Let’s plant some hedgerows.

Five places to put a boundary around AI before you scale it — and the conversations that get you there.

In the 1930s, Kansas learned that the Dust Bowl was not only a drought. The state had converted an enormous amount of prairie into cultivated ground and had scaled production faster than it scaled stewardship. The response was not to stop farming. It was to change practices, build institutions, and plant shelterbelts — hedgerows — across open country.

A hedgerow does not stop the wind. It slows it down. It is worth knowing what kind of thing it is, because the same properties are what you want around AI.

It is planned
It forms boundaries
It costs you acres
It is precautionary
It is a community effort

Every one of these starts the same way.

Gather a team, and have a conversation. Not a committee and not a policy document — the people who actually touch the work, in a room, once. Everything below assumes you did that first.

Strategic Adoption

Find where AI creates real value before you scale it anywhere. Most AI projects that fail, fail at the choosing rather than the building.

Define a test. What would success look like, as a number?
Run a prototype. Small, cheap, and reversible.
Scale only what earned it.

Guardrails

A guardrail is not a refusal. It slows a failure down enough that you can survive it, which is the only thing a shelterbelt ever did for the wind.

Identify the risks. What could this get wrong, and who pays for it?
Assign accountability. A named person, not a team.
Define boundaries. Say plainly where the tool stops.

Governance

Write it down. One page is enough, and one page that people read beats twelve that sit in a shared drive.

Who should use these tools? Who shouldn't?
How should they be used, and on which data?
When do we revisit this?

Precautions

Transitions are hard on people, and the people who are worried are usually worried about something real.

Listen to concerns. They are information, not resistance.
Prepare a transition. Retraining, reassignment, timelines.
Remember that transitions end. Say when.

Coordination & Community

A technology arrives for everyone, but its costs and its benefits rarely land on the same people.

How will this affect my team?
How will this affect my community?
How will this affect me?

We scaled production faster than we scaled stewardship.

The technology worked. The system around it was not ready for the scale at which it was adopted. That is the failure worth avoiding this time, and it is avoidable.

Open & print the PDF