Special ── 2026-10-26
KPMG: 86% of Established-ROI Firms Run an AI Management Layer
KPMG surveyed 2,131 leaders: mature AI programs have formal management structures.
What changes
KPMG says competitive advantage in AI is shifting from adopting it to running it well.
Executives and HR leaders can use the presence of an AI management structure as a gauge of maturity. That structure covers controls, accountability and orchestration. For policymakers, the origin of models, their security and cost accountability are the issues in play.
Why it matters
According to KPMG, planned AI investment is rising. As investment grows, the ability to operate AI at scale comes under test.
KPMG asks whether organizations can govern where their models come from and secure what those models can reach. It also asks whether they can orchestrate how the models are used and account for what they cost. Scaling becomes hard if these four do not hold.
What the report says
KPMG states that competitive advantage is shifting from adoption to running AI effectively. It also says the maturity gap is narrowing.
Among organizations reporting established ROI, 86 percent operate a formal cross-functional or enterprise-wide AI management layer. In this context, the layer is the formal structure that governs AI use. Of those organizations, 65 percent have deployed it across the enterprise.
Among organizations at the experimentation stage, 31 percent report such a layer. Only 10 percent operate it enterprise-wide.
Who and what scope
KPMG published the report. It is the third-quarter edition of its quarterly global study, the Global AI Pulse.
The survey covers 2,131 senior leaders from organizations in 20 countries, territories and jurisdictions. KPMG's page does not give a breakdown by industry.
Why this report
KPMG asks what running AI at scale demands as investment rises. The question breaks down into governing, securing, orchestrating and accounting for cost.
KPMG also examines why AI maturity gaps between regions are narrowing. It looks at how organizations treat cybersecurity, model sovereignty and provider dependency as linked challenges. Model sovereignty is the question of whether an organization can keep control over its models.
How they studied it
According to KPMG, the study is based on 2,131 senior leaders. Findings for Q3 2026 were collected between 23 July and 26 August 2026.
The research examines AI adoption, maturity, governance, investment, challenges and business impact. KPMG's page does not describe the survey method in further detail.
Key points
- KPMG surveyed 2,131 senior leaders across 20 countries, territories and jurisdictions
- Among established-ROI organizations, 86% have a formal AI management layer, and 65% run it enterprise-wide
- Among experimentation-stage organizations, 31% have one, and 10% run it enterprise-wide
Conclusion
KPMG expects AI advantage to depend on running AI well rather than adopting it. Organizations with established ROI were much more likely to have a formal management layer.
The figures show a correlation only. KPMG's page does not say whether the management layer produced the ROI.
Quotes and sources
“Among organizations reporting established ROI, 86 percent operate either a formal cross-functional or enterprise-wide AI management layer, including 65 percent that have deployed it across the enterprise.”── KPMG, 2026-10-22
“The KPMG Global AI Pulse is a quarterly global study capturing perspectives from 2,131 senior leaders representing organizations across 20 countries, territories and jurisdictions.”── KPMG, 2026-10-22
“Findings for Q3 2026 were collected between 23 July and 26 August 2026.”── KPMG, 2026-10-22
“The question is whether organizations can govern where their models come from, secure what those models can reach, orchestrate how they are used and account for what they cost.”── KPMG, 2026-10-22
Source articles
- KPMG 22 Oct 2026
Written by AI (Claude) from the report and its coverage, then checked by Jev (TypeSafe's judgement model), a Gemini fact-check and a verbatim check of every quote.