81% of shoppers have used AI in the last three months. 79% use it for research. 66% use it to compare products before buying a thing according to Inside Retail.

81% of shoppers have used AI in the last three months. 79% use it for research. 66% use it to compare products before buying a thing according to Inside Retail.

This confirms that AI models have gradually become the default first step in a lot of purchase journeys.

People want AI to inform the decision, not make it for them. Which means the research and comparison stage is where the real commercial battle now happens. For now, humans are still the ones deciding, but AI is the one handing them the shortlist.

MARK measures exactly that shortlisting moment: how often AI models actually mention and cite a brand. A few patterns keep repeating across the audits I've run:

→ Category leaders can hold a meaningfully wider lead in AI citation share than they do on Google. In one category set I audited, the gap between #1 and #2 ran as high as 2.4x.

→ Some brands get recognised by name, then drop off sharply the moment AI is asked to compare them against competitors. They're present, but effectively invisible at exactly the point 66% of shoppers say they're now shopping.

→ Machine-readable product data can beat brand scale outright. I've seen small, local brands out-cite national chains purely because their product feeds are structured and accessible to AI, while the competitor's isn't.

If AI is already the research assistant for 8 in 10 shoppers, the question that moves right to the top has to be: 'does it even know we exist, would it recommend us?'.

This is the gap MARK was built to close — measuring and improving how AI models perceive, cite and recommend brands, the same way SEO measured search a decade ago.