Why HFS Research Named Hazel a Hot Tech Company in the Services-as-Software Category

David Cushman runs emerging technology coverage at HFS Research, and he is one of the people who put a name to Services-as-Software, the shift HFS now sizes as a $1.5 trillion market. When HFS added Hazel to its Hot Tech list, the call came from months of his own diligence. Recently, in the heart of London, he sat down with our CEO, Sneha Shah, to say why.

What Services-as-Software means, and where Hazel fits

Services-as-Software has been read as software replacing services firms. It is services firms becoming software companies.

David puts the mechanism in plain terms. The old model is labor arbitrage, work priced by the hour and the headcount. The new one is technology arbitrage, the same work carried by software the firm owns. A lot of the work once done by people, he says, is now done by AI-led software.

That is the correction the market keeps missing, and it is the one HFS put behind Hazel.

The problem Hazel attacks: 18 months and a million dollars

Enterprise modernization has been slow and costly for a plain reason. It is human-labour heavy. Teams of analysts go in to capture how a business actually works, and the delivery that follows can run for a year or more.

HFS framed the baseline in its note: modernization that runs 9 to 18 months and costs $250,000 to over $1 million. The figure HFS recorded for Hazel is about ten weeks.

David's own list of what gets in the way is worth hearing: technology debt, process debt, talent debt, and data debt. Agents cannot function until a firm's context is captured and made usable. That capture is the first step, and it is the one Hazel is built to do.

The cost of waiting is real too. Firms that stall in pilot purgatory, David notes, watch the rest of the market move while their own AI spend climbs. The way out is a delivery system that proves itself on work already in motion, where there is a real baseline to beat.

Why David picked Hazel: control the enterprise can trust

The reason David singled Hazel out is control.

Most agent programs stall where the enterprise cares most, on governance, auditability, and security. His framing is processes that keep a human at the helm, with the controls an enterprise actually requires. "They've managed to bring in the kind of controls required to make it trustworthy in the enterprise."

Control is what makes autonomy usable inside a regulated program. The agents' output has to be predictable, reviewable, and auditable, with the firm's own approval gates intact and cost and security under the firm's hand. That is the layer Hazel builds around the delivery system, and it is what lets a firm run this on a program its name is on.

David's close says it best: "Being born in AI, being born in agents, being born thinking Services-as-Software, that does give an advantage to firms like Hazel."

What it does to the P&L: revenue that scales past headcount

David has a plain test for whether a firm is really making this shift: revenue per employee.

The deepest shift is not speed. It is the shape of the business.

When delivery becomes software, output stops scaling with people. Revenue decouples from headcount, and the firm starts to look like a software company: higher margin, revenue per employee on the way up, and a method that accrues to the balance sheet instead of leaving with attrition. That is what becoming a software company actually means in the P&L, and it is the prize behind the Hot Tech call.

What the designation holds us to

A Hot Tech call from the analyst who created the category is a benchmark. We clear it in front of the person who set it.

Watch the full conversation above, and read HFS's Hot Tech note on Hazel.