I Had Never Added Up My Own AI Usage Until a Job Posting Made Me Look

I was researching a new category of roles, AI enablement leads, AI operations leads, AI strategy positions, when one posting said something plainly enough to stop me: own Claude adoption and AI knowledge-sharing.

That phrase sent me down a research process before it sent me anywhere else. I asked AI to scan recent postings in this new category, compare them against roles I'd already studied, and pull out the actual requirements rather than the marketing language surrounding them. Then, before trusting any of it, I verified the posting dates, the company's stage, and their remote policy, because a role description means something different from a six-month-old listing at a company that's since changed direction.

The pattern across listings was consistent enough to be the real finding: what they wanted was proof of daily, hands-on AI use. Not certificates. Not course completions. Receipts.

That's the question that made me actually look at my own numbers for the first time, instead of assuming they existed somewhere without ever adding them up.

Skills built, apps deployed, sitting there the whole time, uncounted, because there was never a reason to count them until a job description asked the question directly.

The lesson isn't really about the job market, even though that's what triggered it. It's that most people doing serious daily work with AI have never quantified their own usage, the way you'd never think to count your own hours of practice at something until someone asks you to prove you've done it. The numbers exist. They're just sitting uncounted in session logs nobody's opened end to end.

Audit your own footprint

If you're someone who works with AI daily and you've never actually added it up, this is your prompt to do it. The explain-usage skill runs the exact same breakdown described here on your own session, if you'd rather have it do the counting for you.