The ROI Horizon: Navigating the Transition from AI Deployment to Enterprise Value
Addressing the challenges of accelerating AI investment and agent deployment, KPMG US tackles the question of why so many organizations are still struggling to realize returns despite surging spending. KPMG’s position is that execution—not capital or technology—now determines outcomes, with workforce readiness, governance, and operating models acting as the true constraints on AI value.
Why is scaling AI value harder than proving it?
This question has become unavoidable because the AI conversation has moved decisively past funding debates and pilot programs. Organizations are now committing serious capital to AI, with U.S. organizations projecting an average of $207 million in AI spending over the next 12 months, nearly double the prior year, while AI agents have crossed from experimentation into day‑to‑day operations.
At the same time, the gap between deployment and outcomes is widening. Sixty‑five percent of organizations now cite difficulty scaling AI use cases, nearly double the prior quarter, and 62% point to skills gaps as a barrier to demonstrating ROI. In other words, investment and deployment are no longer the limiting factors. Execution is.
Why It’s Harder Than It Looks
Turning AI agents into consistent business performance is difficult because deploying technology is easier than redesigning how work actually gets done. AI agents increasingly sit across functions, route decisions, and automate workflows, but most organizations still operate with structures, incentives, and accountability models designed for human‑only work.
The challenge compounds as agents move deeper into core operations. As reliance grows, questions of trust, control, and responsibility become operational issues, not abstract governance concerns. Leaders must decide who owns outcomes when humans direct agents, agents act autonomously within defined bounds, and decisions span multiple teams—decisions many organizations have not previously been forced to make.
The Evidence
1
2
3
4
5
KPMG’s Answer
KPMG’s position is that AI returns are no longer constrained by ambition, capital, or access to technology—they are constrained by execution across people, governance, and operating models.
As AI agents move into production, value depends on whether organizations can reengineer work at enterprise scale, not just deploy tools.
The Pulse data shows that AI agents are already coordinating work across functions, routing decisions, and supporting shared knowledge, yet organizations are still structured around fragmented ownership and unclear accountability. Without clearly defined human‑led oversight, decision rights, and escalation paths, agents accelerate activity without accelerating outcomes.
Governance is now inseparable from performance. Requirements for human validation of agent outputs have nearly tripled year over year (63% now require human validation, up from 22% in Q1 2025), signaling that trust and control are operational necessities. Organizations that treat governance as an early‑stage checkbox struggle to scale, while those that embed risk, security, and accountability into everyday workflows create the conditions for sustained AI performance.
Treat AI execution as an operating‑model redesign, not a technology rollout. Leaders should map where AI agents sit in workflows, who is accountable for outcomes, and how decisions move across teams before scaling deployment.
Prioritize workforce readiness alongside investment decisions. Upskilling, role clarity, and human‑agent oversight are now gating factors for ROI, and delaying them increases the risk that AI spend outpaces results.
Explore more
Get in touch
Start the conversation
Connect with our team today to learn how we can help you realize the full potential of GenAI.
