Most AI productivity talk is about speed. How fast did the draft get written? How quickly was the report summarized? How many minutes did the tool save?

That matters. But I think there’s a better question: what work no longer needs to be done?

Faster is useful. Removed is better.

If AI helps someone finish a task 20% faster, that’s a win. But if the task still needs the same setup, the same checking, the same formatting, and the same handoff, the workflow hasn’t really changed.

The bigger opportunity is when part of the work disappears. An executive no longer turns meeting notes into follow-ups by hand. A partnership lead no longer rebuilds the same recap structure every week. A manager no longer reads a 40-page document to find the five things they need. A team stops explaining the same context to AI every time. That’s not just faster work. That’s capacity being created.

Time fills back up fast

Save 15 minutes on a task but keep a clunky process, and that time usually vanishes into more email, more cleanup, more admin, more back-and-forth. So don’t stop at “did AI make this faster?” Also ask: “did we redesign the work enough that something can now be skipped, automated, or handled once instead of repeatedly?” That’s a much better test.

Next time you evaluate an AI use case, don’t just measure how fast the tool spits out an answer. Ask what part of the process should no longer exist. That’s where AI starts creating real leverage.