Cutoff Culture with Theo Lindqvist

At 3:40 on a Thursday afternoon, with twenty minutes left before the bank’s same-day wire cutoff, Theo Lindqvist stood behind Mateo Villaseñor’s chair and read a number out loud. It was $1,184,250, the total for a syndicate closing into a seed round in Austin. The agent had built the wire, matched it to 212 investor commitments, and flagged one line: a founder’s updated wiring instructions had arrived by email that morning, and the account number did not match the one on file. Theo did not look at the agent’s reasoning first. He picked up the phone.

The founder’s CFO answered on the second ring. The instructions were real; the company had changed banks the week before. Mateo released the wire at 3:52. Theo went back to his desk and wrote one line in the team channel: “Call-back done, verbal match, released.” Nobody replied, because nobody needed to. The same thing happens, in some form, about four times a week.

I spent two weeks with Theo’s team in San Francisco and on video with the half that works from New York. I came to see how a fund operations group handles the volume of a venture platform with a team that fits around two tables. I left with a clearer picture of something else: a team that has decided, line by line, which mistakes it can afford and which it cannot.

Fund operations at Brookvale is the work behind a deal that an investor never sees. A lead posts a syndicate. Investors commit. Each one must pass KYC, sign the subscription documents, and send money. The vehicle then wires into the startup on the closing date, records the position, and reports on it for years. Rolling funds call capital each quarter. Roll-up vehicles pool many small checks into one line on a cap table. In March, every one of those vehicles needs a K-1.

Theo describes the job as a calendar with hard edges. Bank cutoffs, closing dates, and IRS deadlines do not move for anyone. The team’s whole design starts from that. “Most ops teams can be late and apologize,” he told me. “If we are late, a founder misses payroll or a lead misses an allocation. There is no apology that fixes that.”

The team is eleven people. Four handle capital movements: calls, wires, and distributions. Three handle investor onboarding and KYC. Two handle tax and reporting, and that group grows to five with contract help from January to April. Theo and one operations engineer, Ines, write and maintain the agent workflows. Last quarter the group closed 1,630 vehicles and moved $412 million in 5,900 outbound wires.

When Theo took over the team, each person had a private tracker. The capital calls lived in one spreadsheet, the KYC exceptions in another, and the tax amendments in an inbox folder that only one person could find. His first rule was that no work could live anywhere a teammate could not see it. His second rule, a year later, was that no work could live anywhere an agent could not read it.

Today every vehicle has one record that shows its state: committed, funded, closed, reported. Agents read that record, do the next step, and write back what they did and why. When an agent cannot finish, the vehicle goes to one of three queues, one for each group. People work the queues. They do not work the vehicles that are moving normally, and they do not look at them unless a weekly sample asks them to.

The Monday queue review in San Francisco. Each column on the board is one of the three exception queues.

The queues are small on purpose. On a normal weekday, the capital movements queue holds about 30 items out of 600 vehicles in motion. The onboarding queue holds 45 to 60. Theo watches one number above the others: how long the oldest item has waited. If anything sits for more than one business day, it is the first topic at the 9:15 standup, with a name next to it.

The change was not free. For the first two months, the agents sent far too much to the queues, because Ines had set every threshold low on purpose. The capital movements queue peaked at 140 items on a Tuesday in May, and Mateo’s group worked late three nights in a row. Theo did not raise the thresholds to make the pain stop. He had the team tag every item that a person cleared with no change, and each Friday Ines raised the threshold only for the tags with more than 50 clean results. By August, the queue was down to its current size, and each threshold had a record of why it sat where it did.

Hana Kurosawa runs investor onboarding. Her agents collect documents, check them against the accreditation rules, screen names against sanctions lists, and chase the investors who stopped halfway through. Two years ago, a new investor took six days on average to become ready to fund. Today the median is nineteen hours, and 81 percent of investors finish without a person on the team ever opening their file.

The other 19 percent are the hard ones, and Hana says they should be. A trust with three layers of ownership. An investor who shares a name with a person on a sanctions list. A letter from an accountant that says “accredited” but does not state the income test. The agent writes a short memo for each one: what it found, which rule applies, and what it would do. The memo is a draft. A person signs the decision.

