I developed a concept for an AI Lab that combines speed and diligence. It turns governance from a brake into an enabler and translates the AI strategy into a clear process:

• Clear roles: the executive board sponsors, the AI board sets the strategy and decides on use cases, the AI Lab steers delivery, and a delivery team builds and runs the solutions.
• Value Sherpas: a newly defined role guides the business units from the idea to the business case and makes sure that only use cases with proven value move forward.
• Buy, adopt, make: three routes from standard solutions via low-code to in-house development. The route gives a first indication of the depth of review required. The actual answer comes from the impact assessment, meaning how strongly customers, employees and other stakeholders are affected, and from the risk assessment of the individual use case.
• Batch size one: each use case is assessed individually for value and risk, but runs through a standardised pipeline. Teams are put together per case from a central pool of specialists.
• Target operating model: four levels from strategy and governance through sourcing and data to operations with MLOps and AIOps, plus a RACI matrix and a steering process with risk assessment and a clear go/no-go decision, also aligned with the EU AI Act.