Press kit and data downloads
What Dopbot measures and does not measure, how to cite it, and the data behind the index.
What Dopbot measures
- How often AI assistants recommend a brand when a shopper asks a buying question.
- Each market is updated once a month: we ask assistants the same shopper questions in every category, with web search on, through their developer APIs, and record which brands each answer recommends.
- Share of answers is the percentage of sampled answers in a category that recommend the brand. We also record the average position in the answer and the web sources the assistants cite.
- Markets: United States, México, Canada. Questions are asked in the market's language, with the search location set to that country.
What it does not measure
- Sales, market share, product quality or customer satisfaction.
- What any one person sees in a consumer app. API answers approximate the apps but are not identical to them, and answers vary from run to run.
- Brands the assistants did not name in the sampled answers.
- A ranking is not an endorsement. Dopbot is independent and is not affiliated with OpenAI, Perplexity or any other AI provider.
The full method, including how brands are extracted and counted, is in the methodology.
The data so far
The first index has not published yet.
These counts appear once it does. 160 categories are set up for sampling.
How to cite
Name the week or month, say what the number is, and link to the page you used. Describe a share as "recommended in X% of sampled AI answers", not as "the best" brand.
Category ranking
Dopbot Index, "[category name]", [week label]. https://dopbot.com/index/[category]
Monthly report
Dopbot monthly report, [month year]. https://dopbot.com/reports/[YYYY-MM]
Data downloads
Monthly CSV files of the public aggregates. They contain brand names, shares and cited domains only. Raw AI answers are not published. Text cells that begin with =, +, - or @ carry a leading apostrophe so spreadsheets do not run them as formulas.
- Brands: week, market, category_id, category, brand, share (0 to 1), avg_position, appearances, samples.
- Sources: week, market, category_id, category, domain, citations (answers that cited the domain).
Downloads appear after the first index publishes.
Contact
Questions, corrections and interview requests: hello@dopbot.com. If a brand name looks wrong, include the category, the week and what you expected to see.