AI ambition at the provincial government of Veldhoven
A provincial government with 12,000 employees spends €38 million a year on IT. Its business information management department has 15 staff, mostly functional administrators. The current information plan dates from three years ago and focuses on replacing an ageing financial system. Sixty percent of the IT budget goes to maintenance and licences; data quality is poor, with three conflicting citizen registries. The new board has announced a 'digital first, AI in every process' agenda and wants a five-year AI strategy within six weeks. The board's letter speaks of 'innovation at speed' and 'data as the new oil'. Department heads are enthusiastic but have no concrete use cases. The CIO warns that any AI initiative will fail without master data management and privacy compliance. The BIM manager has been asked to lead the strategy, with two analysts and no extra budget.
1. How should the BIM manager respond to the six-week deadline and the board's 'AI in every process' framing?
Read the worked answer
Accept the deadline for a first version, but redefine the deliverable. An information strategy is a direction-setting document: it states which business goals information provision must serve, what capabilities are needed, what principles govern decisions, and how progress will be measured. It is not a catalogue of AI projects, and six weeks is not enough to produce credible use cases across 12,000 employees.
The BIM manager should interview department heads to convert 'AI in every process' into two or three concrete business problems, such as faster permit handling or better fraud detection, and test each against data availability and privacy constraints. The strategy can then present a phased roadmap: foundations first (master data, data governance, privacy compliance), then a small number of pilots with clear success criteria, then scaling. This respects the board's ambition while making it deliverable. The manager should also state explicitly what will not be done in the first year, because a strategy that promises everything is not a strategy.
2. Explain the relationship between information strategy, information planning and portfolio management in this situation.
Read the worked answer
The three processes operate at different horizons and must connect.
- Information strategy answers 'where are we going and why', aligning information provision with the provincial government's goals and setting principles such as 'citizen data has one master' or 'buy before build'.
- Information planning translates that direction into a multi-year plan: which capabilities, projects, budgets and dependencies are needed, and in what order.
- Portfolio management then governs the resulting set of investments, assessing business cases, balancing risk and capacity, and stopping or re-prioritising work as circumstances change.
In this case, the AI strategy is the information strategy layer; the ageing financial system replacement and master data work belong in the information plan; and the choice between AI pilots and other investments belongs to portfolio management. If the board treats the AI document as a project list, it will bypass planning and portfolio governance, and the result will be disconnected pilots. The BIM manager should make this hierarchy visible in the document itself, so the board understands what it is approving.
3. Which prerequisites and risks should be flagged to the board before any AI investment is approved?
Read the worked answer
The board should see four categories.
- Data: three conflicting citizen registries mean any model trained on them will produce inconsistent or unfair outcomes; master data management and data quality are prerequisites, not side issues.
- Legal and ethical: privacy impact assessments, transparency about automated decisions, retention rules and bias testing are mandatory in a public body, and they take time.
- Capability: 15 mostly administrative staff cannot deliver AI without new skills, whether hired, trained or partnered; vendor lock-in and dependency are real risks.
- Change capacity: sixty percent of the budget is already committed to maintenance, and departments are enthusiastic but have no use cases, which usually means adoption will be slower than planned.
The manager should recommend funding foundations before pilots, defining two or three pilots with measurable outcomes, and setting a review gate before scaling. Flagging these risks is not obstruction; it is what makes the ambition credible, and it protects the board from announcing a strategy it cannot deliver.