Remote crossed $300 million in annual recurring revenue. It's cash-flow positive now. And the headline stat that makes you lean in: revenue per employee jumped 50% after the company adopted AI across every function — without adding a single headcount.
Let's be honest about where these numbers come from. Remote's CEO Job van der Voort reported them himself in a TechCrunch interview, and the company hasn't provided independent verification for some of its more aggressive claims — like a 300% year-over-year growth rate in core payroll. That's worth flagging, because self-reported milestones from startups are the wild west of financial journalism.
But even with that caveat, the underlying pattern is hard to dismiss. Remote is a seven-year-old Amsterdam-based payroll service provider that serves tens of thousands of companies navigating global employment compliance. It's generating more revenue per person on the payroll, and it's doing it without layoffs or even the hiring sprees most companies in its position would be running.
That's not a cost-cutting story. It's something more interesting — and more unsettling for anyone who still thinks AI's main business use case is making existing workflows slightly faster.
AI Isn't Just for Engineering Anymore
Here's what most companies get wrong about AI adoption: they treat it like an engineering tool. You bring in the LLMs, you let your developers write code faster, you call it a day.
Remote didn't do that. Van der Voort told TechCrunch he runs five different Claude instances simultaneously on his second monitor — some for himself, most for Remote. There's a Slack agent that summarizes discussions across the company. But the real signal is what happened next: employees across all functions started launching apps in Remote Labs, an internal marketplace built on the company's own technology.
That's the part that separates Remote from the typical AI adoption story. This isn't a pilot program where engineering shows HR what's possible and everyone claps politely. This is people in non-technical roles building real products on infrastructure that the company already owns.
Remote then did something most companies wouldn't think to do: it opened that same capability up to customers through Remote Build. Think of it as forward-deployed engineers, but instead of helping you set up a data warehouse, they're helping your teams build custom AI workflows inside their own organizations. Van der Voort said the company considers itself ahead of most in this regard, which is either confidence or a polite way of saying competitors are behind.
The pattern here matters because it suggests AI adoption at scale isn't about giving tools to the people who already know how to use them. It's about creating platforms where anyone can build something useful.
The Coding Revolution Nobody Saw Coming
Remote's engineering team has seen contributions rise more than 60% year over year. In the most recent month, more than 85% of all code written at the company was generated by AI.
Let that sink in for a second. Eighty-five percent. That's not "AI assisted." That's AI-written, with humans reviewing and directing.
The hiring impact is real but nuanced. Remote deferred some hiring plans — it wasn't running a big recruitment campaign anyway — and hasn't cut a single job. Van der Voort said the company is actively evaluating whether to hire more people or spend more time upskilling existing staff and investing directly in AI tools. That's a genuinely interesting decision point, because most companies haven't reached it yet.
Spotify has done similar things with AI-powered coding, and Remote is clearly in that same camp. But the payroll context makes it different. Payroll isn't a consumer app where you can ship fast and break things. It's financial infrastructure — sensitive personal data, compliance requirements across dozens of jurisdictions, real money moving in and out of accounts. Doing 85% AI-generated code in that environment says something about both the maturity of these tools and the confidence of a company that's willing to bet on them.
Van der Voort noted that AI has made the repetitive, bureaucratic work of paying workers in almost every country "easier, and arguably more fun than ever before." There's something almost poetic about a payroll CEO saying the boring stuff is now fun because an AI does it.
Remote as Infrastructure: The MCP Play
Here's where the story gets genuinely forward-looking. Remote recently launched what it calls "Remote MCP" — an interface built on the Model Context Protocol, a standard that lets AI agents securely interact with external software.
What does this mean in practice? If you use ChatGPT or Claude, you can control all of Remote through those agents. You don't have to interact with Remote's platform directly anymore, according to Van der Voort. Platforms like BambooHR and Workday are already using Remote as an underlying engine for payroll and compliance data.
"I think that's where the future goes," Van der Voort said. And he's not wrong, though it does sound almost too clean. The idea that your AI agent handles all the payroll complexity while you never touch a UI is compelling — but it also means Remote is positioning itself to become invisible infrastructure. That's a bold move for a company whose brand name suggests it wants you to know exactly who's paying your workers.
The security angle is critical here. Payroll data is deeply sensitive — financial information, personal details, tax records across multiple countries. Van der Voort described his own experience building a secure interface for his personal AI agent, named Jim (it runs on OpenClaw, an open-source project). The design philosophy: give the agent access to what it needs, but prevent destructive actions. "He cannot do crazy stuff and mess things up," Van der Voort said.
That's the kind of constraint that makes or breaks agentic AI in enterprise contexts. You can let an AI agent do anything, or you can give it a clearly bounded set of actions with hard guardrails. Remote is choosing the latter, and that choice will probably matter more than any feature launch.
The Bigger Picture: AI Scaling Without Headcount
Remote's trajectory is one of the cleaner data points we have on AI's real business impact. Not because it's a perfect case study — self-reported numbers, single primary source, the usual caveats — but because it demonstrates a specific pattern that's hard to find in isolation.
More revenue per employee. Deferred hiring without layoffs. An expanding product surface area without proportional headcount growth. That's the operating model many companies are chasing, and Remote is one of the few that seems to be actually building it.
The company's competitors largely went a different direction, adopting an "all-in-one" HR platform model. Remote stayed focused on the hard payroll problem. Van der Voort sees the current AI wave and the subsequent commoditization of software as validation for that decision. Whether he's right remains to be seen, but there's something to be said for a company that built deep expertise in one area and then used AI to make that expertise infinitely more distributable.
AI spend is increasing at Remote, but Van der Voort said it's manageable — efficiency gains provide room to invest in AI initiatives without breaking the bank. The company tracks it carefully, he said, and the trajectory feels sustainable.
This isn't a story about AI replacing workers. It's about AI changing the relationship between output and headcount, and that distinction matters more than most people realize. Remote isn't firing its way to profitability. It's growing revenue per person, deferring hiring where it makes sense, and investing in tools that make existing people more capable. That's a different playbook than the one most companies are writing — including Oracle, which cut 21,000 jobs while pivoting to AI infrastructure, a stark contrast to Remote's approach of growing without reducing headcount.
What This Means for the Rest of Us
There's a tendency to treat stories like Remote's as either utopian or dystopian. AI will save us all, or AI will destroy the middle class. The reality is almost always more boring and more complicated.
Remote's CEO said AI has added "a whole new fun angle" to his role. He runs five Claude instances while he works — the same Anthropic models that recently shipped Claude Fable 5 with expanded guardrails for general use. His personal AI agent has a name — Jim, apparently — and it can interact with Remote's systems securely. The company's engineering team writes 85% of its code with AI assistance, and nobody lost their job because of it.
That's not a revolution. It's an evolution, and it's happening in a company that does payroll for everybody — not just remote workers, despite the name. Van der Voort was explicit about this: the vast majority of Remote's clients employ people in offices. The name is misleading, but the business isn't.
The question Remote's trajectory raises isn't whether AI will change businesses. It's how fast companies will adapt their operating models to match what AI actually enables, and whether they'll make the uncomfortable decisions about hiring, upskilling, and investment that come with that adaptation.
Remote seems to be making those decisions in real time. Whether other companies will follow is the story worth watching next.