The last mile of every leadership program.
Six months after the offsite, I could tell you attendance, completion, and a satisfaction score. Nothing about whether anything actually changed.
You know the offsite. Two days, good facilitator, real energy in the room. People make commitments. They write them on cards. Someone takes a photo of the wall.
Six months later, what happened?
I ran programs like that for years, and I'll tell you what I actually knew six months later: how many people attended, how many completed the follow-up modules, and what they scored the experience on a five-point scale. I could tell you nothing about whether a single manager was having a different conversation with a struggling direct report on a Tuesday afternoon.
That Tuesday is the last mile. It's the only mile that matters. And it's the one nobody walks.
Here's why. Every measure a program produces is a measure of the program. Attendance, completion, satisfaction. To know whether behavior changed, you'd have to observe the behavior, at the level of the person it was supposed to affect, and the person it was supposed to affect is not the manager. It's the people who work for the manager. Nobody asks them. Managers self-report, which tells you what they believe about themselves, which is not the same thing.
So the honest answer to "did it work" is almost always "we don't know," dressed up as a completion rate.
AI is about to hit the same wall
I bring this up now because AI is about to run into the same last mile with a much bigger budget.
When employees have both Copilot and ChatGPT available, three quarters pick ChatGPT. The dashboards measured provisioning and called it adoption. And I don't think the value gap is because the tools don't work. It's because nobody walked the last mile to find out whether the behavior took hold.
What walking it looks like
Asking the people whose work was supposed to change, in their own words, whether it did. Not "did you complete the training." Not "rate the program." Did anything about the way you work on an ordinary day actually shift, and if so, what, and if not, why not.
Not whether it launched. Not whether training was completed. Whether the behavior the organization invested in is being lived. That question is answerable. It just isn't being asked.