Before you introduce AI, name the work.

When a team asks which AI tool to use, I want to understand the work they hope it will improve.

What takes too long? What gets repeated? Where does information get lost? What decision needs better support?

Those questions make the conversation specific enough to act on.

Start with one task.

“We want to use AI” leaves a lot unresolved. “Our managers spend time turning meeting notes into actions, and we want to test a better way” gives us a task to examine.

We can look at what goes in, what comes out, who checks it and what happens if something is wrong. We can also decide whether AI is useful for that task at all.

Keep judgement visible.

Before trying a tool, decide who remains responsible for the work. A draft may be quick to produce, but someone still needs enough knowledge and time to check it.

Ask what information the tool may use, what an acceptable result looks like and where a person needs to make the decision. Work within the organisation’s approved tools and processes.

Listen to the people doing it.

A manager might see time saved. The team might see another system to open, another output to check or a new question about accountability. Bring those perspectives into the same conversation.

The questions people raise help define a better test. They tell us where the workflow is unclear and what support needs to be in place.

Make the first test small enough to learn from.

Choose a bounded task, agree how to judge the result and decide when to review it. Look at the effort required to use and check the output, as well as the time taken to produce it.

Then decide whether to adapt the approach, extend it or stop. Each can be a sensible result.

Before the next AI discussion, write one sentence describing the work you want to improve. Start there.