saymetry

Open data

State of AI Visibility

Aggregate numbers from every real Saymetry scan to date: 6 paid scans across 5 markets, 1,162 live answers from 7 AI engines, each answer stored word for word. No company-level data appears here, ever. Updated 2026-09-02; refreshed monthly as the dataset grows. Quote freely with attribution (CC BY 4.0).

3.2%

median mention rate

How often the median business is named at all in AI answers about its own market.

2.6%

median recommendation rate

How often the median business is actually picked, not just listed.

33.3%

never recommended

Share of scanned businesses that were not recommended in a single answer about their own market.

66.7%

never cited

Share of scanned businesses whose own pages were never used as a source by any engine.

32.7%

top-3 concentration

In the median market, the share of all recommendations taken by just three names.

How these numbers are made

Every scan asks all 7 engines 60 questions written the way buyers ask them, and classifies each answer for who was mentioned, recommended, and cited as a source. Each rate in a scan carries a 95% confidence interval; the aggregates here are medians across scans, so no single market dominates. The full method, including its limits, is at saymetry.com/methodology. The dataset is small and young; it says so rather than pretending otherwise, and the figures firm up as it grows.

Where does your business sit against these numbers?

The $249 scan measures your market the same way: all 7 engines, every answer quoted, an honest plus-or-minus on every number.

Run my scan, $249 →