Wil Reynolds made an uncomfortable argument at SEO Week 2026: search could be measured as a performance channel, but AI visibility can't — at least not cleanly. Buyers ask an assistant, form an opinion, and may never click. When teams assume they can track everything, they start growing whatever is easiest to track instead of what actually moves buyers.
Marketing still has three jobs
Reynolds framed marketing's job as getting a business seen, believed, and chosen. AI answers mostly affect the first two. That makes visibility an opportunity metric rather than a revenue metric, and it should be reported that way.
What to report instead of clicks
- Are we showing up? Presence and citation share for a fixed set of buyer prompts.
- Are we described accurately? Accuracy score across assistants.
- Are we standing out? How we're described compared with competitors.
- Is it reaching the business? AI-referred sessions, self-reported attribution ("how did you hear about us?"), and pipeline trends alongside visibility.
Ryan Jones made a similar case: the useful questions are whether you show up, whether you stand out, and whether there's business impact, direct or indirect. He also shared client work where modelling showed AI Overview presence contributing to offline sales, though results varied by client.
Talk to buyers, not just dashboards
Some of the most useful AI visibility research Reynolds described was qualitative: interviewing real buyers about how they use AI when researching a purchase. A handful of conversations will tell you which questions matter and how much buyers trust AI for each step.
How to put it to leadership
Be honest that AI visibility is an influence channel. Report it next to pipeline, not as the cause of it. Agree on a small set of prompts that represent real buying decisions, and show progress on accuracy and presence quarter by quarter.