The same unbranded buyer questions go to three AI answer engines twice each month. Every Shortlist Share ships with a 95% interval, and a brand is called the leader only when it ranked first in at least 97.5% of bootstrap draws; otherwise every brand whose rank range includes #1 is shown as tied for first.
How often each tracked brand is recommended, mentioned without being recommended, and cited from its own website in answers to a fixed set of unbranded buyer questions. Every engine is reported separately and combined with equal weight. Every Shortlist Share is published with a 95% interval, and every engine's score with its number of eligible answers.
Personalized results in consumer apps, market share, product quality, sales impact, why an engine chose a brand, or future rankings. The index measures answers returned to API requests, not what one person sees in a chat window.
Mention Rate = Shortlist Share + Recommendation Gap. There is no composite visibility score; each metric is a share of eligible answers you can quote directly.
How often AI puts a brand on the buyer's shortlist.
Definition and formulaHow often AI mentions a brand without recommending it.
Definition and formulaHow often a brand's own website is part of the evidence AI shows.
Definition and formulagpt-5.6-terra
Responses API with the web_search tool.
sonar-pro
Sonar chat completions with native search.
gemini-3.5-flash
generateContent with google_search grounding.
Engine names describe the APIs we query. OpenAI, Google and Perplexity do not endorse, review or sponsor the index.
| Category | Status |
|---|---|
| Project management software | Question set and brand universe ready; measurement not yet scheduled |
| Email marketing software | Question set and brand universe ready; measurement not yet scheduled |
| CRM software | Building the brand universe with a discovery pass |
| Help desk and customer support software | Building the brand universe with a discovery pass |
| HR software (HRIS) | Building the brand universe with a discovery pass |
| Product analytics software | Building the brand universe with a discovery pass |
Each category tracks up to 15 brands. RecoProof does not index its own category, AI visibility tools. How brand universes are built.
Measurement has not started yet, so no edition is scheduled. The first edition's collection and publication dates will be announced here before collection begins. Each edition is collected from the first Monday of the month and published on the Tuesday eight days later. A category that fails a quality gate is not published that month.
How to read this index. Each month we ask three AI answer engines (OpenAI with web search, Perplexity Sonar, and Google Gemini with Search grounding) the same 24 unbranded buyer questions per category, twice each, through their public APIs. Shortlist Share is the percentage of eligible answers that recommend a brand, shown with a 95% interval. A brand is shown as #1 only when it ranked first in at least 97.5% of bootstrap draws, so its whole 95% rank range is #1; otherwise every brand whose rank range includes #1 is shown as tied for first.
What it does not show. Results come from API requests, not from the consumer apps of these engines, so they do not reflect personalized results or any individual's chat history. A fixed set of 24 questions cannot represent every buyer question. Being recommended does not mean a product is better, and the index does not explain why an AI recommends a brand.
Independence. Scores and badges cannot be bought, RecoProof customers are measured exactly like everyone else, and RecoProof does not rank its own category.
The methodology lists every threshold, formula and quality gate. Earlier studies are on the research page.
Analytics preferences
RecoProof uses optional Google Analytics to understand acquisition, product activation, and purchases. Advertising storage and personalization remain disabled. Rejecting analytics does not affect the product.