The scene, May 2026
It is performance-review season, and somebody is staring at a spreadsheet.
Somewhere in the People Operations wing of a fast-growing company - the kind with a kombucha tap and an HRIS no one fully understands - a compensation lead is preparing to recommend raises for 3,400 employees. Five years ago, this would have been a doomed exercise in cell formulas and stale survey PDFs. Today, she opens a browser tab. The tab is Pave.
The market data on her screen was pulled this morning, not last quarter. The merit budget knows which engineers are out of band. A panel labeled Paige answers questions in plain English about senior staff comp in New York. The review will be done by lunch. She will not have to call her contacts at three other startups to ask what they pay.
This is what Pave built. This is also what it broke.
The problem they saw
Pay is the largest line item in any business, and almost nobody manages it like one.
Until very recently, the way companies decided salaries was, to put it generously, vibes-based. HR bought a survey from a consulting firm. The survey had been compiled six months earlier. The data was sliced into categories that did not quite match anyone's actual job. Then a finance team merged it with three spreadsheets and emailed the result to managers who promptly ignored it.
The cost of all this was not just inconvenience. It was rotation. It was lawsuits. It was the slow, expensive erosion that happens when your best engineer learns - usually from a recruiter, often on a Tuesday - that the company down the street pays 22% more.
Matt Schulman watched this happen at Facebook, where he wrote software for a living and noticed that the people writing the software had no idea what they were worth. He did the unfashionable thing: he quit a comfortable job at a famous company to build software for a department most engineers had never met.
The founders' bet
Real-time data, plumbed directly from the HRIS, would beat any survey ever sold.
The bet was almost embarrassingly simple. If you could connect to Workday, BambooHR, ADP, Rippling, and the dozen other systems that already held the world's actual pay data, you would not need to ask companies to fill out a survey. You could just read it. Sliced cleanly, anonymized, refreshed continuously.
Investors agreed - eventually. Y Combinator took an early bet in 2020. Andreessen Horowitz wrote the seed and the A. Bessemer joined the B. By June 2022, Index Ventures led a $100 million Series C at a $1.6 billion valuation, and Pave joined the unicorn club roughly 18 months after it had finally secured the pave.com domain name. The order of operations was very 2022.
In the same week as the C, Pave acquired Advanced-HR from Morgan Stanley - a quiet but enormous deal that brought the legacy OptionImpact and OptionDriver datasets under one roof. The bank, in the end, sold the data to the startup.
Milestone reel
The product
One platform, four jobs, eventually one agent.
Pave is what software people call a workflow product wrapped around a data product. The data product is the benchmarking engine: live market pay, pulled from thousands of payroll integrations and segmented by job family, geography, level, and company stage. The workflow products sit on top.
Benchmarking
Live market data sourced directly from HRIS integrations. No more six-month-old survey PDFs.
Planning
End-to-end merit cycles, promotions, and adjustments without an offline spreadsheet phase.
Total Rewards
Branded employee portals that explain pay, equity, bonus, and benefits - so the recruiter on LinkedIn loses the information advantage.
Paige (AI agent)
Ask in English. Get the chart. The compensation analyst's new junior associate.
The proof, in numbers
Who runs comp cycles on Pave
The mission
Build the world's compensation infrastructure.
The phrase is a little grandiose, which Pave's marketing team will admit on the right day, but it is also accurate. Compensation is plumbing. It is invisible when it works. It is the entire story when it does not. The argument the company makes - to investors, to customers, to anyone who will sit through the slide deck - is that pay decisions should run on infrastructure, not on instinct, and that the infrastructure should be continuously updated rather than annually patched.
There is a moral case threaded through this, too, though Pave is too polite to lead with it. Live data closes pay gaps faster than HR policies do. When a manager can see, in real time, that a high-performing employee is 14% under market, the conversation gets easier. When she cannot, it does not happen at all.
Why it matters tomorrow
Because the next compensation cycle is always closer than you think.
The interesting question is not whether Pave wins comp software. By most measures it already has. The interesting question is what happens when every meaningful pay decision in the United States moves into a single category of tooling, and that tooling starts talking back. Paige is the first version of that. Future versions are easier to imagine than to build.
The skeptic's case is straightforward and worth airing: comp is sensitive, regulated, and political. Real-time data invites real-time arguments. Companies may decide they prefer the ambiguity. A few will. Most, judging by the customer roster, will not.
Back to the scene
She closes the tab. The raises are submitted.
Nobody walks past her desk to ask why review cycle took three months this year. It did not. The kombucha tap is still leaking; that is a separate problem. But the spreadsheet on her old laptop - the one with the merged cells, the conditional formatting, the column nobody had touched since 2021 - is closed for good.
Pave did not make compensation simple. Compensation is not simple. It made it a system. The difference, if you are the one making the decision, is the difference between guessing and knowing. Most people, given the choice, will choose knowing.