Start with one useful task.
Our perspective on making the first conversation about AI smaller, clearer, and closer to everyday work.
The phrase “adopt AI” can describe almost anything. A new tool, a new workflow, or a change in how an entire team works. Our preferred starting point is more specific: one task, one person who knows it well, and a clear idea of what a useful result would look like.
Describe the work before the tool.
We begin by walking through the task as it happens today. What starts it? What information does it use? What decisions require someone’s judgment? Where does the result go?
That description is valuable even before a tool enters the conversation. It helps everyone see what the work actually involves and what should stay in human hands.
Define “better” in plain language.
We want to know what improvement means to the people doing the work. Less time preparing a first draft? A more consistent starting point for answering a question? Less effort gathering information from several places?
A clear aim gives a pilot something to test. An impressive demonstration is interesting; a result the team can review against its original need is more useful to the decision.
Keep the pilot small enough to understand.
Our approach is to start with a bounded workflow and examples the team is authorized to use. We discuss what information belongs in the tool and how a person will check its output.
The pilot is a learning exercise. If reviewing and correcting the result takes more work than the original task, that is useful information too. The goal is to discover whether the approach belongs in this particular business.
Build adoption into the conversation.
A workflow needs to make sense to the people expected to use it. We think about clear instructions, who owns the process, and how the team will flag an unexpected result.
Sometimes the next step is to expand the pilot. Sometimes it is to simplify it or try a different idea. Starting small leaves room for that judgment.