GEO glossary

AI share of voice

AI share of voice is the proportion of AI-generated answers in a category that mention a brand, measured against competitors across a fixed set of prompts. It shows how often ChatGPT, Gemini, Perplexity or Google's AI features recommend a brand compared with its rivals.

Reviewed by Gareth Hoyle, Managing Director
Last reviewed 31 October 2026

Two versions are common. Mention rate is the percentage of responses that name the brand at least once. Share of voice divides the brand's mentions by all brand mentions in the same responses, so the figures for every brand in the set add up to 100%.

Some trackers weight mentions by position in the answer or by whether the brand is also linked as a source. Whichever method is used, it has to stay the same between reporting periods, with the same prompts, platforms and location settings.

AI answers vary from one run to the next. Research by Rand Fishkin of SparkToro and Patrick O'Donnell of Gumshoe.ai, reported in January 2026, ran 12 recommendation prompts 60 to 100 times each per platform on ChatGPT, Claude and Google's AI. ChatGPT and Google's AI returned the same list of brands less than 1% of the time, and the same list in the same order less than 0.1% of the time.

How often a brand appeared was steadier. Across varied, human-written prompts about headphones, Bose, Sony, Sennheiser and Apple appeared in 55% to 77% of responses. The study was not peer reviewed, but it explains why share across many runs is a more reliable measure than a single position.

How to improve it

Share of voice rises when a brand appears in more of the sources AI systems draw on: comparison articles, review sites, trade press and forums, as well as its own pages. In our experience, digital PR and reviews move it more than on-site changes alone.

Example

Take 50 prompts about UK running shoes, each run five times. A retailer named once in 90 of the 250 answers has a 36% mention rate. If those 250 answers contain 600 brand mentions in total, the retailer's 90 give it a 15% share of voice.