[CHECK: GEORDIE — quick read on the opening. I led with the conversation as you described it. Tweak any wording that is not how you would say it.]
I was on a call with a founder I work with, and he was walking me through the jobs his team dreads. He got to one I had barely heard of: K-1s. I had to ask him what they even were.
What stuck with me was not the form. It was the workflow around it. The other side, the people who produce these documents, were handing over giant files of raw data for his team to go through and clean by hand. My first question was the obvious one. Why? Why is the group generating this not automating it before it ever reaches you? Why is a person on your side reading rows in a file at all?
There was no real answer. It was just the way it had always been done. That was the problem I already knew about, and the one I knew I would end up fixing.
And he said it first. Before I pitched anything, the founder himself said this was the perfect place for AI to come in and take over. When the customer is the one pointing at the fire, you do not need a sales deck. You need a build plan.
So let me stay on the what and the why, and leave the how for the work.
What a K-1 is, in plain terms: a tax document that reports each partner's share of an entity's income. If a business has a lot of partners, that is a lot of near-identical documents, each one carrying numbers that have to be pulled out and reconciled against a master record. Multiply that across entities and you have hundreds of them, every one needing a human to read it, extract the figures, and check them.
Why it is so painful: it is high-volume, deadline-bound, and unforgiving. One transposed number matters. It is exactly the kind of work that burns out good finance people, because it demands total accuracy on a task that is pure repetition. Nobody grows a career reading row 47 of a data file. They just have to get it right, hundreds of times, under time pressure.
That combination, high volume, high stakes, zero judgement once the rules are set, is the cleanest case there is for automation. Same idea as the bank reconciliation I took live last week: read, match, draft, route the exceptions to a human. Different document, same method. It plugs straight into the AI controller we are building, and it is the next set of manual hours we take off the table.
The how comes later, in the work. For now the point is the one the founder handed me himself. The most hated job in the building is the first place AI should go. That build runs most of the rest of the year. If you want to know why I do any of this, it is in the fund I started for my son.
Monthly Revenues $9,200 | Clients 2 | Prospects (Meta to WhatsApp campaign built, WABA pending) | Team: Me + Jan (CTO)
Day 73 of 365.