Hana measures her agents the way she measures a new analyst. Each month she pulls 150 files the agents cleared with no review and has two people check them again from scratch. In the last six months, the re-check found four errors, all in how a document was labeled, none in a decision. When the agents disagree with a person, she reads the case herself. About a third of the time, she told me, the agent was right.

The scene at 3:40 shows the one rule the team will not change. An agent prepares every outbound wire, reconciles it against the commitments, and checks the instructions against the history of that recipient. Then two people approve it, and if the instructions changed, one of them calls a phone number the team already had, not the number in the email. Theo does not plan to change this rule, even as the agents get better.

His reason is not that the agents make mistakes. In fourteen months, the agents have built more than 22,000 wires, and the second reviewer has caught eleven real errors before release, most of them a wrong closing amount from a late change to the round. His reason is that wire fraud is a contest with a person on the other side. The attacker writes the email to fool whoever reads it. “A call to a number we trust is the one check they cannot write their way past,” he said.

The team has a short list of other work that agents draft but people own: any distribution above $250,000, any change to an investor’s bank details, any amended K-1, and any message that tells an investor their money is late or lost. Theo keeps the list on one page and reviews it each quarter. In two years it has grown by one item and lost none.

Adaeze Okonkwo leads tax operations, and she plans K-1 season the way a product team plans a launch. From October to January, her agents collect the inputs for every vehicle: final cap tables, fees, expenses, and the transactions that closed during the year. They compare each one to the year before and flag what changed. By the time the tax preparers start their work, the team has already found most of the surprises.

This year, the team issued 148,000 K-1s, 9 percent more than the year before, with the same five people at the peak. The share that went out by the target date rose from 71 percent to 93 percent. The amendment rate, the number Adaeze cares about most, dropped from 2.8 percent to 1.1 percent. Each amendment is a message to an investor who already filed, so she treats each one as a failure to study.

The busiest part of the season is not the forms. It is the questions. Between February and April, the support inbox for tax gets about 9,000 messages. Agents now answer the common ones directly: where is my K-1, why does this vehicle have a loss, when will the late ones go out. They send the rest to a person with the vehicle record and a draft already attached. Adaeze reads a sample of 40 replies each week, and she rewrites the templates when a reply sounds like a form.

Theo’s hiring has changed more than any workflow. He used to hire for throughput: people who could process a high volume of documents with few errors. Now the agents do the volume. He hires for the exceptions, which means people who can read an agent’s memo, find the weak step in its reasoning, and make a decision they can defend to an auditor.

His interview reflects this. A candidate gets ten real cases with the investor details removed, each with the agent’s memo. Some memos are right. Two are wrong in ways that are easy to miss: a trust document from the wrong year, and a wire amount that matches the commitments but not the final round size. The candidate has forty minutes. Theo does not score how many cases they finish. He scores whether they find the two bad memos and say why.

He also hires fewer people. The team has grown from nine to eleven while the volume of vehicles more than doubled. The two new roles were Ines, the operations engineer who maintains the agent workflows, and a second capital movements analyst, added because the team wanted three people who could approve a wire on any day, including holidays.

New hires do not touch a live queue in their first two weeks. They work last quarter’s exceptions, with the real decisions hidden, and compare their calls with what the team did. Theo says the gap between the two lists is the best training material he has. A new analyst who disagrees with a past decision must write down why, and twice that note has changed a rule.

On my last afternoon, I asked Theo what he would build next if he had another quarter and no constraints. I expected an answer about tax, or a new type of vehicle. He talked about the calendar again. He wants every vehicle record to show, at all times, the next hard deadline it faces and how many hours remain. He wants the queues sorted by that number, not by when an item arrived.

It is a small change on paper, and Ines has already started it. To Theo, it is the whole idea of the team in one screen. The agents make the work fast. The people make it safe. The calendar decides what matters today. “An error, we can fix,” he said. “A missed close, a missed cutoff, a late K-1 – those are the ones we cannot take back.”

At 3:55 that day, the board showed seven wires left before the cutoff, all approved, all with two names. Mateo released the batch. Theo watched the count go to zero, closed his laptop, and went to find coffee. The queue still had 26 items in it. None of them was older than four hours.