Amplifying human potential with AI.

Find a useful first project—and keep people involved from the start.

Look for bounded, repeated work

A useful first AI project has a clear task, a defined set of inputs, and someone who can judge the output. Start with a draft, summary, or internal knowledge task where review fits naturally into the workflow.

Set boundaries before scaling

Decide what information the tool can access, which actions need approval, and how people will report mistakes. Give your team a clear way to override or correct the result.

Evaluate usefulness, not novelty

Try representative examples, including difficult cases. Compare the output with your existing process and ask the people doing the work whether it actually helps.