Why Botkeeper went out of business

AI bookkeeping software company Botkeeper is no longer. Over the weekend of February 7–8, in a post on its home page, the company announced that it was going out of business.

A well-known name in accounting circles, Botkeeper, founded in 2015, secured almost $90 million in funding from investors including Gradient Ventures, formerly part of Google. It hosted a conference called AI Unchained, was twice named to Inc. 5000’s Fastest Growing Companies lists, and boasted among its hundreds of clients Withum, the nation’s 22nd-largest accounting firm.

And it was known for innovation. Botkeeper “blazed a lot of new trails,” David Cieslak, a CPA and chief cloud officer and EVP at RKL eSolutions, and a frequent speaker on accounting software, told CFO Brew. Byron Patrick, a CPA and senior project manager at Karbon, and a Botkeeper employee from 2019–22, agreed. “Botkeeper, I think, was one of the earliest AI-focused tools in the profession,” Patrick told CFO Brew.

What went wrong? In a statement to the Botkeeper community, CEO and founder Enrico Palmerino blamed the company’s downfall on a “‘perfect storm’ of macro-economic shifts that arrived more swiftly than we could course-correct.” Toward the end of 2025, he said, it faced “a series of unexpected industry consolidation that significantly impacted our largest clients.”

The accounting profession has certainly seen its share of consolidation in recent years. But for Botkeeper, the cracks may have been showing before the late-2025 period. “There were a lot of signs in the last couple years that things were slowing down for them,” such as layoffs and reduced visibility in the marketplace, Patrick said.

The market for accounting software is also “astronomically different” from when Botkeeper launched, Patrick added. AI, especially the type of AI that Botkeeper offered, is no longer a differentiator. There are “new product announcements every single day on the AI front,” Cieslak pointed out. Incumbents like Sage Intacct and QuickBooks have added AI capabilities, while “AI-native” startups like Campfire, Rillet, Accrual, Maxima, Light, Pennylane, and Pigment are bringing in millions in venture capital, for valuations in the billions.

By the end of 2025, Palmerino wrote in his statement, Botkeeper had become an “AI powerhouse” that was “capable of cleaning up years of messy data in minutes, autonomously reconciling accounts, and coding 80%+ of transactions with a staggering 98% accuracy.” New features were on the way, including automatic check scanning and a voice-activated assistant.

But it may have been too little, too late for a company best known for its innovations in an older generation of AI: machine learning and automation. “Using pure machine learning in an accounting environment is just not as complete of a solution as we now have with generative AI,” Patrick said. “Compared to the AI that we’re using today, that was kind of a butter knife to a steak knife.”

Lessons learned: Botkeeper’s closure offers lessons for both software startups and buyers. Don’t confuse funding with functionality, Patrick said. Though getting venture capital money is exciting, “that is not an indicator of a workable product. That just means somebody is betting on—gambling, frankly—on those horses that they’re going to win the race,” he said.

Smaller niche players may run into trouble in the market if they are “tightly focused on a particular use case” that a larger competitor could easily add to a core solution, Cieslak said.

Buyers should know the signs of companies that are trying to go to market with software that’s not quite ready, Patrick said, such as giving buyers limited ability to try out tools during sales conversations or pressuring them to commit to contracts too early.

Accounting software buyers also need to have a solid platform as their core, Cieslak suggested, and shouldn’t focus on the solution so much as the use case for it. Such focus can help an organization avoid putting together a “frankensolution” composed of too many different products.

“You want to be careful you’re not just creating something that’s got so many moving parts that it becomes fragile,” he said.