Finance software has been stuck in a model for twenty years. A ledger over here, a reporting tool over there, a planning tool bolted on top, and — when the seams show — a spreadsheet taped across all of them. The pattern was born in an era when the internet was optional and AI was a research curiosity.
What the stack looks like today
Walk into any mid-market finance team. You'll find QuickBooks or NetSuite running the books, a model in Excel or Mosaic layered on top, Gusto or Rippling handling payroll, a tax firm invoicing quarterly, and Ramp or Brex riding on the edges. Every one of those systems has its own data model. Every seam is a reconciliation risk. Every new question requires somebody to join the data by hand.
The response from vendors has been more tools, not fewer. New AI features layered on old data models. New "all-in-one" platforms that turn out to be loosely federated acquisitions. None of it collapses the fundamental fragmentation — because none of them were built to.
The pattern that always wins
Every category of business software eventually gets rebuilt from the ground up by someone who refuses to accept the fragmentation. Sales had Salesforce, then the modern AI-native challengers. Design had Photoshop, then Figma. Engineering had Jira, then Linear. The rebuilds always come from operators who understood the domain deeply enough to design for what it should be, not what it was.
Finance is overdue.
What an operating system for finance looks like
One data model. One authentication. One source of truth. Products specialized by job-to-be-done — planning, ledger, ops, payroll, tax — but stitched into one fabric by a shared intelligence layer that reads across all of them.
The difference shows up in the questions you can ask. "Why did gross margin drop in Q2?" Today, that's two analyst-days — pulling data from three systems, joining in a spreadsheet, interpreting, writing a memo. In a real operating system, it's thirty seconds, with the exact transactions and customers driving the movement cited inline.
That compression is the unlock. When the operating system handles the plumbing, finance teams get to do the work they were trained for.
Why now
Two things shifted. AI became capable enough to do the work — not summarize it, not retrieve it, but do it. And the fragmented vendor landscape finally broke past the point where buyers would rather consolidate than add another tool.
The next decade of finance software is going to be defined by the company that takes this seriously. We're betting it's us.