Dataiku Inc., a leader in enterprise AI platforms, launched Agent Management, a solution designed to discover and monitor all AI agents a business operates, regardless of the vendor platform. Alongside this, the company expanded Cobuild, its no-code natural-language environment for building AI projects, extending its usability across multiple developer workflows.
- Agent Management inventories AI agents across multiple vendor platforms in one interface.
- Cobuild evolves with enhanced natural language querying and integration into developer coding agents.
- The new tools aim to improve AI governance, risk assessment, and operational transparency in enterprises.
What happened
Dataiku released Agent Management as a standalone product to help enterprises find, inventory, and track every AI agent they run, including those built on other companies' platforms. This product integrates with major platforms such as Salesforce, AWS, Microsoft, Google, Databricks, and Snowflake, as well as custom environments via OpenTelemetry.
In parallel, Dataiku expanded Cobuild, its natural-language interface for creating AI projects. The updates enable all employees to ask questions of governed company data and receive high-quality responses, while developers can use Cobuild within popular coding agents like Anthropic's Claude Code, OpenAI's Codex, and SpaceX's Cursor to build or modify AI workflows and pipelines.
Why it matters
Most enterprises today struggle to maintain a complete and current inventory of the AI agents they deploy, despite having detailed records of other IT assets. According to industry research cited by Dataiku, fewer than 20% of organizations keep an up-to-date catalog of their AI systems, presenting risks in governance and audit readiness.
By offering a vendor-agnostic inventory and continuous risk monitoring with test scheduling and certification tracking, Agent Management delivers transparency and an audit trail that enterprises need to manage AI at scale and reduce compliance risks. Coupled with Cobuild’s enhancements, the combined tools reduce barriers for both business users and developers to manage AI projects effectively and align them with organizational governance standards.
What to watch next
Dataiku plans to make Agent Management generally available in October 2026, with access fees charged annually per instance and usage metered by the number of agents monitored. Early adopters will likely assess the solution's ability to deliver portfolio-wide visibility across heterogeneous AI environments.
The expanded Cobuild capabilities signal Dataiku's strategy to democratize AI project creation and embedding intelligent workflows within existing developer tools. Observers should watch how enterprises leverage these features to streamline AI adoption, governance, and integration across business and technical teams, potentially setting new standards for enterprise AI management.