Dataiku announced Cobuild, an AI building tool designed to turn plain-language business objectives into governed AI projects, including data preparation flows, machine-learning models, AI agents, and applications.
Dataiku announced Cobuild, an AI building tool that the company says will become generally available on June 18, 2026.
In a PR Newswire release published by Dataiku, the company described Cobuild as an “AI building agent” intended to convert plain-language business goals into governed, production-ready AI projects. Dataiku said those projects can include data preparation flows, machine-learning models, AI agents, and applications.
Dataiku’s own press release uses similar language, saying Cobuild is designed to help users create AI project flows from business needs while keeping governance controls in place. The company framed the launch around a common enterprise problem: organizations want more people to build with AI, but still need oversight, permissions, and controls before projects are used in production.
On its Cobuild product page, Dataiku says the tool generates “complete, governed, inspectable visual flows” from plain-language requirements. The company also says governance and permissioning are built into the experience.
That positioning matters because many enterprise AI tools are being marketed around speed and automation, while buyers also ask how generated systems can be checked, modified, and approved. Dataiku’s product materials emphasize inspectability, suggesting users can review what Cobuild creates rather than treating outputs as a black box.
According to Dataiku’s announcement, Cobuild is meant to support the creation of production-ready work rather than only prototypes. The company says the tool can help assemble assets across the AI project lifecycle, including data work, machine learning, applications, and agents.
Dataiku, which provides an enterprise AI and analytics platform, is presenting Cobuild as a way to reduce the gap between building AI systems and governing them. In the PR Newswire announcement, the company described this as closing the “build-govern gap,” a phrase that reflects its claim that AI development and AI oversight should happen in the same environment.
The launch also fits a broader market trend in which enterprise software vendors are adding AI-assisted creation tools. Dataiku’s materials focus on business users expressing objectives in natural language, while the system generates structured project components that technical teams can inspect and manage.
Dataiku has not, in the provided materials, published independent benchmark results or third-party evaluations for Cobuild. The available sources are company-issued materials and should be read as Dataiku’s description of the product’s capabilities.
The key test for Cobuild will be whether enterprises find that its generated project flows are useful, auditable, and secure enough for real operational use. Dataiku’s launch materials say Cobuild is designed for governed production work, but adoption will depend on how well the tool fits existing data access rules, approval processes, and collaboration between business and technical teams.
For now, the announcement gives Dataiku a clearer answer to a question many enterprise AI vendors are addressing: how to let more employees create AI-enabled systems without separating that work from governance, permissions, and review.
Dataiku announced Cobuild, an AI building tool that the company says will become generally available on June 18, 2026.
Dataiku said those projects can include data preparation flows, machine learning models, AI agents, and applications.
Dataiku’s own press release uses similar language, saying Cobuild is designed to help users create AI project flows from business needs while keeping governance controls in place.
Continue reading
