answer is usually built from review sites and directories you've never claimed, not the homepage you've rewritten four times.
#start
"How do I rank in AI?" is the wrong question. AI doesn't rank you, it cites sources. And in most categories, those sources are review sites, directories, and industry blogs, not your homepage. You can't optimize your way into an answer built from places you're not on.
#startup