Only 14% of companies have a coherent AI strategy and have redesigned accountability to match, according to a new study.

getty

Companies are adopting AI so fast that Claude-maker Anthropic’s run-rate revenue exploded from $9 billion at the end of 2025 to more than $30 billion now, in just four or five months. But there’s a big problem, according to a global survey of 500 companies: no-one’s really in charge. Too often, there’s no human making decisions about who’s actually in charge, and who’s accountable if things break. Or, when they break.

“Too many organizations are scaling AI without redesigning accountability, which risks scaling bad decisions faster," Raj Sundaresan, CEO of Altimetrik, which commissioned the study, said in a statement. "Putting humans at the helm is about ensuring every AI-driven decision is governed with the same engineering rigor, ownership, and scrutiny we expect from any critical business system. Without that accountability, you’re scaling risk instead of intelligence.”

The top-line numbers from the report are not pretty:

• 14% of companies have a clear AI strategy
• 53% rely on humans as trust mechanisms in AI systems, but in unclear ways
• More than half of companies expect AI to trim workforces, with most planning to let that happen via natural attrition
• 52% of employees are scared AI is only here to replace them
• 7% of employees feel in control
• 80% of companies say accountability is unclear
• 72% of employees fear being blamed if AI experiments fail
• 6% of CEOs own AI strategy

In other words, we’re running full speed ahead but we’re not sure where the finish line is, or whether the course is clearly marked out, or what success actually looks like. AI is getting deeply embedded in operations but is structurally unmanaged. And AI is making decisions across workflows and influencing outcomes, but in most companies, ownership of those decisions – and their consequences – remains diffuse, unclear, or reactive.

That leads to what Altimetrik calls the “AI velocity gap,” the “widening distance between how quickly enterprises deploy intelligent systems and how slowly they redesign the human systems required to govern them.”

AI maturity in companies ranges from isolated deployment where people or teams pick up ChatGPT to help with a few tasks to managing AI across the company and continuously optimizing it. Only 13% of companies operate at the highest level of maturity, where they actually give AI high levels of autonomy in managing critical workflows.

That is, the report says, where actual “compounding advantage” lies, but also where building in human authority, checks and feedback loops matter the most.

Getting that part right is what actually delivers significant performance enhancements in customer impact, where 88% of high-AI-maturity companies get double-digit customer experience improvements, 41% report double-digit revenue growth and a quarter report much faster execution and idea-to-deployment timelines.

Most aren’t seeing those benefits, and the reason is a lack of AI leadership from humans who are supposed to be in charge, the report says.

“Enterprises are scaling AI faster than accountability, and that gap is now a workforce crisis,” says Phil Fersht, chief analyst of HFS Research, which conducted the research. “When leaders don't define what AI decides and what humans own, employees stop questioning it. That's not augmentation, it's abdication. Fix it now, or you're not building an intelligent organization. You're scaling unmanaged risk."

Ironically, the solution to employing AI better is better human management.

That’s good news: there are important roles for humans in the AI-driven company of the future. But the caveat is that people who are scared AI is going to take their jobs are probably not going to be the best at helping to figure out where AI will provide the most corporate advantage.

Solving that fear factor is probably a good first leadership step in transforming an enterprise for AI.