Share of Model: a 2-minute briefing
Share of Model = the new Share of Search. Here’s what you need to know.
What is Share of Model (SOM)?
Share of Model is the term being used to describe a new metric that tracks how often a brand appears in AI-generated answers.
It’s the AI equivalent of ‘share of search’ – describing how often a brand is mentioned or referenced by large language models (LLMs) like ChatGPT, compared to its competitors. Just as traditional SEO optimises for Google rankings, Share of Model optimises for visibility in AI-driven search.
Welcome to the new battleground for brand visibility. Many of the rules are the same. But there are some crucial differences.
Why it matters
For years, we optimised for Google’s SERPs. Now, AI-powered search is reshaping discovery. Unlike traditional search, LLMs don’t rank websites – they generate the answers. That means fewer links and more direct responses where brands either appear – or don’t.
As people use AI chatbots for more and more queries, LLMs will increasingly influence customer behaviour. Brands that dominate AI-generated answers could win market share simply by being the name the AI remembers.
Just like SEO, Share of Model gives us insights on how a brand ecosystem compares to competitor brands, what drives the interest of consumers online in a given category, and how to develop content on the different platforms analysed by the different LLMs to stand out.
How it works
1. AI prioritises associations, not just keywords. Traditional search engines rank pages based on keyword relevance. LLMs, however, generate responses by recognising patterns and contextual links between concepts. If your brand frequently appears in trusted sources, AI models are more likely to reference it.
2. AI learns from broad datasets. Instead of simply pulling up ranked web pages, LLMs synthesise information from a mix of structured data, news, academic papers, and industry sources. Brands with strong media presence and thought leadership in key topics (e.g., cybersecurity, eco-friendly fashion) become more embedded in AI-generated responses.
3. AI values concise, high-authority information. LLMs favour facts, statistics, and quotable insights that can be easily integrated into responses. Studies show that AI is more likely to cite pages containing structured, authoritative data rather than generic marketing copy.
4. AI references structured databases and content. Wikipedia, WikiData and high-authority repositories play a significant role in AI training. If your brand has a strong presence in these sources, it increases the likelihood of being included in AI responses.
5. AI models evolve with user feedback. Many LLMs adjust outputs based on engagement signals and feedback loops. This means that over time, brands that consistently appear in user interactions and AI-driven research can shape how they are perceived and referenced by AI models.
The bigger picture
A year ago, we wrote in BN Edition about how search was likely to be affected by AI. Since then, Jellyfish launched a Share of Model platform marketing it as the first ever solution to track how LLMs perceive brands, products and services.
LLMs are no longer just tools; they’re becoming a critical part of the customer journey, as well as an audience in their own right.
But the core challenge for brands remains the same: cutting through the noise in a crowded market where consumers are overwhelmed with choices. AI-powered search offers customers a much higher level of personalisation than has been possible before. If you’re shopping for some new jeans and you can brief an AI on your own specific style, give it some reference images, maybe even upload photos of yourself, why wouldn’t you trust the recommendations more than a generic Google search where the top result is usually simply the highest bidder? Brands have an opportunity to really differentiate – get niche and specific about who they’re for so they always appear to the right people.
The bottom line is that if AI chatbots are answering consumer queries, brands need to be part of the answer.