Who do the robots recommend?

Every week we ask ChatGPT, Claude, Gemini and Perplexity the questions buyers actually ask, and keep score of who gets named.

Top of the charts

How the score works

For each category we keep about 20 questions a buyer might ask, from “what are the best tools” to specific use cases. Every week we send each one to four AI assistants and note which products they name and in what order. A week's score is how often a product is named, weighted by position: first place counts in full, fifth place about 60%, tenth or lower 40%. The number you see is the average of our last four runs, because a single run of answers is noisy, and we only call a product climbing or sliding once it moves by 5 points or more. No sponsorships, no reviews, no votes, just what the assistants actually say.

The full method, including what these numbers cannot tell you →

We ask ChatGPT through its consumer app and the other three through their APIs. Answers can differ from what you see logged in, since the apps add memory, personalization and their own instructions. Questions about alternatives to a specific product show up on product pages but don't count toward scores.