AI is no longer operating at the margins of organisations - it is actively shaping how decisions are made, how information is interpreted, and how work gets done every day.
KPMG’s Emma Coogan examines how AI is challenging traditional governance models, and the frameworks needed to keep pace.
Across the organisations we work with, one pattern is increasingly clear: adoption is accelerating rapidly, but governance is not keeping pace. AI is already influencing outcomes long before organisations have aligned on how its use should be overseen, challenged or documented.
This creates a growing disconnect - between how leaders believe decisions are governed, and how they are actually being influenced in practice. Governance has always depended on clarity: clarity of responsibility, transparency of process, and confidence that decisions can be explained. AI is now testing each of these fundamentals in real time.
AI is no longer confined to specialist teams. It is summarising regulatory material, shaping client deliverables, informing internal recommendations and influencing operational decisions, often without being formally recognised as part of the governance landscape.
As a result, governance models - designed for more controlled and deliberate technology adoption - are being stretched in ways they were never intended to handle:
This is not a theoretical risk. It is already embedded within day-to-day operations across organisations of all sizes. Smaller businesses may adopt AI quickly because it helps them scale, but this speed can expose them to governance gaps if expectations aren’t defined early.
Those who respond well map where AI is already used, assess where influence is concentrated, and ensure governance grows alongside that capability.
AI fundamentally changes the governance equation. It increases both the volume and speed of decisions influenced by systems that are inherently probabilistic and not always fully explainable. It compresses decision cycles, making retrospective oversight more difficult.
And it challenges traditional assumptions about what “good judgement” looks like when outputs are co-produced by people and technology.
Without clear leadership direction, inconsistency emerges quickly. Different teams develop their own norms. Definitions of appropriate use drift. Some individuals rely too heavily on AI-generated outputs; others avoid using it altogether due to uncertainty.
That inconsistency, in itself, becomes a governance risk.
In practice, effective AI governance is less about the technology itself, and more about how it is used within the organisation. Three shifts are particularly important:
At KPMG, our Trusted AI approach focuses on embedding these behaviours into governance, risk and operational decision-making - ensuring that trust is both designed and demonstrated in practice.
The actions required are practical rather than transformational.
Organisations that get this right establish a strong leadership tone, grounded in Trusted AI principles. They embed AI into existing governance processes rather than creating parallel systems. They validate how decisions are made, not just how tools perform. They view AI as a contributor to judgement, not a substitute for it.
Most importantly, they build governance habits that scale: repeatable, practical behaviours that reinforce responsibility every time AI is used.
The actions required are practical and rooted in how transformation programmes are designed and delivered.
Waiting for “full clarity” is no longer viable; organisational exposure grows every time AI use outpaces governance.
The governance challenges associated with AI are not emerging — they are already embedded within how organisations operate. The question for leadership teams is not whether uncertainty exists, but how confidently it is being managed.
Organisations that lead in this space are not those with the most advanced tools, but those with the clearest expectations, strongest leadership behaviours, and most consistent application of Trusted AI principles.
The decisions being made now will determine whether AI becomes a source of resilience and trust - or a source of unmanaged risk. Our AI consulting team can help you navigate this landscape with confidence.
