> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://help.sigmacomputing.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://help.sigmacomputing.com/_mcp/server.

# Configure a warehouse-hosted model as AI provider

> Use warehouse-hosted AI models from your connected data platform as the AI provider for Sigma organization features.

When you [configure AI features for your organization](/docs/configure-ai-features-for-your-organization), you can use the AI models available in your connected data platform as your AI provider to power Sigma AI features. For the models that Sigma uses with each data platform, see [Supported AI models](/docs/supported-ai-models).

> **Info**
>
> This feature isn't supported by all data platform connections. To check if your connection supports it, see [Supported data platforms and feature compatibility](/docs/region-warehouse-and-feature-support#supported-data-platforms-and-feature-compatibility).

> **Note**
>
> The use of AI features is subject to the following [disclaimer](/docs/notice-for-enabling-ai-enabled-features-in-sigma).

## User requirements

You must be assigned the Admin [account type](/docs/create-and-manage-account-types).

## BigQuery requirements

If you use BigQuery as your AI provider, Sigma uses [Vertex AI](https://docs.cloud.google.com/vertex-ai/docs) to access Google-provided models. You cannot choose which model to use. For the models that Sigma uses, see [Supported AI models](/docs/supported-ai-models).

> **Info**
>
> Sigma uses the [global endpoint](https://docs.cloud.google.com/gemini-enterprise-agent-platform/resources/locations) to access the supported model. If you have regional restrictions for AI data processing, consider using a different AI provider.

To use BigQuery as your AI provider, the following requirements apply:

* You must set up a connection to BigQuery in Sigma. See [Connect to BigQuery](/docs/connect-to-bigquery).
* Billing must be enabled for your Google Cloud project. See [Enable, disable, or change billing for a project](https://docs.cloud.google.com/billing/docs/how-to/modify-project) in the Google Cloud documentation.
* Your [BigQuery service account](/docs/connect-to-bigquery#create-a-bigquery-service-account) must be assigned the following IAM roles:

  * `Vertex AI User` to grant access to use Vertex AI
  * `BigQuery Data Viewer`
  * `BigQuery Job User`
  * (Optional) `BigQuery Data Editor` to enable [write access](/docs/set-up-write-access) to your connection. Required if you [configure a usage dashboard for Sigma Assistant](/docs/configure-ai-usage-dashboard).
* (Optional) Identify a region to access the Vertex AI API from. Sigma uses the [Vertex AI API](https://docs.cloud.google.com/vertex-ai/docs/reference/rest) to access the model. Google provides several different endpoints to access models located in different regions. For a full list of supported endpoints, see the [Service endpoint](https://docs.cloud.google.com/vertex-ai/docs/reference/rest#service-endpoint) section in the Vertex AI API page of the Google Cloud documentation. To reduce latency, Sigma recommends identifying an endpoint that is located near [the region where your Sigma organization is hosted](/docs/region-warehouse-and-feature-support#supported-cloud-platforms-and-regions).

> **Info**
>
> Performance of the Vertex AI service used by Sigma depends on the current traffic and usage of the service. You might experience higher latency during busy periods. For more details, see [Dynamic shared quota (DSQ)](https://docs.cloud.google.com/vertex-ai/generative-ai/docs/dynamic-shared-quota) in the Google Cloud documentation.

## Databricks requirements

If you use Databricks as your AI provider, Sigma uses models made available as Foundation Models through Mosaic AI Model Serving. For more details, see [Deploy models using Mosaic AI Model Serving](https://docs.databricks.com/aws/en/machine-learning/model-serving/) in the Databricks documentation.

To use Databricks as your AI provider, the following requirements apply:

* You must set up a connection to Databricks in Sigma. 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 `model-serving` scope to the PAT.

* The region of your Databricks workspace must support the Foundation Model API. See [Model serving features availability](https://docs.databricks.com/aws/en/resources/feature-region-support#model-serving-features-availability) on the Features with limited regional availability page in the Databricks documentation.

