Deciding where AI acts, delegating without losing control, reading an operations report, assessing a supplier, knowing your obligations.
- What an agent can do, and what it must never do alone
- Ranking actions by reversibility
- Costs, risks, dependencies: the questions to ask
- AI Act and GDPR: what concerns the leadership team
Managers · 1 day
Delegating to AI without losing control
Bringing AI into a team's work: which tasks to hand over, how to check, how to respond to a mistake.
- Identifying the tasks that can be delegated in your department
- Organising human approval
- Measuring the real gain, not the advertised one
- Supporting a team through a change of method
Technical teams · 3 days
Building and running AI agents
The flagship programme: designing, coding, connecting and running agents in production, using real incidents as material.
- Agent architecture: tools, memory, orchestration, MCP protocol
- Choosing the model per task, keeping costs under control
- Safeguards: scope, human approval, logging
- Operations: daily checks, failure tracking, incident management
- Practical work: an agent built and put into service during the session