Remote, the seven-year-old Amsterdam payroll provider, says it crossed $300 million in annual recurring revenue and turned cash-flow positive. But according to TechCrunch AI, the headline number isn’t the ARR. It’s this: Remote grew revenue per employee by 50% after wiring AI into every level of the company, without adding headcount to get there.
CEO Job van der Voort told TechCrunch he runs five Claude instances on his second screen at any given moment, building tools for himself and for the company. That’s not a one-off founder flex. It’s the operating model.
What actually changed
The efficiency gains came from spreading AI past the CEO’s office and the engineering team. Employees across functions build their own apps in Remote Labs, an internal marketplace running on the company’s own tech. A Slack agent summarizes discussions. The repetitive, bureaucratic grind of paying workers in nearly every country got automated further.
The coding numbers stand out most. Van der Voort says engineering contributions rose more than 60% over the past year, and over the last month, more than 85% of Remote’s code was written by AI.
A few claims deserve a flag. Van der Voort says the core payroll business grew more than 300% year over year and credits AI for much of it, but TechCrunch notes the company hasn’t offered independent verification. The ARR milestone and the “tens of thousands of companies” customer count also come from Remote itself.
The hiring story is the real signal
What stands out here isn’t faster shipping. It’s that Remote used AI to defer hiring without cutting jobs. Van der Voort says some departments planned to hire more people than they ended up needing. The company is now actively asking a question more leaders will face soon: do we add people, or do we upskill the team we have and spend that money on AI instead?
This is significant because it’s one of the cleaner public data points on AI’s business impact. Plenty of companies claim productivity wins. Remote is putting a structural change on the table: more revenue per employee, slowed hiring plans, and a wider product surface, all at once.
Betting against the all-in-one playbook
Most of Remote’s competitors went the other way and built sprawling all-in-one HR platforms. Remote stayed narrow, owning the genuinely hard problem of global payroll and compliance. Van der Voort frames the AI wave and the commoditization of software as proof that focus was the right call.
That focus shows up in two new products:
- Remote Build puts what investors call “forward-deployed engineers” directly alongside customers, helping them build the same kind of custom AI workflows Remote built for itself.
- Remote MCP is an interface built on the Model Context Protocol, the emerging standard that lets AI agents securely talk to outside software. It gives agents and platforms like BambooHR and Workday direct access to payroll and compliance data, letting them run Remote as the engine underneath.
Why this matters for the industry
The MCP move points at where things are heading. “If you use ChatGPT or Claude, you can control all of Remote,” van der Voort says. “You don’t have to interact with our platform anymore.” In other words, the app layer fades and the company becomes infrastructure that agents call on demand.
He’s been testing this himself with an open-source personal agent named Jim, built so it can access what it needs but can’t do anything destructive with sensitive payroll data. That security boundary is the whole game when agents start touching financial and personal records.
Van der Voort isn’t worried about climbing AI bills. Spend is rising, he says, but the company tracks it, and growing efficiency leaves room to reinvest in exactly these initiatives.
The takeaway for practitioners: AI adoption that stays locked in engineering leaves most of the value on the table. Remote’s gains came from putting build tools in the hands of every function, then exposing its core data to agents through an open standard. Expect more focused, infrastructure-first companies to follow this path.
Full details are available at the original source, TechCrunch AI.