Concepts
Recommendation
A recommendation is an AI answer going beyond naming your brand to endorsing it as the right fit for the person asking. Being present and being chosen are different outcomes, and the gap between them is where most brands lose: mentioned in the list, but never the one the assistant suggests.
Why it is separate from presence
A brand can appear in almost every answer in its category and still never be the one recommended. Treating the two as one number hides that, and it is the more expensive of the two problems to have, because it means buyers are seeing you and choosing somebody else.
What it takes to measure
Recommendation is a harder measurement than presence. It depends on how the question was asked, what constraints the buyer gave, and how far the conversation had got — so a recommendation reading needs more samples, and more carefully designed prompts, before it will support a trend. Treat a single endorsing answer as an anecdote.
Present in every answer, chosen in none
A brand appears in almost every answer about its category, and its visibility figure looks strong. Read the answers and it is always third in a list, described accurately but without enthusiasm, while a competitor is the one the assistant suggests when asked what to pick. Presence is not the problem. Being the recommendation is.
Related terms
- Mention — The weaker outcome, and the one most tools stop at.
- AI visibility — A brand can score well here and never be the one recommended.
- Within measurement noise — Why a single endorsing answer is an anecdote, not a trend.
Last reviewed . Definitions are reviewed quarterly, and whenever the underlying measurement changes.
Don't let AI decide your brand's future without you.
See exactly where you stand vs. competitors—and what to do about it.