AI Metrics
Every platform shift rewrites the scoreboard.
Web 1.0 lived on hits and traffic; Social on Monthly / Daily Active Users.
Benedict Evans now asks the essential question: what number will define generative AI adoption, usage and scale?
One metric we’re focused on working with in house marketing teams is “time spent per day” using AI which tracks the “front-of-house” usage. In Figure 10 you can see the breakdown of light / medium / heavy AI users. We’re seeing real value coming from getting users to 30 mins a day using AI tools and emerging cohorts of experts who are all over the 1 hour a day mark.
Two further areas worth exploring in your AI strategy
The heavy lifting sits backstage: agents and APIs firing off silent LLM calls. We’ll need throughput metrics - tokens-per-workflow, calls-per-employee, latency-to-decision—to see that engine turning.
Taking a search mindset to “share of model”: getting brand data into training sets, embedding signals, and surfacing in agent answers.
Link to full AI Metrics article from Benedict Evans
