Companies Say AI Needs Managers and Offices to Actually Work, New Data Shows

Companies have spent two years throwing money at AI. The data now coming in says the tech isn't the hard part. The people running it are.

Gallup surveyed 102 chief human resources officers through its Global CHRO Roundtable and found 99% call AI important or very important to their company's strategy. That's not surprising. What is surprising: half of those same CHROs say they're not confident their managers can actually guide employees on using AI at work.

Gallup's Q1 2026 workforce study found AI adoption doesn't automatically make workplace culture better or worse. It just makes things move. In companies that haven't touched AI yet, 59% of employees say culture stayed the same over the past year. In companies that have adopted it, that drops to 51%, with the rest split almost evenly between culture getting better (24%) and getting worse (25%).

The deciding factor, according to Gallup, is the manager. Employees who strongly agree their manager champions AI are far more likely (33%) to say AI actually transformed how work gets done, versus just 4% among those who don't feel that management support. Employees whose managers actively back their AI use are also more likely to report their culture improved (31% vs. 21%).

Companies know this. Gallup found 57% of CHROs are now running AI training specifically for managers, and 62% are standing up internal AI centers of excellence or champion programs. Whether it works is a separate question nobody can answer yet.

The Agent Problem

Deloitte's research, reported by HR Dive, found only 1 in 5 organizations are actually prepared to move toward autonomous AI agents, the next stage past chatbots and copilots. The holdup isn't the AI. It's the plumbing: poorly documented processes, fragmented data systems, and work habits nobody's bothered to update in years.

Most leaders don't even want full automation. Deloitte found 75% of business leaders surveyed believe the bigger payoff is human-agent collaboration, where employees direct the AI, check its work, and step in when judgment calls come up, rather than just letting the machine run.

But it's also harder to execute. More than 4 in 10 leaders in Deloitte's report expect "a lot" to "extreme" disruption in the next 12 to 18 months as job roles get rewritten around this collaboration model. Deloitte recommends companies treat AI as a layer on top of existing processes first, not a wholesale replacement, giving employees real hands-on time and trust-building before expecting them to hand off real authority to a bot.

A separate Culture Amp report cited by HR Dive found only about a third of HR professionals say they've actually moved beyond basic task-level AI use toward these agentic workflows. Most companies are still using AI as a smart assistant, not an autonomous operator with defined authority.

Insurance And The Office Floor

The theory plays out concretely in two very different corners of the economy. David Fishel, chief growth officer at Higginbotham, a Texas-based insurance and financial services firm that crossed $1 billion in revenue for the first time in 2025, told Insurance Business that AI is compressing how fast new producers learn the technical side of the job, freeing up veteran advisors to spend more time coaching instead of training. "AI is accelerating how quickly someone can learn the technical side of this business, but it doesn't replace wisdom, judgment or relationships," Fishel said. He's describing a labor market problem, not a tech problem: insurance has an aging workforce, and AI is one lever to move new hires up the learning curve faster.

Forbes contributor Ryan Anderson made a related but separate argument: as AI absorbs routine and information-heavy tasks, the value of human work shifts toward judgment, creativity, coaching and relationship-building, which Anderson argues require in-person collaboration. He points to firms like Atlassian and Zoom redesigning offices around mentorship and team co-creation rather than solo desk work.

It's also an argument from a workplace-design contributor with an obvious professional stake in offices mattering more, not less. Nothing in the sourcing here proves remote or hybrid teams can't build the same judgment and mentorship AI supposedly demands. It's a plausible theory, not a settled finding.

What Remains Unresolved

None of this data answers the question every executive actually wants answered: does AI adoption pay off in revenue or productivity terms strong enough to justify the disruption Deloitte's leaders expect. The Federal Reserve Bank of St. Louis released a study in late July 2026 finding U.S. executives are largely optimistic about AI's future impact on productivity, according to HR Dive, but optimism isn't a balance sheet. The gap between what CHROs say they're planning and what Deloitte finds companies are actually structurally ready for, 1 in 5, is the number worth watching over the next 12 to 18 months as that disruption window closes.