Skip to main content
Recommended rate is the share of AI answers that recommend your brand over the alternatives. Visibility tells you whether you’re in the answer. Recommended rate tells you whether the answer tells the buyer to pick you.

​
What counts as a recommendation

An answer recommends a brand when it clearly prefers it over the others. For example, it calls the brand the best option, or tells the reader to go with it. These don’t count:
  • Being mentioned. Showing up in a list is visibility, not a recommendation.
  • Being listed first. Coming first in a list is Position, not a recommendation.
  • Being described as big or popular. “The market leader” or “the largest player” describes a brand. It doesn’t recommend it.
Some answers don’t pick anything. They say “it depends on your needs” or ask a follow-up question. Those answers recommend no brand. When an answer recommends one of your products, it counts as a recommendation for your brand.

​
How it’s calculated

Every non-branded answer counts toward the total, including answers to prompts that only ask for a list. Each prompt is scored on its own, and the overall rate is the average across prompts. Every prompt counts equally, however many AI models answered it.
Branded prompts are left out. A prompt that names your brand tends to get an answer about your brand, which would inflate the rate.
Your competitors get a Recommended rate the same way, so you can compare yourself against them.

​
Where you’ll see it

  • Overview: a Recommended rate column in the competitors table, and a switch between visibility and Recommended rate on the development chart and the Topics Leaderboard.
  • Leaderboard: a switch on the Models, Topics, Audiences and Markets tabs.
  • Prompts: a Recommended rate column for each prompt and topic.
  • Prompt details and answers: which brand each answer recommended. A dash means it recommended none.
  • Products: how often AI recommends each product.
Hover the i beside any Recommended rate heading for a short explanation.
Recommended rate is counted from 22 September 2026. Earlier dates show no data.

​
How to read it

A low rate overall isn’t always bad news. Many prompts ask for a list, not a choice, and many answers decline to pick. Compare your rate against your competitors’, not against 100%.

​
How to improve it

  • Make your case explicit. State who you’re the best choice for, and why, on pages AI can read.
  • Win comparison and “best of” sources. AI draws its verdicts from them.
  • Back claims with proof. Pricing, outcomes and reviews give AI a reason to pick you.
  • Visibility: are you mentioned at all
  • Position: where you appear inside an answer
  • Rank: your overall standing against competitors