Bolt.new, the AI app-building platform made by StackBlitz, said on September 29, 2026 that it had acquired Dokai, a San Francisco startup that built autonomous AI agents for enterprise sales and marketing workflows. It is Bolt.new's first acquisition. Terms were not disclosed.
Dokai co-founder Gerry Fernando Patia and the Dokai team are joining Bolt.new's AI and machine learning organisation, and Dokai's agent technology becomes part of the Bolt.new platform.
Bolt.new said enterprise customers will see the first capabilities built on Dokai's technology in late autumn 2026.
What Dokai built
Patia and Rio Atmadja founded Dokai in San Francisco to explore how AI agents could automate complex, repetitive work inside enterprise software. It first focused on teaching AI to learn business workflows by watching demonstrations.
It later became what it called an AI go-to-market engineering platform for enterprise sales teams. Its agents carried out multi-step work such as prospect research, ranking and tiering accounts, enriching data and updating CRM records, turning instructions written in plain language into actions inside the systems companies already use. Dokai's orchestration layer was designed to let agents plan and complete that work without a person in the loop.
“Enterprise buyers don't care whether an agent demos well. They care whether it does the work every day, inside the systems they already run.”
How the two products fit
Bolt.new said both companies built autonomous, action-taking AI systems: Bolt.new lets anyone build software through conversation, while Dokai applied the same idea to enterprise sales. What it acquired, it said, is the team's experience building agents that act inside enterprise systems rather than simply making recommendations to employees.
Simons said the Dokai team had been doing that kind of daily work for customers, "and it's why we wanted them building with us." The acquisition also adds to the AI and machine learning team that builds the agents behind Bolt.new's platform, which customers use to create complete applications from text prompts.
Why Bolt.new is buying
Bolt.new lets users build, run and deploy full-stack web applications by describing them in natural language. It integrates with Microsoft Azure and Microsoft 365 for enterprise teams. Both companies, it said, were built on the idea that natural language can replace code as the main interface between people and complex workflows; the difference was the domain.
Bolt.new plans to use Dokai's orchestration work to improve how its own agents handle complex, longer-running builds and how the apps customers create connect to existing enterprise systems such as CRMs and internal data sources.
"Enterprise buyers don't care whether an agent demos well. They care whether it does the work every day, inside the systems they already run," said Eric Simons, Bolt.new's co-founder and chief executive. He said the Dokai team had been doing exactly that.
The founders and the funding
Dokai was backed by Character Capital and ERA, according to the announcement. Its funding amounts were not disclosed.
"We built Dokai on the belief that natural language could run production workflows inside enterprise systems. Joining Bolt.new gives that belief the biggest canvas there is: software itself," Patia said.
Dokai's agents turned natural-language instructions into work carried out directly across enterprise systems, from research on prospects to CRM updates. Bolt.new did not say how many people joined from Dokai, what role each founder will take, or whether Dokai's existing sales product will continue to be offered to customers on its own.
Bolt.new's enterprise push
Bolt.new is pushing beyond individual users into large company contracts, and it said that enterprise business is growing: roughly half of its enterprise customers start building before their contract is signed, and the average customer deploys its first app within 34 days.
Buying a team with experience of agents operating inside corporate systems is a way to address one of the main concerns of enterprise buyers, which is whether AI-built software can work reliably with the data and tools they already run.
