By Michael Domanic, Section Head of AI

If you’re serious about AI transformation, every employee at your company should be able to see their AI fitness score, and it should be explicitly tied to their performance.

I'm not talking about a token leaderboard. I'm talking about a methodology that measures whether they're using AI in ways that drive value.

At Section, we use our own product (Section HQ) to assess employee AI usage and assign people into four buckets: novice, experimenter, practitioner, and expert. The score is based on depth of usage - not just how often people are using AI, but whether they're using it to build and share agents, engage in multi-step reasoning, create scalable solutions, and so on.

Today, the overwhelming majority of people at Section are operating at or near the expert level. That means they are not simply using AI to improve their own productivity. They are building agents, workflows, and reusable solutions that create leverage across the company.

We don't publicly publish individual scores because we don't want people optimizing for the metric. We want them optimizing for real AI proficiency that drives business impact. I tell every employee: if you want to know your score, reach out to me and we'll discuss it. And when they do, I don't just say “you're a practitioner.” Based on the data, I'm able to tell them why they scored the way they did, what specific behaviors put them at that level, and what they'd need to do to change the way they use AI in a way that drives greater value to the whole of the business.

Right now, the low scores are more actionable than the high ones. When someone scores low, the signal is clear - they haven't built the habit yet, or they don't have the support they need, or nobody told them what was expected. I can work with that. 

When someone scores high, I need to dig deeper. Heavy AI usage doesn't automatically mean high business impact. The score gets me to the conversation, and the conversation is where I find out whether someone's building things that matter or creating automations that run every 15 minutes for no reason. 

The methodology isn't perfect. It doesn't fully account for AI used in point solutions outside our primary platforms. But it shows how deep people are going in our primary LLM (Claude) - which, at the end of the day, is the behavior we want to scale and incentivize. 

If AI proficiency matters at your company (it does), you need to say so explicitly. Measuring it behind closed doors and never telling employees where they stand sends a worse message than being transparent because you're making judgments about people without giving them a fair opportunity to improve.

Put it in performance conversations. Make it part of how managers evaluate their teams. Be upfront that this is a real expectation, not a nice-to-have.

Most importantly, invest in helping people succeed. Give every employee access to training, coaching, practical examples, and time to build new skills. Hold people accountable, but provide the enablement and support they need to meet the standard.

If you're thinking about doing this, my advice is to start with individual conversations rather than a company-wide ranking. Give people context, not just a label. 

Be willing to stand behind the methodology while being honest about its limitations. And connect it to performance directly - because if you don't, you're measuring something you're not willing to act on, and your employees will figure that out fast.

See you next week,

Michael 

Your fellow Head of AI