I've had the same conversation three times in the last five weeks.
Most recently, over the weekend in New Orleans, sitting with a group of business leaders comparing notes on AI.
Different industries. Different companies. Same question, every time:
Why aren't our people using this?
They've bought the tools. Sat through the demos. Signed off on the investment.
And still—nothing changes.
The mistake shows up the same way every time
They think they're making a productivity decision.
They're not.
They're dealing with an adoption problem and calling it something else.
So they start in the wrong place:
Will this make us faster?
Their teams are asking a different question entirely:
Why would I risk what already works?
That's the gap.
And that's where this breaks.
Not because people are resistant. Because they're being asked to make a bad trade.
Learn something new. Change how you work. Deliver the same output.
No one says it out loud, but everyone understands the deal.
Keep your head down. Stick with what works.
What leaders call resistance is usually self-preservation
Most teams are not slow to adopt AI.
They are protecting a system that already produces results.
That's not irrational. That's the job.
The failure isn't on the employee side.
It's in how the change is being introduced.
Where adoption actually starts
The companies getting traction aren't pushing harder.
They're doing something quieter.
They're not asking people to work differently.
They're showing them exactly where AI fits inside the work they already do.
Not a new workflow. Not a separate system.
A layer.
When AI shows up as one more thing, it gets ignored. When it shows up inside the work, it gets used.
That's the difference.
What this looks like when it works
I see this in my own work.
I write a column for the Dallas Morning News. AI doesn't write a word of it.
But it does something else.
It pressures my thinking.
I've built a simple system that holds my past columns, the standards I care about, and the patterns I'm trying to avoid.
It flags weak logic. Pushes on claims that aren't earned. Calls out a close that doesn't land.
The workflow didn't change.
The work got sharper.
That was enough to keep using it.
Start where the work already exists
This is where most organizations lose momentum.
They introduce AI as a capability.
What they need to design is a behavior.
Start with one task people already do every day. Insert AI into one step. Make that step easier the first time it's used.
Stop there.
No rollout language. No mandates. No transformation decks.
Just a better way to do something familiar.
How it actually spreads
If it works, people come back. If they come back, they push further. If they push further, it spreads.
Not because leadership forced it.
Because the work improved.
People don't adopt tools.
They keep what works.
The point most organizations miss
Productivity is what you measure.
Adoption is what you design.
Get the design right, and the productivity follows.
Get it wrong, and you've bought a tool nobody touches.