Every founder in our category has heard the same pitch, because most of us have given it. Connect the accounts. Sync the feeds. Render the dashboard. The work disappears.
The work did not disappear. We’ve spent this series explaining why — that the dashboard is a rendering on top of data somebody else still assembles by hand, that the structured database underneath is mostly empty, mostly manual, mostly held together by very expensive humans. This post is about those humans, and about the number nobody in wealth tech wants to say out loud.
For every dollar a firm spends on the software, it spends about six on the servicing of that software. The chart on the screen is the cheap part. The labor underneath it — the $100K analysts, the $500-an-hour CPAs, the family-office accountant logging into a GP portal at 11pm to pull a number that arrived as a PDF — is where the real money goes. The software was never the cost. The work was.
And the work is not going anywhere, because the hardest parts of a financial life are not data entry. They are judgment. Listen to the people who actually live in this problem. A family-office COO told us, plainly, that he had no clean way to tell his principals at year-end what their aggregated performance actually was. A founder who’d made dozens of investments through his own network put it like this: all of it is a mess, and there’s no one to help on the private side. An investor with a couple hundred K-1s landing every spring wasn’t worried about his dashboard at all — he was worried that if something happened to him, his family wouldn’t know where to begin. Another family office described their reality bluntly: the public positions pull through fine, the private ones pull through wrong, so they hand-key every alternative investment.
None of that is a feed problem. It is a judgment problem. Someone has to read the K-1 and decide the mark. Someone has to know which of the eleven entities actually owns the position. Someone has to look at a number that just changed and know whether it’s a correction or an error. Software renders. It takes whatever sits beneath it and makes it look finished. When the foundation is unverified, the clean dashboard isn’t clarity — it’s a confident-looking guess, and a wrong number that looks right is the one that gets used to make a real decision.
The fashionable answer in 2026 is that the AI agents will simply do all of this. We don’t believe that, and we’ve said so: AI alone is too sloppy and too indefensible to be trusted with a family’s balance sheet, and humans alone are too expensive and don’t scale — that’s the six-to-one. The only thing that works is the combination, engineered. Real domain experts and a structured framework of AI agents, with a logged handoff between them, producing a structured data layer rather than a one-time report.
That is why we productized the service instead of selling the software. We’ve given the model a name — SPaS, service presented as a software — because it is the thing that actually works in finance. You like your interface. You like that everything sits in one place, in the same hundred fields, consistent every time, so the family you manage money for stops calling you with a thousand questions. What you don’t like is filling it out. What you don’t like is the labor. So we took the labor.
Here is the part the market keeps mispricing. Our services are not a drag on the model that we apologize for on the way to a clean multiple. They are the model. The hard work is the moat. Anyone can ship a dashboard. Almost no one is willing to build — and keep staffed, and keep honest — the rigor that makes the dashboard true every single day.
The people who built the first wave of financial technology understand this in their bones. The clean systems we now take for granted ran, in the early days, because someone stayed late and reconciled by hand and caught the break before the customer ever saw it. The automation came after. It always comes after. The human does the hard part first, and then teaches the machine to do the rest.
That is, almost literally, how we built Abacus. Our AI agents learned the work by watching ex-investment-bankers, ex-private-equity professionals, and ex-family-office CIOs do it — in one case, onboarding more than five thousand alternative positions for a single family in the Midwest. The judgment came first. The automation is the residue of the judgment, and it’s what keeps each client’s Financial Digital Key correct and current every day, so they can permission it anywhere, keep their existing tools right, and get back to the part of the job they actually want to do.
The wealth firms that win the next decade will not be the ones that bought the best software, and they will not be the ones that fired the most people. They will be the ones that understood the six-to-one — that put the human exactly where judgment lives, the software everywhere else, and refused to pretend the labor wasn’t the whole game.
Finance doesn’t have a software problem. It never did.