ChatGPT vs Gemini: which brands each one names

Each question is asked as written, from a logged-out United States session on ChatGPT, Gemini and Google AI Mode.

How is a score calculated?

On a category board, the score is the share of every valid run across that category's questions that names the brand, over a rolling four weeks, with a 95% Wilson interval. Rows are ordered by the low end of that interval, so a brand named on one question cannot outrank one named across the category. A brand named in fewer than 10 runs, or on only one question, sits below the line as low confidence. On a question page the share is of that question's runs. Where a name sits in the answer is not a rank.

Which brands count where?

Names are matched to a reviewed brand list, and a brand counts only on questions in its own category. Names that are ordinary words, such as Streak, Loops and Instantly, count only when the answer clearly means the product. A CRM named as an integration in a meeting-notes answer is not a point on the meeting-notes board. A brand named in the question itself, as in “alternatives to Profound”, is not scored on that question, so those runs count neither for nor against it. A name that is not on the list stays off the board until a person adds it.

Will this match a signed-in chat?

A logged-out session has no memory and no personalization, so a signed-in person can see a different answer.

Cited pages are the sources that engine marked as citations, not every link in the raw response.

Run by Agnostify. Agnostify builds custom AI systems for teams, and measures how AI recommends brands. This board is a measurement, not a promise that a brand will be recommended.