What share of model actually measures
So the idea is this: your target buyer opens ChatGPT or Perplexity and asks something like "which RevOps consultancies work with early-stage SaaS companies." Your brand either appears in that answer or it doesn't. Share of model is the percentage of relevant AI-generated answers that cite, reference, or recommend you across a defined query set.
That's a different thing from share of voice. Share of voice, in the traditional SEO sense, is the percentage of keyword search volume your domain captures in organic rankings. It assumes a results page the buyer scrolls through and picks from. AI search doesn't give them a page. It gives them a synthesized answer, and the sources it draws from are largely invisible to the person reading.
If you're not in the model's response, you're not in the consideration set. The buyer never saw your listing. They never had the chance to choose you.
Why the old measurement breaks down
Share of voice was built for a world with a list. Position one got more clicks than position five. There was a conversion logic you could trace from keyword to click to lead.
AI search collapses that. The model reads content, forms a view, and either trusts you enough to include you or doesn't. The buyer sees a paragraph, not a ranking. And for B2B buyers doing early research, that paragraph is increasingly where the consideration set gets formed, before any vendor interaction.
We've heard this directly from buyers we work with: "We've been investing in SEO for eighteen months. Organic rankings are fine. But leads from organic have been declining for over a year." The channel isn't broken. The assumption that keyword rankings still describe buyer behavior, that's what's broken.
The platforms are responding. Several AI visibility tools have emerged this year to measure exactly this. One AI visibility tracking tool launched an MCP integration this past quarter to pipe citation data directly into other tools in your stack. The market is building the measurement layer because practitioners need the number, and they need it in their existing workflows.
Share of model as a RevOps number
Here's where most teams get this wrong. This metric tends to land in marketing dashboards, next to brand lift surveys and social impressions. That's the wrong home for it.
If your ICP is running AI-assisted research before they ever fill out a form, then your citation rate is influencing top-of-funnel pipeline before any CRM record exists. That makes it a revenue metric, not a brand metric.
What this really is: the awareness-to-pipeline conversion point that attribution can't yet see. A buyer asks an AI "what's the best way to structure revenue reporting for a post-Series A company" and your name doesn't come up. You don't get the demo request. You can't track the loss because there's no session, no UTM, no form submission. But the loss is real.
Share of model belongs in that same conversation as pipeline coverage, channel mix, and conversion rates, not on a separate awareness dashboard that nobody acts on.
Most RevOps teams I talk to haven't started tracking this yet. It's sitting with the content team or the SEO function, sometimes with no owner at all. My feeling is that's a mistake that compounds over time: the earlier you make share of model a tracked RevOps metric, the earlier you start accumulating the baseline you need to act on it.
The market is noticing this shift. An AI search agency that repositioned its entire practice this year now describes its core work as engineering brand visibility in ChatGPT, AI Overviews, and AI Mode, rather than optimizing for blue-link rankings. That repositioning signals where the revenue concern has moved.
How share of model gets measured
There are two ways to do it, maybe three.
First: build a query set yourself. List the thirty to fifty questions your ICP is most likely to ask an AI assistant during their research phase. Run those queries monthly across ChatGPT, Gemini, Perplexity, and Google AI Overviews. Track whether you appear, whether you're cited, and how prominently. This is manual, but it's yours and it doesn't require a new contract.
Second: use one of the AI visibility platforms that automates this. A few of them track citation rates across the major AI platforms on a recurring basis. The methodology differences matter here: which query set, which models, how often, how the scoring works. I'd ask those questions before you buy.
Third: work backwards from your CRM. If buyers occasionally mention "I found you through ChatGPT" or "you came up in a search I was doing," log that. It's noisy and incomplete, but it's real signal about where the funnel actually starts.
My recommendation is to start with the manual query set before adding another tool. Build it from actual ICP research, not keyword tools. Run it quarterly. You want a directional number you can track over time, not a perfect number that requires a six-figure annual commitment.
The honest part about attribution
I don't have a great answer for how share of model connects to closed revenue yet. For better or worse, the attribution problem here is the same one we've had with brand advertising for decades, just in a new channel.
You can measure citation rate. You can watch it move when you publish more authoritative content. What you can't cleanly do, at least not yet, is draw a straight line from "we appeared in 40% of relevant AI answers last quarter" to a specific pipeline number.
That matters for how you position this internally. If you're putting share of model in front of a CRO, be honest that it's a leading indicator, not a trailing one. It tells you whether you're visible in the channel where early consideration is increasingly happening. It does not yet tell you the exact revenue impact.
That doesn't make it a vanity stat. It's similar to top-of-funnel organic traffic or brand search volume: a number that matters and requires interpretation. The difference is that AI search is growing fast enough that treating this as a future concern carries real pipeline risk today.
The teams getting this right are treating share of model the way they'd treat any early-stage pipeline signal: watch it, understand what moves it, and connect it upstream to the content and positioning decisions that eventually drive revenue. For more context on how this fits into a modern GTM measurement model, our posts on AI visibility and GTM measurement go deeper.