Devplan raises $2.5M to take on product integration work AI is leaving behind – GeekWire

Devplan, a Seattle startup that tries to eliminate the meetings and status reports that consume a product team’s week, is going private Thursday with $2.5 million in funding.

The company, founded by two industry veterans with experience at companies such as Uber, Amazon, Snap and Meta, is expanding the availability of its workflow software to all product and engineering teams, after quietly testing it with a small group of customers.

Devplan traces its roots to a youth soccer field, of all places, where founder Anton Safonov coached the daughter of his future business partner, Chris Bee. They had children in the same school, and they found that they were frustrated by managing engineering teams: most of the work had nothing to do with building a product.

The bee calls you an invisible tax.

“As AI fundamentally changes the way we do software development, we’ve seen this tremendous acceleration in coding and engineering work,” he said in an interview. “But the other coordination function, that whole tax – ‘work for work’ – hasn’t changed much, frankly.”

They founded Devplan in 2025, with Bee as CEO and Safonov as CTO.

How does this work: Devplan’s core product, Weaver, connects to commonly used tools including GitHub, Jira, Slack and meeting note takers. The user can ask Weaver through the built-in chat function, the Slack bot, or through a direct link to AI coding tools like Claude Code, asking about product status, features, or who is responsible for which features of the project, for example.

Weaver also works in the background, generating daily alerts for each individual and tracking projects on its own, flagging risks and progress without having to fill out an update.

The idea is to avoid scheduling a meeting or creating a status report.

Bee said queries run faster and cheaper with Weaver than pointing an AI tool at the raw data each time, because Devplan processes the information in advance and stores it in a knowledge graph rather than scanning code and documents throughout the application. He said queries are twice as fast and three times cheaper in token costs in internal testing.

Sponsorship: The company’s $2.5 million seed round was led by AI2 Incubator, with participation from Acequia Capital, Mighty Capital, Grand Ventures and eLab Ventures.

Based at AI House at Seattle’s Pier 70, Devplan employs six people, with a seventh hire in the works. The seed money goes toward hiring developers and deep integration, Bee said.

Founders’ domains: Bee was previously CTO of Lessen, where he helped grow the real estate services company from a $20 million startup to a $2 billion valuation, and previously led product and engineering teams at Zillow, Uber and Amazon.

Safonov spent seven years as a principal software engineer at Snap, where he was a lead engineer on the company’s infrastructure team, and previously worked on the Meta systems that handled real-time traffic for Messenger, Facebook and Instagram. You also create programs on LinkedIn.

“Chris and Anton have lived this problem at scale, and have the technical depth to solve it,” Yifan Zhang, managing director at AI2 Incubator, said in a statement.

Competitive landscape: Devplan’s bet is that a tool designed specifically for product and engineering teams will outperform general-purpose AI assistants.

Glean, the AI ​​search firm, might be the closest comparison, Bee said, but it works more broadly across the company’s experience while Devplan goes deeper into software development. Devplan also integrates with Linear, a project management tool, making it more of a partner than a competitor, he said.

Increasingly, given the capabilities of AI coding assistants, the competition includes companies deciding to build a coding tool in-house rather than buying one.

Current status: Devplan has many business customers who pay on annual contracts, and hundreds of users have tried it so far. Pricing is based on usage, with companies quoting a lower price based on group size and expected usage.

What’s next: The company is currently focusing on business customers, with plans to eventually open up the product to the public on a paid basis.