What is Agent Studio?

Agent Studio is an AI platform that empowers users to rapidly create AI agents using the trusted context already available in Alation.

We’re here to supercharge your AI agents with Alation’s trusted metadata.

AI agents often fail in the enterprise for a simple reason: they don’t have the right context. That gap typically shows up in three ways:

Fragmented knowledge across systems and teams
Conflicting or outdated context that leads to inconsistent outputs
Unsustainable manual effort to keep agents accurate and useful

Agent Studio is built to solve these problems by grounding agents in a trusted, continuously maintained knowledge foundation—built on top of the context already available in Alation.

This page covers:

The foundation: tools
What Agent Studio provides
Where you can build and run agents
Common usecases
Who uses Agent Studio?
How do I use Agent Studio?

Tools are the foundation of Agent Studio

In Agent Studio, tools are the foundation for everything you build and deploy.

Our out-of-the-box tools are built on top of the trusted context you already have—your Alation catalog, data products, and even your databases—so agents can take real actions using reliable enterprise knowledge. Because tools can be accessed both inside and outside of Alation, they make it easy to build agents in a variety of ways depending on where you want those agents to run.

A complete list of Alation-provided tools is available here.

What Agent Studio provides

Agent Studio provides a seamless way for customers to unlock trusted context in Alation and put it to work in AI agents. Depending on how you want to build and deploy, we support a few paths:

Tools (the building blocks) We provide tools that can be accessed via MCP servers (hosted or local). You can use these tools to build agents in whatever platform you like—inside Alation or outside of it. A complete list of Alation-provided tools is available here.
Out-of-the-box agents We provide pre-built agents that solve common tasks on top of the Alation catalog. These agents can be used as-is or serve as starting points for further agent customization. They can be accessed externally via MCP (hosted or local) or REST APIs. A complete list of Alation-provided base agents is available here.
Custom agent building and evaluation We provide the ability to build custom enterprise agents on top of our tools. You can build from scratch, or customize and edit one of our out-of-the-box agents to your match your enterprise’s needs.

To ensure high quality agents, Agent Studio provides a custom evaluation interface that enables you to create, monitor, and track evaluation sets for each agent you build—so you can ensure your agent keeps performing well for your enterprise over time.

Like out-of-the-box agents, these agents can be accessed externally via MCP (hosted only) or REST APIs.

No matter which approach you choose, Agent Studio makes it easy to move agents from prototype to production and bring them to wherever your AI already runs today. Our philosophy is simple: Get what you need from Alation, then deliver value where the work happens.

Where you can build and run agents

Agent Studio supports both building in Alation and building outside Alation, while still using the same underlying tools and trusted context.

Build inside Alation (Web UI)

In the Agent Studio Web UI, you can:

Use default (out-of-the-box) agents to solve common catalog tasks
Build custom agents on top of Agent Studio tools
Evaluate and iterate using the built-in evaluation interface

When you’re ready, you can publish your agents so they can be used outside of Alation via MCP or REST APIs.

Build outside Alation (your agent platform)

If you prefer to build and run agents in your own environment, you can still leverage Agent Studio:

Use Agent Studio tools directly via MCP
Use the Agent Studio SDK to access tools and agents from Python

This approach lets you bring Alation context into the workflows you already run—without forcing you to move your entire agent stack into Alation.

When should I use Agent Studio?

Use Agent Studio when you want to build agents on top of the metadata and context inside the Alation catalog—especially when you need customization beyond “one-size-fits-all” catalog experiences.

Common reasons include:

You need a custom agent built on Alation metadata and context
You want to bring Alation catalog context outside of Alation to power other agents or workflows

Next, we’ll walk through a few common use cases.

Data Insight Agents

Data insight agents help teams answer business questions using structured data—faster, more consistently, and with less back-and-forth between analysts and stakeholders.

As a business analyst:

I want to ask questions in natural language directly against data in a database, without needing to know where it lives or how it’s modeled.
I want to automate repetitive, data-driven tasks and routines powered by structured data in a database (for example, a weekly supply chain workflow that checks inventory).

Why this matters:

Reduces time spent translating business questions into SQL or dashboards
Improves trust by grounding answers in governed definitions, lineage, and approved data products
Enables operational workflows that depend on structured data, not one-off analysis

Metadata Insight Agents

Metadata insight agents help users understand data—what it means, where it comes from, and how it should be used—by using the trusted context stored in Alation.

As a data analyst or data app developer:

I want to generate or augment a semantic layer using my context in Alation
I want to pull Alation context into other tools (for example, Claude or ChatGPT) so answers are grounded in our enterprise definitions, policies, and governance

Why this matters:

Makes “enterprise-safe” AI possible by grounding responses in trusted metadata rather than guesswork
Reduces ambiguity by standardizing definitions and business context across teams
Helps external agent experiences (like Claude or ChatGPT) stay aligned with what the catalog says, even when users interact outside of Alation

Curation Agents

Curation agents automate and scale the work required to keep your catalog accurate, complete, and useful—especially when you need custom enrichment logic.

As a data steward:

I want to build a custom curation agent to enrich data in my catalog, and I can’t use Alation’s curation automation tool to do this

Why this matters:

Keeps the catalog continuously improved without relying on manual updates
Helps enforce organization-specific standards for documentation, classification, and governance
Enables enrichment workflows tailored to your enterprise (instead of generic curation rules)

Who uses Agent Studio?

Agent Studio provides different entry points depending on how much you want to build, but we believe strongly that with AI, everyone is a builder.

Agent Studio is designed for:

Business analysts, who primarily use the UI to run and customize agents for common workflows.
Data analysts, who use the UI and/or MCP servers to build, iterate, and operationalize agents on top of structured data.
Data stewards, who use the UI and/or Python SDK to apply governed context and ensure agents stay aligned with trusted definitions and policies.
AI developers, who use MCP servers, REST APIs, and/or the Python SDK to build and deploy agents into existing AI workflows and platforms.

All of these users are looking to build or use agents on top of structured data and the Alation catalog.

How do I use Agent Studio?

You can access Agent Studio through:

The Web UI
Hosted MCP servers
The Python SDK

Each entry point gives you access to the tools and agents available in Agent Studio, but each is optimized for a different workflow:

Web UI: The easiest way to get started inside Alation. Create and customize agents in a UI, test quickly, and build evaluation sets. When you are ready, publish agents for use outside of Alation.
Hosted MCP servers: Best for building agents outside of Alation, or for directly using existing tools and out-of-the-box agents when you do not need to customize and publish a new agent.
Python SDK : Best for developers who want to use Agent Studio tools and agents from a Python development environment.