McKinsey’s annual State of AI survey is out, and the headline looks encouraging: 88% of organizations now regularly use AI in at least one business function. But the fine print tells a more complicated story — and if you’re a business leader trying to make sense of your AI investments, this is the report worth reading.

Surveying 1,719 professionals and executives globally, McKinsey found that AI adoption has genuinely accelerated. Generative AI use jumped from 33% in 2024 to 72% in 2026. That’s not a marginal shift. Businesses have moved.

The problem is what comes after deployment.

The ROI Question Remains Open

Only 37% of respondents say they can attribute any EBIT (earnings before interest and taxes) impact to their AI use. And that figure is essentially unchanged from 2025. Despite two years of broad adoption, the number of businesses actually seeing financial returns from AI has not grown.

Worse, only around 6% qualify as what McKinsey calls “high performers” — organizations attributing significant company-wide profit to their AI capabilities. That’s a small slice of the market doing the hard work of turning pilots into financial outcomes.

This doesn’t mean AI isn’t working. It means most businesses haven’t figured out how to make it work at scale.

Productivity Is Real — But It’s Staying on the Desk

80% of survey respondents say AI has improved their individual productivity. That’s a genuinely high number. People feel the difference in their day-to-day work.

But individual productivity gains and organizational financial gains are different things. What McKinsey is observing is a classic implementation gap: the benefits exist at the task level but aren’t flowing through to business outcomes. The reasons are predictable — workflows haven’t been redesigned, processes still have human bottlenecks, and most AI investments are still sitting on top of old operating models rather than replacing them.

The Workforce Reality Is Shifting

39% of respondents now expect their employer to cut jobs because of AI in the coming year. That’s up from 32% in 2025 — a meaningful increase in a single year. Whether that expectation materializes will depend heavily on how fast agentic AI systems can handle end-to-end tasks rather than isolated steps.

Why the Scaling Gap Matters

Nearly two-thirds of organizations in McKinsey’s survey haven’t yet begun scaling AI across the enterprise. They’re in function-level deployments, scattered pilots, or proof-of-concept territory. And the report identifies the main barriers to going further: security and risk concerns top the list, ahead of regulatory uncertainty and technical limitations.

This is the interesting part. The technology isn’t the bottleneck anymore. Governance, trust, and operational readiness are.

What This Means for Business

McKinsey’s report confirms what we’ve been seeing in practice: the businesses getting real value from AI are the ones who treat it as an operational redesign, not a tooling decision. They’re not asking “which AI tools should we buy?” They’re asking “how should our workflows, roles, and data systems change?”

For businesses currently in the pilot phase, the message from this report is clear: the window for experimentation is closing. Early movers who scale now will establish the operating advantages that are genuinely hard to replicate later. Those who keep adding point solutions without rethinking the underlying operations will continue to see productivity improvements at the individual level that never convert to competitive advantage.

The 6% who are getting it right share a common trait — they’re not deploying AI in isolation. They’re rebuilding how work gets done.

The Agentic Shift Coming Next

McKinsey flags agentic AI as the next critical scaling vector for enterprise. This is the shift from AI as a tool you use to AI as a system that runs processes. Rather than an employee using an AI assistant to answer emails faster, an agentic system handles the entire triage, response, and follow-up cycle with minimal human intervention.

This is exactly the territory where businesses need to move carefully. The security and risk concerns that are already the top barrier to scaling AI will intensify as AI systems take on more autonomous decision-making. Getting governance in place before scale — not after — is what separates the high performers from the rest.

---

Want to move from the 94% to the 6%? At Enterprise DNA, we work with businesses to build agentic AI systems that operate end-to-end, not just assist individual workers. Talk to us about an Omni advisory session to map where your business is leaving AI returns on the table — and what it would take to actually capture them.