How do you prioritise AI investments?

The value-feasibility-risk prioritisation framework

Applying the value-feasibility-risk framework starts with a list of candidate AI use cases — all the ways AI might help the organisation. Each candidate is scored on three dimensions. Value: how significantly does this use case advance our strategic priorities — growth, efficiency, risk reduction, client experience? Feasibility: do we have the data quality, the governance structures, the technical capability, and the implementation capacity to make this work in the next six months? Risk: what EU AI Act classification applies, what data privacy risks exist, what are the consequences of errors, and what is the reputational exposure if the AI fails publicly? Use cases that score high on value and feasibility, and low on risk, are the first tranche of investment. Use cases that score high on risk, or low on feasibility, are deferred until the governance or capability gaps are addressed.

Why governance readiness affects prioritisation

Governance readiness is a feasibility constraint. A use case that would otherwise be high value and high feasibility becomes infeasible if the governance foundations needed to deploy it safely are not in place. For example: an AI system that assists with credit assessment may be technically straightforward to deploy, but if it is a high-risk AI system under the EU AI Act and the organisation has not completed the required conformity assessment, it cannot be legally deployed. A use case that generates client-facing documents using AI may look low-risk, but if the organisation has no AI use policy, no verification protocol, and no oversight mechanism, it is higher-risk than it appears. Governance readiness therefore feeds directly into prioritisation: building governance foundations unlocks use cases that would otherwise be too risky to progress.