The score is the input. Citations are the proof.
OpenAEO measures results the only way that matters in AI search: by testing whether assistants actually name your business. On the day we deliver a scorecard, we also capture a citation baseline — your real target questions, asked across ChatGPT, Claude, Gemini and Perplexity, recorded verbatim with the date. Then we re-test weekly and track AI-referral traffic from chatgpt.com, claude.ai and perplexity.ai, so the story is a before-and-after you can audit rather than a number you have to trust. We would rather show you nine of twelve questions citing you today, up from zero, than a flattering score with nothing behind it. This page explains exactly how that proof is built, and why we do not publish invented ratings.
The same-day citation baseline
Every engagement starts with a locked baseline. Before a single fix ships, OpenAEO asks your four to twelve target questions across each assistant and records the exact answers, including whether your site is named, linked, or absent. That snapshot is hashed and stored so it cannot be quietly rewritten later; if a client ever disputes the story, the original baseline is the evidence. Because it is captured on day zero, every improvement afterward has an honest starting line — you can see the week your first citation appeared, and the flat weeks in between where nothing moved, without anyone pretending the curve was smooth.
Weekly testing and referral tracking
Scores drift, models update, and competitors ship their own pages, so a one-time audit goes stale fast. On the Business plan, OpenAEO re-runs your target questions weekly across all four assistants, repeating each question several times so a single lucky answer never reads as a trend. Alongside the answers, we segment AI-referral traffic in your analytics — visits arriving from chatgpt.com, claude.ai, perplexity.ai and Copilot — and watch your bot logs, which move first because crawlers visit before citations follow. Together these turn "are we being recommended?" from a guess into a weekly number with a direction.
Why there are no star ratings here yet
You will not find an aggregate rating or a wall of testimonials on this page, and that is deliberate. OpenAEO launched recently, and we will only publish reviews that come from verified customers, attributed by name and date, once they choose to share them. Inventing a "4.8 from hundreds of businesses" would be exactly the kind of unverifiable claim our own audit flags and penalizes — schema that does not mirror a true, checkable fact is a negative signal to assistants and a broken promise to you. When numbers appear here, they will be real, sourced, and something an AI could confirm.
Updated