Study of 87 AI visibility tools finds no proof their tracked prompts reflect real user queries
dejanseo · x · 2026-10-28
AIMCLEAR published an investigation of 87 AI visibility (AEO/GEO) tools and found none offer published proof that their tracked prompts accurately represent the vertical conversations customers actually have with AI, or at what scale.
Key points
- AI companies don't release prompt data, so vendors estimate: most synthesize prompts to some degree, some melding traditional search data
- Some buy clickstream and panel records of real AI conversations (claimed PII-free), but filtering by industry, buyer, location, and intent can shrink billions of records to a handful — or zero
- Single synthesized prompts miss how LLM users research via multi-turn, account-wide context; wrong prompt choices skew visibility scores
- The article maps which tools sell what and where their prompts come from
A useful reality-check for anyone doing AEO/GEO measurement.
More from Venture
- Polymarket opens AI bubble burst market, traders price it at just 8% — Polymarket · 2026-10-28
- Agents hate latency: Claude Opus drops APIs past 7.2s, and 'agent marketing' is born — danshipper · 2026-10-28
- Polsia claims $25M run-rate but where are the customer case studies? — petergyang · 2026-10-28
- Silicon Valley influencer lists 39 Agent-era startup ideas: stop building software — FinanceYF5 · 2026-10-28
- OpenAI DevDay's 20+ Launches Point to Subscriptions Becoming a Universal Credit — FinanceYF5 · 2026-10-28
- OpenAI president: AI security should dwarf software — three camps are already cashing in — FinanceYF5 · 2026-10-28