Training decays fast. The difference between leaders who get value from an AI course and those who do not is almost entirely whether they attached the learning to a live problem while it was still fresh.
Pick the problem before week one
Choose one recurring, annoying, measurable task in your own function before the course starts. Report assembly, triage, first-draft writing, data cleanup. Write down how long it currently takes and how often it happens. That single number turns the whole course into a project.
Apply one thing per week
After each session, ask: what is the smallest thing I can change at work this week? Not a strategy. One change. Draft a use-case one-pager. Run a prompt comparison. Interview one person who does the task. Small weekly applications compound; a single grand plan at the end does not.
Bring your own data
Every exercise a course gives you can be re-run on your own context. Swap their sample case for your department. It takes twenty extra minutes and produces something you can actually use afterwards.
Build the artefacts you will need later
By the end you should hold three documents: a use-case shortlist, a risk and governance note, and a one-page business case. These are what get quoted in a promotion conversation - see explaining your new AI skills to your employer.
Tell someone you are doing it
Brief your manager in week one. Public commitment raises completion rates, and it means the work lands as visible progress rather than a surprise. It also sets you up for starting small with AI rather than proposing a large programme.
Choose a course built for this
Applied, blended courses make the habit easier. Leading AI Innovation & Organisational Transformation runs over six weeks precisely so each module lands against live work, and Strategic Application of AI in Business is structured around your own use cases.
The bottom line
Choose the problem before you start, apply one small thing a week, use your own data, and finish with three artefacts. That is the whole method.