Q2 2026: Everyone Adopted AI. Almost Nobody Is Getting Paid for It.

The headline number

Enterprise AI adoption has flattened out at the top. Stanford's 2026 AI Index, published this spring, puts organizational adoption at 88%, with 70% of organizations using generative AI in at least one function. Three years ago those figures sat near a third. The adoption race is effectively finished, and most companies won it.

The Federal Reserve, tracking the same question from the labor side in an April 2026 note, found that 54% of workers are now employed at firms using large language models, and individual work-related use of generative AI reached roughly 41%. Whichever angle you measure from, AI is inside most companies.

Look at the last bar. Adoption is near universal, but fewer than 10% of organizations have fully scaled AI in even one function. That is the story of the quarter: everyone is using it, almost nobody has built the muscle to run it at scale.

The gap that matters

Adoption is not the same as advantage. The widely-cited MIT figure from late 2025, that 95% of enterprise generative-AI pilots produce no measurable P&L impact, still describes most companies going into mid-2026. Gartner projects worldwide AI spending will hit $2.59 trillion in 2026, a 47% jump over 2025. Spending is not the constraint. Turning spend into results is.

The gap is not a technology problem. The models work. The gap opens because most companies bought tools without changing how work gets done, who owns AI decisions, or how risk gets managed. Buying access to a model is easy. Rebuilding a workflow around it is the hard part, and it is the part most organizations skipped.

Three signals from Q2

1. Agents got capable faster than companies got ready

The 2026 AI Index makes the point sharply on the model side: on the SWE-bench coding benchmark, performance jumped from 60% to near 100% in a single year. The capability is there. Deloitte's enterprise survey found only about 21% of organizations have a mature governance model for autonomous agents, so the readiness is not. Companies are deploying systems that act on their own faster than they are writing the rules for what those systems can decide alone.

2. The failure incidents are now being counted

The 2026 AI Index documented 362 AI incidents in 2025, up from 233 the year before. And 74% of organizations name inaccuracy as their top AI risk, which tracks with the finding that even leading models still produce wrong answers a meaningful share of the time. These are the predictable results of tools reaching production ahead of policy.

3. AI has started to reshape the org chart

The signal that matters most for how companies are organizing: AI governance roles grew 17% over the year, and postings for agentic-AI roles surged. Meanwhile the Federal Reserve data shows adoption climbing fastest in the most recent quarter. The companies pulling ahead are the ones putting someone in charge of AI, not just buying more of it.

What it looks like by industry

The averages hide wide variation. The Federal Reserve's April 2026 business data puts professional services at the front of firm-level adoption at 33%, financial services close behind at 30%, and manufacturing lower but growing fast, with AI use in finance functions up 127% year over year. Here is how the picture reads across the sectors we work in most.

The read for the next quarter

The competitive question has shifted. It is no longer whether a company uses AI. Almost all of them do. It is whether a company has built the operating model to turn that use into results: someone who owns AI decisions, governance that keeps pace with what gets deployed, and workflows redesigned around the tools rather than bolted onto them.

The companies in that under-10% who have fully scaled AI are not using better models than everyone else. They organized around AI on purpose. That is the whole difference, and it is the work we do inside our clients' companies as their AI Office.