Neil Murray does not sound like an executive preparing to automate away a workforce. As the head of Real Estate Management Services at commercial real estate giant JLL, he runs an organization of more than 88,000 people charged with keeping nearly 6 billion square feet of buildings operating day to day. Yet when the conversation turns to artificial intelligence, he recoils at corporate jargon.

“I happen to despise the term ‘headcount’ in the world of AI,” Murray told Fortune. “I think it sort of dehumanizes what we’re talking about.”

Murray rejects the notion that AI will shrink his workforce, but JLL’s internal staffing record tells a more cautious story. Asked whether headcount in his 88,000-person division has actually grown over the past three years, the company declined to provide figures ahead of earnings, calling headcount merely “stable” and projecting it to stay flat or expand only modestly in 2027.

The company instead offered a candid redefinition of corporate growth. “Stability does not equate to stagnation,” a JLL spokesperson told Fortune. “In an environment where AI is genuinely improving individual productivity, holding headcount flat while increasing output is actually a meaningful result.”

Murray, for his part, advances an article of faith increasingly rare among corporate leadership: automation will not shrink his division. “This is my profound belief,” he said. “I say that as a leader and as a parent—a profound belief that we will create better jobs, more impactful jobs, more human-centric jobs—and automate some of the jobs that people didn’t enjoy doing anyway. The notion that AI will somehow reduce the number of people we have serving this industry is just not right.” JLL does not typically report year-over-year changes in headcount for Murray’s division. For outside observers, the earnings report the company declined to pre-empt is the nearest scheduled checkpoint for the division’s staffing trajectory.

It is an earnest, deeply appealing defense of physical work. But a few pages into JLL’s own research, provided exclusively to Fortune ahead of its release, the arithmetic begins to push back.

The cold logic of procurement

The tension sits squarely inside JLL’s 2026 Global State of Facilities Management Report, a survey of 1,321 organizations across 15 countries. While Murray called technology an unalloyed amplifier of human talent, the organizations surveyed are planning an explicit shift away from payroll.

Eighty-nine percent of respondents agreed that facility management budgets will shift from labor toward technology and automation. More critically, 87% expect outsourcing contracts to abandon labor-based models in favor of outcome-based pricing by 2030.

That subtle contractual change rewires the financial incentives of commercial property. Under traditional facilities contracts, providers bill clients for the hours and people they deploy on-site; more staff translates directly into more revenue. Under outcome-based agreements, clients pay fixed fees tied to performance metrics—unbroken uptime, completed work orders, or reduced energy consumption. In that structure, every hour of human labor ceases to be a revenue stream and becomes a direct cost against the provider’s profit margin. At the scale JLL operates—more than 88,000 people across nearly 6 billion square feet—that distinction determines whether each deployed worker appears on the revenue side of a contract or the cost side.

The wrong kind of robot

Murray dismissed fears of job replacement by pointing to the limits of current robotics. “Humanoid robots doing engineering work are a long way off,” he said, adding that this was his own opinion.

He is very likely right that an android will not be diagnosing a complex commercial boiler system anytime soon. But corporate real estate does not need humanoid robots to trim staffing levels.

JLL’s report notes that client adoption of robotics is poised to nearly double, jumping from 37% today to 71% by 2030. The company concedes that this surge will come from “purpose-built technologies, not humanoids”—specifically autonomous scrubbers, robotic lawn care, and remote security systems. These machines do not displace licensed HVAC engineers; they displace cleaners, security personnel, and access-control staff. Both the robotics adoption curve and the pricing-model shift are set against the same 2030 horizon, according to the survey’s respondents.

Murray’s confidence rests heavily on demographic scarcity. In key markets, he noted, JLL cannot find enough candidates to fill open roles, with 68% of organizations reporting that skilled-trades shortages disrupted operations over the past year. Across the broader economy, employers require an estimated 1.7 million skilled-trades workers annually through 2035, even as nearly a quarter of them are 55 or older.

“The trades have probably been out of vogue for a generation,” Murray said. “Ironically, with AI, people are seeing the value of those skilled trades now … I do see a golden age for trades.”

That golden age remains some distance off. Research from the Ad Council and Ford Motor Co. found that only 9% of high school students are actively considering entering a trade.

Murray plainly wants to believe that technology will lift the physical workforce rather than discard it. His defense is heartfelt, rooted in community, and delivered with evident warmth. “I certainly hope,” he added, “those comments don’t age badly.”

This story was originally published on Fortune.com.