Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

IT now has three roles

The role no one's managing yet is the riskiest.

[Photo: Getty Images]

Profile photo of Carey Kolaja

Chief Executive Officer · Versapay

A few years ago, I became obsessed with a question: What if we treated the people inside our companies with the same care we reserve for our customers? I wrote about it at the time, the idea of “productizing” the company and putting employees first. Back in 2019, Gallup’s State of the Global Workplace report found that 85% of the workforce was disengaged, costing the global economy an estimated $7 trillion in lost productivity. I argued that the fix was to invert customers-first into employees-first. I believed it. But honestly, the tools to deliver it at any real scale didn’t exist yet.

I think they finally do. And the place I see it happening first is the one we used to think least about: IT.

HR FOR AGENTS

When Jensen Huang said the IT department of every company would become “the HR department of AI agents,” some people heard a provocation. I heard something true. We are starting to hire agents. And here’s the part we discuss less: You also have to fire them. Firing an agent is easy—far easier than letting a person go. The real risk is that in most companies, no one can answer the basic question: Who manages this agent?

Think about how many are already running around your company, spun up by well-meaning teams, doing real work, and owned by no one. HR for agents isn’t a metaphor. It’s a governance job we haven’t designed or staffed yet. Before hiring and firing, you need clarity on who’s accountable for each one, what good performance looks like, and signals for when you should deprecate. We’re still building that muscle, but a rule we’re moving toward is that no agent goes live without an owner.

Agents are half the story, and the other half is human.

PRODUCT TEAMS FOR EMPLOYEES

For decades we defined IT by what it issued—laptops, logins, devices—and what it closed: tickets. That model treated employees like petitioners at a counter, and it was never going to survive a workforce that, off the clock, summons a car, a mortgage, and a doctor in three taps. AI is what finally broke the counter. When everyone has a capable problem-solver in their pocket, the slow ticket queue is frustrating, but also it’s irrelevant. People route around it.

So we stop asking “how fast do we close tickets?” and start asking “how do we make the ticket unnecessary?” That’s a product question and answering it forces IT to behave like a product and engineering team instead of as a service desk.

Employees are the customers seeking solutions; the internal experience is the product. For us, that wasn’t a slogan. It changed how the team actually works. We started from user research instead of assumptions, asking who our employees are and where the friction is. We stopped managing a backlog of complaints and started working a roadmap, watching telemetry, adoption, time to resolution, and self-service rates, the way a product team watches its funnel. And the definition of success flipped: The best support experience is the one an employee never has to ask for.

We don’t publish this; we try to do it. My team recently came off an AI sprint where we inventoried everything we’re running—every tool, every workflow, every place someone was reaching for automation and intelligence. It was humbling in both directions. There was more underway than I’d have guessed, and far more of it was the same work repeated across departments. Once we stripped out the duplication, a long list collapsed into a much smaller set of recurring needs, with only a handful genuinely unique to a single function. Naming that was the whole unlock. A workforce is a community of personas and you can’t productize what you don’t understand intimately, or spot patterns you aren’t looking for.

GATEKEEPERS FOR INTELLIGENCE

There’s a third shift, maybe the most consequential. As agents multiply, what separates a brilliant one from a useless one is the data it can reach. IT is becoming the layer that governs how data flows and who—human or agent—can access it. The more context an agent can see, the smarter it gets, and the more productive the person standing behind it. Increasingly, IT isn’t issuing hardware, but intelligence.

This raises a question. In most companies, the people with the greatest information access are the most self-directed, curious, and the ones who see the most. What happens to productivity and fairness when we govern access as deliberately as we govern headcount? Too much control of access, which limits knowledge and timely discovery; too little control in hygiene, format standards, and review systems, which hinders consistent, quality results.

I don’t think IT is one thing anymore. It’s HR for the agents we hire and manage. It’s the product team for the employees it serves. And it’s the gatekeeper of the data that makes both smarter. None of that looks like issuing laptops.

I won’t pretend I have this figured out; we’re learning as we go. But I’m increasingly convinced the teams that treat their employees like customers, and their access like strategy, will pull ahead. AI exposed and made this thesis buildable.

Creating operational leverage in the AI era is the opposite of replacing people with agents. It’s about giving people back the hours we used to lose at the counter. And it’s holding IT to the same standard of craft, data, and delight we’d never compromise on for a paying customer.

Here’s my challenge to every leader still treating IT as a cost center that hands out hardware: Stop running a help desk. Start running a product team. Your employees are your most important users. Serve them the way you’d serve your best customer, because the companies that do will be the ones people want to be a part of and want to buy from.

Carey Kolaja is CEO of Versapay.