KPMG study links trusted AI to stronger performance

KPMG has published a global study linking stronger AI transformation results to trust and governance. The survey covered more than 1,750 senior leaders across 20 countries.

The findings highlight a gap between rising AI adoption and broader business results. Many organisations are expanding AI use in specific functions without changing the wider operating model needed to turn those efforts into enterprise-level gains.

While 58 per cent of leaders consider enterprise-wide systems, processes, people and technology critical to transformation, only 12 per cent said their organisations deliver them effectively. The study also found that risk-led transformation produced the strongest performance improvement, at 14 per cent.

Workforce readiness emerged as another weak point. While 75 per cent of respondents expect benefits from humans and AI working together, only 19 per cent reported having a workforce ready for that shift.

Risk concerns were widespread, but integration remained limited. Nearly three in four respondents cited risk, security and privacy as major concerns, yet only 24 per cent said those issues are embedded in strategy and technology.

Measurement was also patchy. Just 28 per cent of organisations track operational or revenue outcomes linked to trusted AI, suggesting many still rely on adoption rates, qualitative signals or no formal measurement at all.

Operating model gap

The research argues that AI deployments often remain confined to individual use cases and are not fully tied to decision-making or end-to-end workflows. In that environment, productivity gains may appear in isolated parts of a business without translating into sustained organisation-wide improvements.

Legacy structures are part of the problem. Many businesses still operate with models built for stability rather than constant adaptation, making it harder to coordinate change across multiple teams and systems.

Adrian Clamp made that point in comments accompanying the research.

“Real value from AI requires operating as an intelligent enterprise – aligning strategy, decisions, and execution. Yet, most organizations have not redesigned themselves to do so, with complexity rising faster than performance. As a result, many risk scaling AI without delivering sustained enterprise impact or meaningful returns,” said Adrian Clamp, Global Head of Consulting Strategy & Investment, KPMG International.

Governance divide

The strongest performers were more likely to treat trust and AI governance as part of day-to-day operations rather than as a separate compliance exercise. The study linked that approach to better outcomes in areas including innovation, investment capacity and stakeholder trust.

Only a minority have taken that route. Most organisations still rely on reactive, siloed or partly integrated approaches to AI risk management.

Samantha Gloede said the issue goes beyond technical oversight.

“Trust is no longer a safeguard; it is a prerequisite for performance. As transformation scales across interconnected systems, organizations must be able to rely on decisions, not just data. That confidence is built through how risk is governed, managed, and embedded into execution. When it is, transformation can be directed, aligned, and scaled. When it is not, it fragments under its own complexity,” said Samantha Gloede, Global Head of Risk Services and Trusted AI Leader, KPMG International.

Broader shift

The study frames the findings as part of a wider change in how businesses compete. Rather than judging success by the number of transformation projects, it suggests organisations are increasingly being tested on whether they can coordinate change across the whole business.

KPMG described this as enterprise orchestration: the ability to align priorities, connect execution and manage trade-offs continuously across different parts of the organisation. The data suggests that without that coordination, AI investment may increase activity without producing equivalent returns.

The survey spanned sectors including technology, financial services, healthcare and manufacturing, indicating that the issues identified are not limited to a single industry. Across responses, a common theme emerged: AI adoption is moving faster than organisational redesign, leaving many companies with more complexity but not necessarily stronger performance.

One of the starkest findings was the contrast between ambition and readiness: 75 per cent of leaders expect gains from human and AI collaboration, but only 19 per cent say their workforce is ready.