1-2-1: Time saved is the wrong number

A sentence I keep hearing from founders this year is "I need two more people and I can't afford them."

You've probably tried to close a gap like that with AI. You may have a number for how much time it saves each week. It's a good number. But ask which hire it replaced, and the answer is usually none.

1 Principle

The honest test of whether it worked is the hire you no longer need to make.

Minutes saved are real, but they arrive in pieces. Ten minutes for one person, twenty for another, and most of it disappears into the next meeting or a bit more polish on work that was already fine. It only becomes money when the capacity adds up to a role and someone decides not to fill it. That decision sits with you.

2 Things Top of Mind

#1 Would I be paying someone for this?

That's the question I ask of my own AI team, one job at a time. In July I wrote a role brief for a part-time content designer. Content was capped at 30% of my week, there was a newborn at home, and visual production was eating most of that 30%. The brief was precise: four visual posts a week, produced in a monthly batch. I realised I'd have to manage the hire either way, human or AI agent, so I put Des, my AI design agent, to the test. The brief I wrote for the human turned out to be exactly what I needed for Des. After some intensive training, I decided I didn't need a human. That's the hire I didn't make.

This week: Write the brief for the next role you want to hire, before you post anything. List the outputs it would own each week. That list is your yardstick for judging whether anything else, a tool included, has already covered the job.

#2 Where the minutes went

It's budget season. Someone has multiplied minutes saved by headcount and put a big number on a slide, and usage looks healthy. Then you turn to next year's hiring plan, and it has the same roles it would have had anyway. Nobody did anything wrong. The time was saved. It went back into everyone's week a little at a time, and no one was ever asked to give a role back.

This week: Book thirty minutes with whoever owns next year's hiring plan. Go through it role by role and ask: "Is this role smaller because of the tools we bought this year?" If every answer is no, that's your real return so far, and those roles are where to point the tools next.

1 Question

Which role on next year's hiring plan should be smaller by now, and isn't?