Enterprises are spending an average of $186 million on AI in the next 12 months. Ninety-five percent have an AI strategy. Thirty-nine percent are scaling AI across the enterprise. But only 8% report tangible return on investment. That's the headline from KPMG's Global AI Pulse Q1 2026 survey of 2,110 C-suite leaders across 20 countries—and it exposes the brutal reality of enterprise AI adoption in 2026.
The gap isn't about technology access, budget constraints, or willingness to experiment. Most organizations have those boxes checked. Fifty-eight percent prioritize IT infrastructure spending. Fifty percent are boosting cybersecurity and data protection. The difference between the 8% seeing ROI and the 92% burning cash comes down to something most CIOs miss: organizational structure, governance maturity, and workforce readiness.
The AI Leaders: 11% Doing It Right
KPMG identified 11% of surveyed organizations as "AI leaders"—companies demonstrating the ability to translate AI investments into measurable business outcomes at scale. These leaders aren't just running pilots or experimenting with tools. They're orchestrating multi-agent AI systems across business functions and delivering enterprise-wide value.
The numbers tell the story. Eighty-two percent of AI leaders report meaningful business value from AI deployments. Among non-leaders still in pilot mode, that number drops to 62%. AI leaders are also 2.5 times more confident in their ability to manage AI-related risks. And according to separate research from an enterprise security company, AI leaders achieve 1.7x revenue growth, 3.6x three-year Total Shareholder Return, 2.7x return on invested capital, and 1.6x EBIT margin compared to laggards.
What separates these leaders from the rest? Three uncommon traits that most enterprises overlook:
1. They Measure AI Performance Across Functions, Not Just at Project End
"What sets AI leaders apart is that they have a clear link between their AI activity and the business results," said Samantha Gloede, global head of risk services and global trusted AI leader at KPMG International. "They use consistent performance metrics across functions, and they have visibility into impact as systems operate—not just at the end."
Most enterprises measure AI ROI the way they'd measure any IT project: Did we deliver on time? Did we stay on budget? Did the pilot work? AI leaders measure differently. They track business outcomes in real-time—revenue per customer interaction, cost per resolved support ticket, time to close deals, error rates in financial reconciliation. They instrument AI systems to surface value continuously, not just in quarterly reviews.
This isn't about dashboards and KPIs. It's about embedding measurement into how AI operates across the business. When a customer service AI agent resolves an issue, the system tracks not just resolution time but customer satisfaction scores, repeat contact rates, and downstream revenue impact. When a procurement AI flags a supplier risk, the system measures both the accuracy of the prediction and the cost avoided by acting on it.
2. They Build Governance Into Agent Design, Not Bolt It On Later
Governance is the defining challenge of 2026. While 52% of surveyed organizations use AI to automate workflows across functions, only 9% have orchestrated multiple AI agents across workflows. The reason? Fragmented systems, unclear ownership, and governance models designed for single-purpose tools—not for autonomous agents making decisions across teams.
"When you start coordinating multiple AI agents across business functions, getting the governance right is both difficult and vital," Gloede explained. "CIOs need to be clear about who owns decisions made by agents, because once agents operate across teams, decisions don't sit in one place anymore."
AI leaders don't treat governance as an afterthought. They design governance frameworks before deploying agents—defining ownership, accountability, risk thresholds, and escalation paths as part of the system architecture. This means:
- Clear decision ownership: Every AI-driven decision has a named human owner, even when the agent operates autonomously
- Real-time monitoring: Observability built into agent workflows, not added post-deployment
- Adaptive controls: Governance rules that evolve as agents scale, not static policies that break under production load
The data backs this up. Eighty-one percent of AI leaders report having the capabilities and governance to manage AI risk at scale. Among non-leaders, that number is 63%. The gap widens as ambitions grow. Leaders who successfully orchestrate multi-agent systems across functions consistently cite governance as the enabler—not the barrier.
3. They Invest in Workforce Readiness Through Real-World Immersion
Only 22% of surveyed organizations are "very confident" their talent pipeline can meet the needs of an AI-enabled workforce. Twenty-five percent identify workforce readiness as a top challenge. The issue isn't hiring AI specialists or recruiting data scientists. The issue is preparing the existing workforce to work alongside AI systems—and most enterprises approach this through abstraction rather than immersion.
AI leaders take a different path. They embed AI skills training into real workflows, creating sandbox environments where employees experiment with AI tools in immersive, real-world simulations. Some launch internal competitions, offering cash prizes to teams that develop AI solutions delivering measurable business value—either for client work or internal operations. This approach builds muscle memory, not just theoretical knowledge.
