// BLOG · 08/11/2026

Share of Model Response (SMR): measuring how often AI names your brand

Share of Model Response is the proportion of AI answers to questions in your category that mention your brand. It plays the role for AI answers that share of voice plays for media, and it is the core number in AEON GPTO.

What goes into it

• A set of prompts. Real questions a buyer in your category would ask, such as "best X for Y" or "alternatives to Z". The set is fixed so results can be compared over time.
• A set of engines. Each engine is measured separately, because each has its own data and retrieval.
• Repeated runs. AI answers vary from one run to the next, so each prompt is asked more than once and the results are combined.
• Brand detection. Each answer is checked for your brand and your competitors, including common variants of the names.

For one engine, SMR is the number of answers that name the brand divided by the number of valid answers. Requests that fail or return no answer are left out rather than counted as a miss.

How to read it

• Per engine first. A high average can hide an engine where you are absent. The per-engine view shows where to act.
• Against competitors. SMR is most useful as a comparison: who is named instead of you, and for which prompts.
• As a trend with a margin. Because answers vary, small movements can be noise. A confidence interval around each figure shows whether a change is real.

What moves it

In practice the levers are the ones described in What is Answer Engine Optimisation?: clear and consistent entity facts, structured data, content that answers the category's questions directly, presence in the sources engines retrieve from, and a site that AI crawlers are allowed to read. SMR is how you find out which of these is working.

See your own figure

The free AI visibility scan runs a short SMR check for your domain. Paid plans measure on a weekly or daily cadence with competitor tracking; see pricing.