Start with something you want your business to do better. A worthwhile AI project begins with the work, not a list of tools.
Notice where the effort goes.
Look for work that repeatedly asks someone to gather, rewrite, search, or reorganize information. That might be a proposal assembled from old documents, a handoff missing its context, or a question that sends everyone looking for the same answer.
Useful work that never gets started matters, too. An idea waiting for a first draft may be as relevant as a task that already consumes your week.
Describe a result you would recognize.
"Use AI" is not a result. A project brief with the necessary details, a draft based on current facts, or an answer with a source you can check is much easier to evaluate.
Bring an example of the current work. Identify what makes the result useful, what usually needs correcting, and what would make the change worthwhile.
Keep the first project focused.
Start with one useful task and compare the result with the way you do it now. That makes it easier to see what helps, what needs checking, and what you would change.
Sometimes the best answer is a simpler process or ordinary automation. AI should earn its place in the solution, not be added because it is available.
Worth keeping in mind
A useful conversation can begin with one task, one example, and one thing you would like to change.
A question to take with you
Which piece of work would you be glad to stop piecing together by hand?
These notes are a starting point for discussion, not a delivery commitment. Sources, permissions, providers, and responsibilities depend on the project.