KPMG uses this model internally. "Approaches like these help us build a workforce that can adapt and thrive as our profession evolves," Gloede noted. The goal isn't to turn every employee into a prompt engineer. The goal is to normalize AI as part of the daily workflow—so employees understand when to lean on AI, when to override it, and how to collaborate with autonomous agents without creating bottlenecks.
The Cost of Getting It Wrong: Why 92% Fail to See ROI
The flip side of the AI leaders story is the 92% burning millions without tangible returns. The reasons are structural, not technological.
First, most AI strategies are performance art. A recent survey by Writer and Workplace Intelligence found that 75% of executives admit their company's AI strategy is "more for show" than actual internal guidance. Thirty-nine percent lack any formal plan to drive revenue from AI tools. Forty-eight percent call adoption a "massive disappointment." When strategy is theater, ROI becomes a mirage.
Second, organizations are creating a two-tiered workforce without addressing the cultural fallout. Ninety-two percent of surveyed C-suite leaders are actively cultivating "AI elite" employees—super-users who leverage AI for 5x productivity gains. Meanwhile, 60% plan to lay off employees who can't or won't adopt AI. This divide breeds sabotage, not transformation. Twenty-nine percent of employees (44% of Gen Z) admit to actively undermining their company's AI strategy. Seventy-three percent of CEOs report stress or anxiety about AI transitions. Sixty-four percent fear losing their jobs over AI failures.
Third, security and governance gaps are creating real business risk. Sixty-seven percent of executives believe their company has already suffered a data leak or breach due to unapproved AI tools. Thirty-six percent lack any formal plan for supervising AI agents. Thirty-five percent admit they couldn't immediately "pull the plug" on a rogue agent. These aren't hypothetical risks. They're production failures waiting to happen—or already happening in the shadows.
Fourth, productivity gains don't translate to enterprise ROI without structural transformation. AI super-users deliver 5x individual productivity gains, yet only 29% of organizations see significant ROI from generative AI. The gap is clear: individual wins don't scale to organizational outcomes without changing how work flows across teams. Pilots succeed. Enterprise transformation fails.
What CIOs Should Do Right Now
If you're a CIO or technical leader evaluating AI investments in 2026, here's what the data says about where to focus:
Build Governance Before You Scale Agents
Don't wait until multi-agent systems are in production to figure out ownership and accountability. Design governance frameworks as part of the architecture—defining who owns AI-driven decisions, how risks escalate, and what controls adapt as agents scale. Treat governance as an enabler, not a compliance checkbox.
Measure Business Outcomes in Real-Time, Not Just at Project End
Instrument AI systems to surface value continuously. Track revenue impact, cost savings, error rates, and customer satisfaction—not just API calls and inference latency. Build performance measurement into how agents operate across functions, so you can see what's working (and what's burning budget) before quarterly reviews.
Invest in Workforce Readiness Through Immersion, Not Abstraction
Stop teaching AI through PowerPoint decks and online courses. Create sandbox environments where employees experiment with AI in real-world workflows. Run internal competitions that reward measurable business outcomes. Normalize AI as part of daily work, not a separate skillset to learn.
Consolidate AI Tools to Reduce Security and Governance Risk
If 67% of executives believe they've already had AI-related breaches, the issue isn't just unapproved tools—it's tool sprawl. Consolidate around enterprise platforms with built-in governance, observability, and security controls. Reduce the attack surface by limiting the number of AI vendors and tools employees can access without oversight.
Don't Pilot Forever—But Don't Scale Without Governance
The window for pilots is closing. Thirty-nine percent of organizations are already scaling AI across the enterprise. But scaling without governance, workforce readiness, and real-time measurement is how you join the 92% burning millions without ROI. The path forward isn't more pilots. It's selective scaling of high-value use cases with governance and measurement built in from day one.
The Bottom Line
Enterprises will spend an average of $186 million on AI in the next 12 months. Ninety-five percent have AI strategies. Only 8% see tangible ROI. The difference isn't budget, technology access, or willingness to experiment. The difference is governance maturity, measurement discipline, and workforce readiness.
The 11% of organizations getting it right aren't smarter or luckier. They're building governance into agent design, measuring business outcomes in real-time, and preparing their workforce through immersion rather than abstraction. They're treating AI transformation as an organizational challenge—not a technology deployment.
For the 92% still burning budget without returns, the path forward is clear: stop piloting, start governing, and measure what matters. The technology works. The question is whether your organization is structured to capture the value.
Want to calculate your own AI ROI? Try our AI ROI Calculator — takes 60 seconds and shows projected savings, payback period, and 3-year ROI.
Continue Reading
- Enterprise AI Strategy: Why 75% of Executives Admit Their Plans Are Just for Show
- The Two-Tiered AI Workplace: How 92% of Companies Are Creating a New Elite
- AI Governance for Multi-Agent Systems: What CIOs Need to Know Before Scaling
