About Measure Protocol | Behavioral Data & Consumer Insights Platform

AI is a measurable shelf. Some brands already own it.

Salesforce's Connected Shoppers Report 2025 found that 39% of consumers — and 54% of Gen Z — have already used generative AI to discover and evaluate products. Gen Z is roughly 10× more likely than Boomers to use AI frequently for product discovery (Salesforce, 2025).

The implication is widely understood: AI is a shopping surface now. The under-discussed implication is that "shopping surface" is a euphemism for shelf — and shelves have winners and losers. Some brands appear when a consumer asks. Others don't.

For an NRF deliverable with Publicis (Jason Goldberg), we classified 674,038 ChatGPT prompts from 8,837 unique users across two categories: coffee and sunscreen. Two findings stood out.

Coffee: a single brand owns the prompt

Across 2,310 coffee-related prompts, the brand mention distribution was startlingly concentrated. Top brands that were frequently mentioned include: Starbucks, Nespresso, Dunkin’, and McDonalds’ Coffee.

Out of the top brands, Starbucks is mentioned roughly 5× more than the second-place brand and 10× more than the third-place brand. In ChatGPT, the coffee category is functionally a one-brand conversation. Whether the prompt is a comparison ("Starbucks vs McDonald's vs Dunkin'"), a hack ("cheap Starbucks drink in a venti cup"), or a product question, Starbucks anchors the discussion.

Sunscreen: nobody owns it yet

Sunscreen looks like a category before a winner exists. The top brand surfaced just 6 mentions across 571 prompts. Consumers are asking about SPF, application, and price — but the brand layer is up for grabs. For sunscreen marketers, that is either the most measurable competitive opening of the year or proof that the category hasn't yet developed AI-shelf habits at scale.

Real users behind real prompts

Most AI share-of-voice products today scrape their data, simulate the prompts, or rely on synthetic users. Our 674K prompts come from real users — 8,837 of them — with attached, demographically-validated panel attributes. So we can not only tell you that Starbucks dominates the AI coffee shelf; we can tell you the demographic of the people having the conversation.

The implication for brand and category teams

If you sell a category that is establishing AI habits — coffee, beauty, electronics, OTC pharma — the leaderboard matters more than the awareness scores you've been tracking for two decades. Awareness was the right KPI when shelf was fixed and the consumer walked to it. AI shelf is dynamic. The consumer asks, and the model surfaces a brand. That brand wins the consideration set in a way no shelf placement could.

If you're not on the leaderboard, you're not on the shelf. And nobody is going to scroll for you.

Two questions for category leaders

What is your category's AI share-of-voice today, measured against real prompts from real users — not simulations?
If your share is low, why? Is the model not surfacing you, or are consumers not asking for you yet? Those are two very different problems with two very different fixes.