Only 11% of enterprises have scaled AI agents to deliver enterprise-wide outcomes, despite planning an average $186 million in AI spend, according to KPMG. The disparity highlights a widening gap between investment and measurable margin impact.
KPMG’s Global AI Pulse survey shows 64% of organizations report “meaningful” results from AI, but most gains remain incremental. Firms categorized as “AI leaders” are deploying agent-based systems that automate decisions and coordinate workflows across functions. Regional spending varies, with Asia-Pacific averaging $245 million versus $178 million in the Americas and $157 million in EMEA.
Why Are AI Agents Failing To Deliver Full Value?
The divide reflects fundamentally different deployment strategies. Among AI leaders, 82% report meaningful value, compared with 62% of their peers, a 20-point gap tied to process redesign rather than tool adoption. In IT functions, 75% of leaders use agents to accelerate development, versus 64% elsewhere, while operational use cases show a similar spread.
“The results reinforce that spending more on AI is not the same as creating value,” said Steve Chase, Global Head of AI and Digital Innovation at KPMG International.
He added that leading firms are redesigning workflows around agents rather than layering tools onto existing systems. Could this architectural shift determine which firms convert AI spend into sustained margin gains?
Infrastructure and governance remain key constraints. Many enterprises underestimate integration costs, including data pipelines, vector databases, and compliance systems needed for auditability. Governance maturity also correlates with confidence, with 49% of AI leaders expressing readiness to manage risks versus 20% among early-stage adopters.
Regional dynamics add complexity to global deployment. Asia-Pacific leads in agent scaling at 49%, compared with 46% in the Americas and 42% in EMEA, while cultural expectations around human-AI collaboration vary significantly. These differences affect how agentic systems are designed and deployed across multinational operations.
Seventy-four percent of respondents expect AI to remain a top investment priority even during a downturn, signaling sustained capital allocation despite uncertain returns. The next catalyst will hinge on whether enterprises shift spending toward operational infrastructure and governance needed to scale agent-driven workflows.
