Approach · July 2026 · 6 min
Ten pitfalls in data and AI strategies
Most strategies fail on missing decisions and a missing link between value and foundation.
Anders Butzbach Christensen
Plan & Proof
Recommendations without a decision
Align the hard recommendations with the CFO and executive team before the final presentation.
Use cases without kroner
A list without value cannot be prioritised and will not be funded.
Foundation without a portfolio
Data and platform without concrete use cases lose support after six months.
Too many initiatives
Choose 3-5 lighthouses and a small foundation programme, not 18 initiatives in three tracks.
Self-assessed maturity as truth
Supplement interviews with a broad survey and a fact check.
Technology choices without options
Show at least two alternatives with total cost before you choose build.
Activity targets instead of impact targets
The number of teams on the platform is a leading indicator, not the goal.
Forgotten running costs
Licences, compute and tokens must be in the business case from day one.
Agents without controls
No agent may act on its own without the seven controls.
A strategy that freezes
Agree the review cadence, e.g. as an annual update.
Sources: Plan & Proof.
The service behind this article: Data & AI strategy →