Boards see their AI oversight. The engineers running it do not.

The accountability gap inside every AI deployment lives between the boardroom that approved it and the engineers who field the alerts. It is a wiring problem, not a policy one, and the policy debates are inheriting the gap.

Virtana's UK survey of 238 enterprise decision-makers put a number on that distance. Asked whether their organization can automatically find the root cause of an AI failure, 59% of executives said yes. The infrastructure engineers who actually field those alerts said 34%. A 25-point gap inside the same firm, on the same question, in the same week.

The underneath reading is an observability story. Boards answer whether they have a process. Engineers answer whether the process works. The two are not the same question, and the policy implications track the engineer answer. With 53% of UK enterprises admitting they cannot fully see the AI infrastructure they run, the precondition for any UK General Data Protection Regulation (UK GDPR) or AI Act accountability is missing before the policy debate starts. Sector oversight in financial services, healthcare, and national infrastructure inherits the same gap.

UK enterprises are ahead of US peers on deployment and behind on the wiring. Every firm that adopts AI faster than it instruments it inherits the same gap. The firms that close it are the ones who let the engineers who catch the failures answer the surveys the boards already filled in.