Share of Model (SoM)
A metric measuring how often a brand is cited or recommended in AI-generated answers across multiple AI platforms, replacing traditional Share of Voice.
Share of Model (SoM) is the percentage of AI-generated answers within a defined query set in which a brand is mentioned, cited, or recommended. It is the AI-era successor to Share of Voice and Share of Search, and it is calculated separately for each model (ChatGPT, Gemini, Claude, Perplexity) because the same brand can dominate one engine while being invisible in another. SoM gives marketing teams a single, comparable number to track AI visibility over time and to benchmark against competitors.
Key takeaways
- 1SoM = your brand mentions ÷ total brand mentions inside an AI answer set, expressed as a percentage.
- 2It must be measured per model - aggregating across engines hides the engines where you are losing.
- 3Query-set design matters: SoM is only useful if the queries reflect real customer intent.
- 4A SoM above ~25% in a category usually corresponds to top-of-mind AI recommendation status.
- 5SoM moves slower than SERP positions because it depends on both retrieval and parametric memory.
Why Share of Model (SoM) matters
SoM is the #1 KPI for AI Visibility, determining whether your brand exists in the AI-powered decision-making process.
When to apply Share of Model (SoM)
Quarterly AI visibility reporting
Track SoM month over month across a fixed query set to show executives whether AI marketing investment is producing measurable lift.
Competitive benchmarking
Compare your SoM against named competitors per model to identify which engines they dominate and which are still winnable.
Campaign attribution
Run SoM measurements before and after a content launch or PR push to quantify the campaign's effect on AI representation.
Best practices
- Lock a stable, representative query set (50-200 queries) before measuring trends.
- Always segment SoM by model and by query intent (transactional, research, comparison).
- Pair SoM with sentiment and rank-position metrics - high SoM with negative sentiment is a warning, not a win.
- Re-baseline after major model refreshes, since underlying weights have changed.
- Treat SoM under 5% as an invisibility problem, not a ranking problem.
Share of Model (SoM) and OGY
OpenGraphy's Share of Model tracker computes your SoM the right way: per model, per query intent, against a stable query set you control. Run a one-click audit and you see exactly which engines mention you, which competitors steal the answers, and how each campaign moves the needle over time. The benefit is a single comparable number you can put on an exec dashboard or in a board deck, broken down into the levers (entity, schema, citations) your team can actually act on. No more spreadsheets, no more manual prompting in four AI tabs. Ready to see your real SoM? Connect your brand to OGY, lock your query set, and the dashboard will start tracking it for you on autopilot.
Track your Share of ModelShare of Model (SoM) in context
As AI-powered discovery replaces traditional search, understanding Share of Model (SoM) is a core competency for modern brands. It sits in the GEO & AI Visibility family of concepts and directly influences how language models like ChatGPT, Gemini, Claude, and Perplexity surface, cite, and recommend brands.
Share of Model (SoM) - frequently asked questions
Quick answers to the questions people most often ask about Share of Model (SoM).
For a defined query set, run each query against each AI model, count how many answers mention each tracked brand, then divide your brand's mentions by total brand mentions across all answers. Normalize per model and per query category.
Share of Voice measures media or SERP visibility; SoM measures inclusion inside synthesized AI answers. SoV is shaped by ad spend and rankings; SoM is shaped by entity authority, content depth, and the model's parametric memory.
Category leaders typically score 25-45% SoM in their dominant model. 10-25% is competitive, under 10% means the brand is at risk of disappearing from AI-driven discovery.
Monthly for retrieval-based engines (Perplexity, AI Overviews) where content changes propagate fast. Quarterly is enough for parametric memory in the base models, since those update on training cycles.
Short-term spikes are possible through coordinated PR or paid placements that influence retrieval, but durable SoM gains require entity consistency, schema, and earned authority that survive across model refreshes.
Early studies from OpenGraphy customers show that a 10-point SoM increase in transactional query sets correlates with measurable lifts in branded search and direct traffic, particularly in B2B SaaS and high-consideration retail.
Related Terms
Generative Engine Optimization (GEO)
The practice of optimizing content and digital presence to rank highly in AI-generated answers from ChatGPT, Perplexity, Gemini, and similar platforms.
AI Visibility Score
A composite metric quantifying a brand's presence, authority, and favorability across AI model responses and training data.
Latent Space
The multi-dimensional representation space where AI models encode concepts, determining semantic relationships and proximity between ideas.