Can You Actually Measure 'Share of Model'?
"Share of model" is pitched as the agentic-era Share of Voice — how often AI assistants recommend your brand. Here's who's selling it, what's genuinely measurable today, and what isn't.
As clicks migrate into AI answers, a new metric is being pitched as the agentic-era equivalent of Share of Voice: "share of model" — how often AI assistants like ChatGPT, Gemini, and Perplexity mention or recommend your brand. A well-funded category of vendors is selling tools to measure it, and the money behind them is real. But before you add a "share of model" line to your dashboard, it's worth separating what is genuinely measurable today from what is still aspiration.
The vendors are real — and one is a unicorn
This is not a phantom category. The clearest signal of investor conviction: Profound, an 18-month-old startup building "answer-engine optimization," raised a $96 million Series C at a $1 billion valuation in February 2026, led by Lightspeed Venture Partners with participation from Sequoia and Kleiner Perkins — bringing total funding to $155 million. Profound says it works with more than 700 enterprise customers and 10% of the Fortune 500, including Target, Walmart, Ramp, MongoDB, U.S. Bank, and Figma. CEO James Cadwallader frames it as an expansion of search, not its death: "There's this idea that SEO is dead, and I very much disagree."
Profound is the scaled leader, but the field is crowded with verifiable players:
Evertune, Daydream, and Goodie — platforms that analyze how brands appear in AI-generated responses, track sentiment across model outputs, and identify which publishers shape model behavior (per Andreessen Horowitz's GEO analysis).
Peec AI — accessible generative-engine-optimization monitoring for smaller teams.
HubSpot's AEO Grader — a free brand-perception check across ChatGPT, Perplexity, and Gemini, with ongoing monitoring at $50/month, scoring brands on sentiment, presence quality, recognition, share of voice, and market position.
Semrush, BrightEdge, and Ahrefs (Brand Radar) — legacy SEO platforms that have added AI-visibility tracking, a sign the incumbents see the category as durable.
What's genuinely measurable today
The honest, defensible part of "share of model" is observational. These tools run large volumes of prompts across multiple AI engines and record the outputs. From that you can measure, with reasonable rigor:
Mention and citation frequency — how often your brand appears or is cited in answers to commercially relevant prompts, and how that compares to competitors. Profound, for instance, analyzes a dataset of more than 400 million prompts and surfaces "query fan-outs" — the multiple sub-queries a model generates from a single user question.
Sentiment — whether mentions skew positive, neutral, or negative.
Cross-model comparison — how ChatGPT, Gemini, Perplexity, and others differ in which brands they surface for the same query, which matters because they draw on largely distinct source sets.
This is real, repeatable measurement, and it is a legitimate proxy for visibility — the AI-era analog to tracking your brand's presence in search results.
What isn't measurable yet
The pitch outruns the proof in three places, and operators should hold the line on each:
Causation to revenue. Visibility in a model's answer is not the same as a sale. There is no clean, industry-standard way to connect "ChatGPT mentioned us more this month" to incremental revenue — the attribution chain that retail media and even traditional search can partially close does not exist here.
Stability. Profound's own research found that up to 90% of cited sources in AI answers can change over time, and different models rely on different sources. Even a well-optimized brand sees its citations fluctuate. A metric that swings that much is hard to manage to a target.
Control over the inputs. Here is the most sobering finding for anyone planning to "optimize" their way to share of model: McKinsey's AI Discovery Survey (August 2025, 1,927 consumers) found a brand's own website accounts for only 5–10% of the sources AI search platforms reference. The other 90% comes from publishers, user-generated content, affiliate sites, and review platforms. You cannot simply rewrite your homepage to win — the model is forming its view from what the rest of the web says about you.
Why the category exists at all
The structural driver is the decline of the click. Gartner predicted (February 2024) that "by 2026, traditional search engine volume will drop 25%, with search marketing losing market share to AI chatbots and other virtual agents." (That forecast predates our 2026 window and is included as the origin of the thesis.) If a quarter of search behavior moves into answers users don't click out of, then being inside the answer — and measuring whether you are — becomes a legitimate marketing concern. Reference rate, not click-through rate, is the metric a16z and others argue now matters.
The operator's verdict
Measure it as visibility, not performance. Track mention frequency, sentiment, and competitive share across models — it's a valid brand-health signal. Do not let a vendor present it as a revenue metric; it isn't one yet.
Start cheap. Use free or low-cost tools (HubSpot's AEO Grader, entry tiers) to establish a baseline before committing enterprise budget.
Invest where you have leverage. Since 90% of what AI says about you comes from third parties, earned media, reviews, and credible third-party coverage move the needle more than on-site optimization.
Expect volatility. Hold "share of model" to a trend line over months, not a daily number — the underlying citations are too fluid for tight targets.
Bibliography
Fortune, "As AI threatens search, Profound raises $96 million to help brands stay visible" (24 Feb 2026). Link
Andreessen Horowitz, "How Generative Engine Optimization (GEO) Rewrites the Rules of Search" (Profound, Goodie, Daydream). Link
Profound, "Best Generative Engine Optimization Tools for AI in 2026." Link
HubSpot, "AEO Grader" (2026). Link
Jarred Smith, "Why Brand Visibility in AI Search Is the Only Metric That Matters in 2026" (citing McKinsey AI Discovery Survey). Link
Search Engine Land, "Generative engine optimization (GEO): How to win AI mentions." Link