> For clean Markdown of any page, append .md to the page URL. > For a complete documentation index, see https://help.sigmacomputing.com/docs/use-warehouse-agents-sigma/llms.txt. > For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://help.sigmacomputing.com/_mcp/server. # Use warehouse agents with Sigma > Use Snowflake Cortex Agents and Databricks Genie Agents in Sigma Assistant and Sigma agents for direct warehouse AI interaction. > **Note** > > The use of AI features is subject to the following [disclaimer](/docs/notice-for-enabling-ai-enabled-features-in-sigma). If you create a [Cortex Agent](https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents) in Snowflake or an [Genie Agent](https://docs.databricks.com/aws/en/genie-agents/concepts) in Databricks, you can interact directly with that agent in Sigma Assistant or in a Sigma agent. By default, warehouse agents such as a Snowflake Cortex Agent or a Databricks Genie Agent use the data in your data platform directly, so any data transformations and restrictions applied to your data in Sigma, such as row-level security, are not applied when you interact with the agent from Sigma. ## Snowflake Cortex Agent requirements To set up a Snowflake Cortex Agent in Sigma, the following requirements must be met: * You must have a Snowflake connection. See [Connect to Snowflake](/docs/connect-to-snowflake). * You must connect to Snowflake using OAuth or key-pair authentication. Basic authentication is not supported for using Cortex Agents. * You must have a [Cortex Agent](https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents) set up in the connected Snowflake account. * Grant the following privileges and roles to the relevant roles associated with your connection: * For key pair authentication, grant roles and privileges to the default role of the service account user. * For key pair authentication and [dynamically assigned roles](/docs/dynamically-assign-roles-used-by-a-connection), grant roles and privileges to all roles. * For OAuth authentication, grant roles and privileges to the current role of all users with access to the connection. If you configure a service account, grant roles and privileges to the role of the service account user. The following roles and privileges are required: * The `SNOWFLAKE.CORTEX_USER` database role or the `SNOWFLAKE.CORTEX_AGENT_USER` database role. * `USAGE` privileges on the agent. For more details and other requirements, see the [Access control requirements](https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-agents#label-cortex-agents-access-control) for the Cortex Agent in the Snowflake documentation. ## Databricks Genie Agent requirements To use a Databricks Genie Agent in Sigma, the following requirements must be met: * You must have a Databricks connection. See [Connect to Databricks](/docs/connect-to-databricks). If you use a personal access token (PAT) to connect to Databricks and restrict the scope of your token, add the `genie` scope to the PAT. * You must have a Genie Agent in Databricks. * The following permissions must be set for each user accessing Databricks (if you use OAuth to authenticate) or the service principal associated with the connection (if you use basic authentication): * **CAN VIEW** or **CAN RUN** on the Databricks Genie Agent. * Permission to access the data queried by the Genie Agent. For more details, see [Share a Genie Agent](https://docs.databricks.com/aws/en/genie-agents/set-up#share-a-genie-agent) in the Databricks documentation. ## Set up warehouse agents with Sigma If you create a Cortex Agent in Snowflake or a Genie Agent in Databricks, after Sigma indexes your connection you can see the agent in the [data catalog](/docs/manage-data-catalog) and use it with [Sigma Assistant](/docs/ask-natural-language-queries-with-assistant) and [Sigma agents](/docs/sigma-agents). > **Note** > > If the agent does not appear in Sigma, sync the connection. When you sync the connection, sync both the database and schema or catalog and schema that contain the agent. For more details, see [Manually sync your data](/docs/troubleshoot-your-connection#manually-sync-your-data). ### Grant access to a warehouse agent in Sigma If you connect to your data platform using key-pair authentication (Snowflake) or a personal access token (Databricks), grant access to the warehouse agent to users and teams in Sigma. You must be assigned the Admin [account type](/docs/account-type-and-license-overview). > **Info** > > If you use OAuth to connect to your data platform, you must manage access to the agent in your data platform. To allow others to use the warehouse agent in Sigma, grant access by doing the following: 1. From Sigma Home, select **Connections** to open the list of connected data sources. 2. Select the connection with the warehouse agent. 3. Search for or browse to the agent, then select it. 4. On the **Access** tab, select **+ Grant access**. 5. Search for users or teams to grant **Can use** access to the agent. 6. (Optional) Select the **Send an email notification** checkbox to notify users that they now have access to the agent. 7. Click **Save**. ### (Optional) Turn off warehouse agents with Assistant By default, you can use warehouse agents like Snowflake Cortex Agents and Databricks Genie Agents with [Sigma Assistant](/docs/ask-natural-language-queries-with-assistant) and Sigma agents. If you don't want to use warehouse agents with Sigma Assistant: 1. Go to **Administration > AI settings**: 1. In the Sigma header, click your user avatar to open the user menu. 2. Select **Administration** to open the Administration portal. 3. In the side panel, select **AI settings**. 2. Select the **Assistant** tab. 3. In the **Agents** section, turn off the **Warehouse agents** toggle. Warehouse agents are no longer available to select in Sigma Assistant. ## Limitations * You can't switch to a different agent within an existing chat conversation. To ask a warehouse agent or Sigma Assistant a question instead, start a new chat. For other limitations that apply when using a warehouse agent with Sigma Assistant, see [Limitations with warehouse agents](/docs/ask-natural-language-queries-with-assistant#limitations-with-warehouse-agents). > **Info** > > If [chat history](/docs/configure-chat-history) is turned on for your organization, conversations with a warehouse agent are stored the same way as conversations with Sigma Assistant. ## Related resources * [Get started with AI in Sigma](/docs/getting-started-with-ai) * [Configure AI features for your organization](/docs/configure-ai-features-for-your-organization) * [Ask natural language queries with Sigma Assistant](/docs/ask-natural-language-queries-with-assistant) * [Configure chat history](/docs/configure-chat-history) * [About Sigma agents](/docs/sigma-agents) * [Build Sigma agents](/docs/build-agents) > Official documentation and guidance for using Sigma to build AI apps, dashboards, and agents on governed warehouse data.