In 1999, I hired my first remote developer. His name was Adam. He was in Virginia. I was in Nashville. And I was completely, embarrassingly unprepared for what that meant.

The one thing I did right was set up a private IRC server. The thing I missed was that having the tool meant nothing if I didn't change how the whole team worked around it. I learned that lesson the hard way. Adam was gracious about it. I was not proud of it.

I think about Adam a lot these days.

Because right now, companies everywhere are rolling out AI tools the same way I rolled out remote work in 1999. Buy everyone a seat and a pail of tokens. Watch the usage dashboard. Announce the transformation. Move on.

No plan for how the software development lifecycle changes. No definition of what "working differently" actually looks like. No measurement baseline set before day one. Just the tool, and the hope that something good happens.

Here is the uncomfortable question nobody is asking at their AI all-hands: how many of your developers have their AI tool open right now and are still working exactly like it's 2022?

Seat counts measure purchasing, not practice. Usage metrics measure opening the tool, not changing the work. If you did not define what behavioral change looks like before you rolled out, you have no way to know whether anything actually changed. You are flying blind and calling it a strategy.

The organizations that pull ahead are not necessarily going to be the ones with the biggest AI budgets. They are going to be the ones that planned the behavior change, not just the purchase. They asked the hard questions before day one: what does "working differently" look like here? How will we know if it is actually happening? What has to change about how we work together for this to stick?

Originally posted on LinkedIn: "You Bought the Tokens, Now What?"