Paul Koullick and the art of making things less taxing

Paul Koullick has a management success story that involves getting lunch. After introducing a library of document templates at Keeper, he noticed fewer meetings, shorter conversations and more useful preparation. The calendar loosened its grip. A meal became possible. In the literature of startup achievement, this is an unusually appetizing result.

He had spent years thinking about the work people do before they can do their work: reconstructing a business expense, deciphering a colleague’s comment, opening a blank document. Keeper, the tax-software company he co-founded, addressed one version of that problem. His writing about running teams addresses another. In both cases, the details are small enough to be overlooked until someone has to live with them every day.

Then, in 2025, Koullick made a larger adjustment. He handed Keeper’s CEO role to co-founder David Kang. A few months later, he announced he was joining Google Meet to accelerate its AI integration. The company he had built would carry on with a different chief executive. He would return to product work.

The tax problem that followed him

Before Keeper, there was a sequence of jobs that put Koullick close to the mechanics of money and software. He graduated from Harvard in 2014 with an applied mathematics degree and joined Square as a product analyst. At Stride, he went on to lead tax product, helping build and grow Stride Tax, an app for mileage and expense tracking.

His March 2017 guest essay for Uber explained an integration with its Driver API. The practical concern was what independent drivers could understand about their earnings and expenses. The essay described generating tax summaries from information already available through the two products. Even then, his subject was the distance between someone doing a job and someone having the records to make sense of it.

At Amplitude, where he worked in product before founding Keeper, he wrote about launch metrics, user research and documentation. It was a useful vantage point for someone interested in whether a product actually helped people. A release could ship; a separate question was what happened after people encountered it.

Keeper took him back to independent workers. Koullick and Kang founded the company in 2018. They were childhood friends and former debate partners. Their working relationship would eventually have to accommodate customers, investors and a CEO succession, giving them rather more to discuss than a tournament motion.

A company built around a smaller question

Keeper’s early product began with a familiar mismatch. Independent workers were running businesses in the eyes of the tax system, even when keeping accounts occupied very little of their attention. A purchase could matter later without looking especially important when it happened. Koullick wanted that connection to become easier to see.

The original service combined machine learning, human bookkeeping and text messages. Software examined transactions and flagged possible business expenses. Uncertain cases could go to a person for review; customers received questions that helped establish what a purchase had been for. The work was distributed across the system instead of arriving in one annual heap on the customer’s desk.

By August 2019, Keeper had passed 1,000 paying subscribers. The company was trying to make an obscure, delayed task feel immediate enough to engage with. Koullick’s explanation was straightforward: the technology itself was easier than explaining its value and creating an experience people wanted to use.

In April 2021, Keeper announced a $13 million Series A. The plan extended beyond finding expenses: the company wanted to build a complete tax platform, including advice and filing. Backers included e.ventures, Matrix Partners and Foundation Capital, alongside founders from other software businesses. The product ambition had grown from noticing a transaction to helping finish the job it belonged to.

Koullick’s own fundraising announcement described the appeal of serving people whose income came through gig and creator platforms. It also described what the capital was for: room to hire and make larger product bets. That distinction matters to his story. An idea about a useful interaction had become a business requiring a team, a system and decisions about where to spend.

By a 2023 interview, his description of Keeper included preparing returns from documents and transaction information, asking users questions along the way, and retaining human sign-off. He talked about tax as a language-heavy domain with extensive documentation. In an AI discussion easily consumed by abstractions, he kept returning to a concrete workflow: information comes in, questions get answered, a return gets prepared.

Lunch, rescued by a Google Doc

Inside a company, uncertainty has its own paperwork. Someone needs a decision. Someone else needs context. A message thread acquires the length and atmosphere of a minor parliamentary inquiry. Koullick had been arguing for better documents well before Keeper.

His 2018 Amplitude essay treated a good product document as a short investment that colleagues could read quickly and use. Formatting, brevity and comments mattered. Writing something down was only the beginning; other people had to engage with it. He made a particular point of leaders commenting, because silence gives a team little reason to keep producing useful documents.

“A good doc is written to be read.”

Paul Koullick, 2018

By April 2025, he had published Templates by Paul, a library drawn from routines he had used and expanded at Keeper. His diagnosis began with the blank page. Starting a document can feel harder than answering a message or scheduling a meeting, so the immediate conversation wins. A template reduces the effort of starting.

He described rolling the library out across Keeper and seeing documents spread beyond leaders and product managers. Meetings came with material to read beforehand. Long Slack exchanges became less common. His favorite hour of the day became time spent reading other people’s documents and leaving comments. This was his account of the improvement, and its specificity is part of the appeal: lunch, an hour of reading, a shorter meeting.

But comments bring a problem of their own. A passing observation from the CEO can sound expensive. Is it an instruction? A suggestion? Does it require another meeting? Koullick’s feedback tags give the reader a clue about the weight of the remark before anyone starts rearranging their afternoon.

His companion decision framework divides work into three approaches. Some decisions can be made and communicated afterward. Others need a proposal and comments. The hardest need the problem framed so other people can help. He has written that he underuses that third approach and is working on it. Even the framework leaves space for its author to improve.

A founder with a parody microphone

Koullick’s extracurricular commentary on entrepreneurship has a less administrative name: Bad Startup Advice. The podcast begins with deliberately dubious counsel, delivered with authority, before turning to a more serious discussion. Founder certainty supplies the joke.

The listed episodes include conversations with Kang about committing to a startup and becoming attached to an idea, plus a discussion with Kapwing founder Eric Lu. Calling an episode “Your idea is the only thing that matters” gives the audience fair warning about the premise. The show’s description explicitly takes aim at founder worship. Koullick has accumulated enough time in the occupation to know where the comedy lives.

There are quieter details, too. His Amplitude biography mentions chess and jogging along San Francisco’s Embarcadero. An Uber author biography records a taste for anti-jokes. These do not need to become a grand theory of his character. They give a little texture to a public working life otherwise populated by retention curves, documents and expense categories.

The handoff, and the next room

On May 6, 2025, Koullick announced his transition to Keeper’s president and board member, passing the CEO job to Kang. He reported seven years of building, more than $15 million raised, profitability and approximately eight-digit annual recurring revenue. These were his account of the business at the handoff.

He also described a change in how he understood leadership. Before becoming a CEO, he had been quick to criticize executives and organizational decisions. Doing the job had made him more sympathetic to the amount happening beyond an employee’s view. He expressed gratitude toward the people who had supported Keeper and confidence in Kang’s ability to lead its next stage.

His next public career announcement arrived in August: Google Meet. Between roles, he had spoken with 18 organizations. He wrote about the process with the candor of someone discovering that an executive résumé does not exempt its owner from interview practice. Running a company and demonstrating your usefulness to a new employer turned out to be separate activities.

The move returns him to product management, a strength he had identified when announcing the Keeper transition. At Google Meet, his stated aim was accelerating AI integration. The setting changes, but the practical questions remain recognizable: what information does someone need, what should the software handle, and how much effort remains for the person?

Across his published work, Koullick keeps examining the small obstacles between intention and action. A blank document. An ambiguous comment. A purchase whose purpose will be forgotten. There is plenty of ambition in that career. There is also a sustained interest in the ordinary afternoon, and whether somebody gets it back.