Skip to content

What is Agent Studio?

Agent Studio lets you build AI agents grounded in the trusted context already in Alation: your catalog, your data products, and your databases.

Enterprise agents usually fail for one reason: they lack the right context. Knowledge is fragmented across systems, definitions conflict or go stale, and keeping agents accurate by hand doesn’t scale. Agent Studio fixes this by grounding every agent in the Alation Knowledge Layer, the continuously maintained metadata your organization already trusts.

Governed tools

Callable actions built on your catalog, data products, and databases. Use the built-in tools in Alation or in any MCP-capable platform.

Agents, no code required

Pre-built agents for common catalog tasks, or clone one and make it yours in the web UI.

Evaluations built in

Measure accuracy with evaluation sets before you ship — and keep measuring after.

Deploy anywhere

Publish what you build via MCP, REST, Slack, or the SDK. Get what you need from Alation, then use it where the work happens.

Inside Alation, the Agent Studio web UI is where you create, test, evaluate, and publish — no code required. Outside Alation, the same tools and agents are available over MCP, the REST API, and the Python SDK, so you don’t have to move your agent platform to benefit from its context.

Everyone is a builder, but the entry points differ. Business analysts run and customize agents in the UI. Data analysts and stewards iterate in the UI and operationalize over MCP or the SDK. AI developers consume tools and agents from MCP, REST, and Python in their existing platforms.