AI competence and resourcing plan

What it does

It matches what the organisation deploys to what it can look after. The plan sets out the capabilities required across the three lines, from model validation and monitoring to audit and legal, measures them against current capacity, and commits resources to the difference. A senior executive owns it with the support of human resources. It is drawn up with the AI strategy, tested at each significant approval and refreshed annually.

How to implement it

• List the roles that AI depends on in each line, and the skills each needs.
• Assess current capacity, including reliance on single individuals and on suppliers.
• Require every business case to fund operation, monitoring and review as well as build.
• Close gaps by training, recruitment or bought-in expertise, with dates.
• Plan succession for key roles and keep knowledge documented.
• Report the position, and slow adoption where capacity is short.

Evidence it operates

Design: a capability assessment and resourcing plan approved at executive level. Operation: budget lines for operating and overseeing AI; training and recruitment records against the plan; business cases showing running costs; evidence of a deployment deferred for lack of capacity.

Where it fits

Set alongside strategy and tested through operation. It is the primary control for inadequate AI resourcing and competence, supports weak independent assurance of AI by equipping the second and third lines, and sits beside the AI literacy programme, which addresses general awareness.

Limits

The skills are scarce and contested, so a plan can be sound and still go unfilled. Capability bought from outside leaves when the contract ends unless knowledge is transferred. Plans count posts more easily than competence. Where the gap cannot be closed, the honest response is to do less, and that is the option most often left out.