Last updated July 2026
Why this metric has three names
The concept arrived before the vocabulary settled. Three terms now compete for the same idea:
- Share of model: emphasises the large language model as the new channel
- Share of voice in AI (or AI share of voice): maps directly onto the familiar marketing concept
- Share of answer: emphasises that the output is a generated answer, not a ranked list
All three measure the same thing. Most AI visibility platforms have settled on share of model or AI share of voice. This post uses share of model as the primary term.
The confusion matters because a search for “what is AI share of voice” and a search for “what is share of model” return overlapping but inconsistent results. Treating all three as synonyms is correct.
The one fact that changes how you think about this metric
That single finding reframes everything. If you rank on page one for your category keywords, you still have roughly an 88% chance of not appearing in the AI answer for those same queries. The reverse is also true: pages that earn high AI citation rates often do not rank prominently in traditional search.
This is why share of model and traditional share of voice are separate metrics that require separate measurement programmes.
AI SoV vs. SEO SoV: 4-row comparison
| Dimension | SEO share of voice | Share of model (AI SoV) |
|---|---|---|
| What it measures | Your % of total organic clicks or ranking positions across a keyword set | % of relevant AI prompts in which your brand is mentioned vs. competitors |
| Output format | A ranked list of 10 blue links | A prose answer, sometimes with citations |
| Signal used | Keyword rankings in a structured index | Training data, retrieval sources, and prompt context |
| Overlap with the other | Very low: ~12% of AI-cited URLs also rank in Google”s top 10 (Ahrefs, August 2025) | Very low: strong organic rankings do not reliably predict AI mention rates |
The practical implication: you need to run two separate measurement programmes. A tool that only tracks keyword rankings tells you nothing reliable about your share of model.
How share of model is calculated
The calculation is straightforward once you have the data:
- Define a prompt set. Pick the questions your target buyers actually type into AI engines, clustered by intent (evaluation, comparison, objection handling, pricing).
- Run each prompt across your target engines. At minimum: ChatGPT, Perplexity, Google AI Overviews, Gemini. Add Microsoft Copilot and others for broader coverage.
- Record which brands each response names. Count mentions, not just citations.
- Divide your brand”s mention count by total responses. That is your share of model for that prompt set.
For example: if your brand appears in 35 of 100 responses run against your category, your share of model is 35%. Run the same set next month to track direction of change.
The number is only meaningful relative to your competitive set and relative to itself over time. A 35% share of model in a category with two competitors is very different from 35% in a category with 20.
Tools that track share of model
Several platforms now automate this measurement across multiple engines. The right choice depends on how many engines you need to cover, your budget, and whether you want execution workflows or monitoring alone.
Temso (from $89/mo) is the easy, all-in-one AI SEO platform that tracks share of model alongside brand mentions, citation rate, and sentiment across eight AI engines: ChatGPT, Perplexity, Gemini, Google AI Overviews, Google AI Mode, Grok, Microsoft Copilot, and Meta AI. It also converts share-of-model gaps into a prioritised fix queue inside the same subscription. That makes it the shortest path from zero data to a concrete improvement plan.
Profound goes deeper on citation attribution and prompt volume data. Its Growth tier (from $399/mo) covers 9+ engines and includes visual citation maps that show which content is driving mentions. It is built for teams with AEO-dedicated headcount.
Otterly.AI covers six platforms at a $29/mo Lite entry and includes a structured GEO audit alongside prompt-level citation tracking. It is a strong starting point for teams on a tight budget.
Peec AI (from €85/mo) covers nine or more engines and offers unlimited seats, which keeps costs predictable for agencies running share-of-model tracking as a client deliverable.
Semrush also offers AI tracking tools for teams already in the Semrush ecosystem. It covers a narrower engine set than the dedicated platforms above but benefits from integration with the broader Semrush keyword and backlink suite.
The full comparison with scoring criteria is at /rankings/ai-visibility-tools. Scoring methodology is at /methodology.
What share of model does not tell you
Share of model answers: “Are we in the conversation?” It does not tell you:
- Why your brand is mentioned (or not)
- Whether the mention is positive, neutral, or negative (that is sentiment)
- Which content is earning the citations (that is citation rate)
- Whether the facts the AI states about your brand are accurate (that is accuracy monitoring)
A complete AI visibility programme tracks all four. Share of model is the headline metric, the one you report to a CMO. The other three tell you what to fix. See the /glossary for full definitions of citation rate, sentiment, and accuracy monitoring.
Start measuring it
The only way to know your share of model is to run the prompts. Gut feel is unreliable: brands consistently overestimate their AI visibility when they have not measured it.
Temso lets you run a full share-of-model baseline across eight engines from $89/mo, with no setup beyond defining your prompt set and competitors. If you want to explore other options first, the ranked list at /rankings/ai-visibility-tools covers every platform with verified pricing and engine coverage.