How AI assistants recommend local businesses: our data
We publish what we measure when we ask ChatGPT, Gemini and Perplexity for local recommendations: real answers, counted, with the method and its limits on every page. Free to quote with a link.
Studies
204 answers analysed
Do ChatGPT, Gemini and Perplexity recommend the same businesses? The AI Agreement Index (2026)
Asked the same question (“best {trade} in {city}”) for 68 trade and city pairs on 2 October 2026, ChatGPT, Gemini and Perplexity named 769 distinct businesses; 81.4% were named by only one of the three assistants, 3.9% by all three, and in 67.6% of the questions no business at all was named by all three.
→136 answers analysed
The words AI uses when it recommends a local business (2026 study)
In 136 real answers from ChatGPT, Gemini and Perplexity about 20 local businesses in 18 cities (27 September – 2 October 2026), the reasons most often given for recommending a business were quality (63% of answers), reviews and ratings (58%), being a specialist (54%) and years in business (51%); price came up in only 17%.
→30 answers analysed
Which AI visibility tools do ChatGPT, Gemini and Perplexity recommend? (October 2026)
Asked five buyer questions about AI visibility tools on 15 October 2026 (30 answers in total), ChatGPT, Gemini and Perplexity named 53 different tools. Only 3 appeared in at least a third of the answers: Semrush (47%), Ahrefs (33%), Peec AI (33%).
→
More data
AI Recommendation Barometer
The same question asked every month to ChatGPT, Gemini and Perplexity for 68 trade × city pairs: how often they name the same businesses, month by month.
→Where AI gets its recommendations
755 sources cited by Gemini across 92 “best {trade} in {city}” questions (September 2026).
→AI rankings of local businesses
Which businesses Gemini names, city by city, from real answers.
→