Every time you take your credit card out and buy something from a brand you’re familiar with, you know exactly why you’re choosing that brand. Sometimes it’s a recommendation. Sometimes it’s trust. Sometimes it’s reputation, and sometimes it’s simple familiarity. You’ve already considered the brand, and then a particular need or product makes you pull the trigger and actually buy it.
Byron Sharp gave marketers two conditions for growth: a brand needs to be easy to think of, which he calls mental availability, and easy to buy, which he calls physical availability. I agree with both. But in the current marketing landscape we can now see AI creating a third condition, and I’m calling it evidential availability.
For the past eight months I’ve worked on everything from AI visibility to brand health, growth and value creation. When I looked at the data, I could see this third condition building. I wasn’t sure how significant it was until I ran a series of reports, category by category, to understand what information a brand needs to be competitive in 2026 so that it can market to humans and machines at the same time.
Initially I wrote about trust. But I could see there were some categories where trust wasn’t as big a driver, and other things replaced it: reputation, data, backing from clinicians or substantiation from the media. There was always something. AI was looking for proof, for evidence that it could use to recommend a brand. After looking at the data and doing a bit of brainstorming, I liked the idea of evidential availability.
Evidential availability doesn’t build in the same way as brand fame, so it helps to start with how fame works. Share of voice creates awareness, awareness creates consideration, and consideration creates demand. When brands pull back from fame building ( advertising ), they lose share of voice, and over time they lose share of market. That’s why the advice has always been that brands should keep investing. Also, this isn’t “GEO” - we are not focused on semantics, agentic website meta data or key word stuffing via content production to appear in LLMs.
A brand can still be very famous and lose share of market, because at the point of consideration it doesn’t have the right product, or it doesn’t have the right evidence. That point of consideration is increasingly where an AI assistant sits. A famous brand can fail to be recommended, or never make it onto the list in the first place.