The cheap option is training yourself. The common outcome of the cheap option is that you come back energised, nobody else has the vocabulary, and nothing changes.
Train yourself first when
- You are still deciding whether AI is relevant to your function at all.
- You need to build the business case before you can spend on anything.
- Your team is under delivery pressure this quarter.
One leader trained is enough to make the next decision. It is not enough to change how work gets done.
Train the team when
- You have already chosen a use case and someone has to run it.
- Adoption is your real risk, not selection. Most AI rollouts fail on people, not technology - see change management for AI adoption.
- Your team is quietly using AI tools already, ungoverned. Then training is a control, not a perk.
Who to include, in order
- The two or three people who will actually use the tool daily.
- Their line manager, so the change is supported rather than tolerated.
- One sceptic. Sceptics find the failure modes and, once convinced, carry others.
Avoid training everyone at once. Broad, shallow training is the most reliable way to spend a budget and change nothing.
Budget it honestly
Per-head fees are the small number. Time off the desk is the big one. A six-week blended course such as Leading AI Innovation & Organisational Transformation is £720 per person, and most providers will discuss group bookings - but the staff hours still need to be planned for.
Make it stick
Pair the training with one real task and a date. Practical patterns are in how to train your team on AI tools and applying course learning immediately.
The bottom line
Train yourself to decide, train the team to deliver. Send the leader first, then the small group doing the actual work - never the whole department at once.