* The region of your Databricks workspace must support both the reasoning model and the embeddings model that Sigma uses. You cannot choose which model to use. For the models that Sigma uses, see [Supported AI models](/docs/supported-ai-models). To determine if your region includes the required models, see [Supported foundation models on Mosaic AI Model Serving](https://docs.databricks.com/aws/en/machine-learning/model-serving/foundation-model-overview).

* Your Databricks account must have access to the `ai_query` function and meet the relevant requirements, such as a running Databricks Serverless or Pro SQL warehouse. For more details, see the [Requirements](https://docs.databricks.com/aws/en/sql/language-manual/functions/ai_query#requirements) section of the `ai_query` function page in the Databricks documentation.

* Users must have access to a Databricks workspace with the supported models. Access requires granting the `workspace-access` entitlement to the service principal that authenticates the connection. If the connection is authenticated with OAuth, grant the entitlement to each user who needs access to the AI provider. Requests from a user without this entitlement fail with the error: `403 This API is disabled for users without the workspace-access entitlement.` See [Manage entitlements](https://docs.databricks.com/aws/en/security/auth/entitlements) in the Databricks documentation.

> **Info**
>
> Databricks models might have a rate limit set. If you see an error: `PERMISSION DENIED: The endpoint is temporarily disabled due to a Databricks-set rate limit of 0.`, contact your Databricks admin about the relevant rate limits set for your workspace and the reasoning model that Sigma uses. See [Foundation Model APIs limits and quotas](https://docs.databricks.com/aws/en/machine-learning/foundation-model-apis/limits/) in the Databricks documentation.

## Snowflake requirements

If you use Snowflake as your AI provider, Sigma uses models made available through the Cortex REST API in Snowflake. For more details, see [Cortex REST API](https://docs.snowflake.com/user-guide/snowflake-cortex/cortex-rest-api) in the Snowflake documentation.

To use Snowflake as your AI provider, the following requirements apply:

* You must set up a connection to Snowflake in Sigma. The connection must be authenticated with OAuth or key pair authentication. See [Connect to Snowflake](/docs/connect-to-snowflake).

* The default role associated with the connection must be granted the `SNOWFLAKE.CORTEX_USER` database role. By default, this role is granted to the PUBLIC role. If you need to grant the database role, see [Setting up authorization](https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-llm-rest-api#setting-up-authorization) in the Snowflake documentation. More specifically:

  * For key pair authentication, the default role of the service account user must be granted the `SNOWFLAKE.CORTEX_USER` database role.
  * For key pair authentication and [dynamically assign roles to users](/docs/dynamically-assign-roles-used-by-a-connection), all roles must be granted the `SNOWFLAKE.CORTEX_USER` database role.
  * For OAuth authentication, the default role of all users with access to the connection must be granted the `SNOWFLAKE.CORTEX_USER` database role. If you configure a service account, the role of the service account user must also be granted the `SNOWFLAKE.CORTEX_USER` database role.

    > **Note**
    >
    > The Cortex REST API is not available in Snowflake trial accounts. If you set up an AI provider that uses a Snowflake trial account, you might see the error `400 status code (no body)`. If you encounter this error when connecting to a Snowflake trial account, contact Snowflake Support.

### Supported models

The models used by Sigma depend on where your Snowflake account is hosted. You cannot choose which model to use. For the models that Sigma uses, see [Supported AI models](/docs/supported-ai-models). The region where your Snowflake account is hosted must support the following:

* The EMBED function model used for embeddings.

  * If your region supports the embed model through the Cortex REST API, the API is used. See the section [Model availability](https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api/embed-api#model-availability) in the Vector embedding API page in the Snowflake documentation.
  * If your region does not support accessing the embed model using the Cortex REST API, the `ai_embed()` function is used to call the model through a SQL query. For details about regions that support this function, see [Region availability](https://docs.snowflake.com/en/sql-reference/functions/ai_embed#region-availability) in the AI\_EMBED page in the Snowflake documentation. If your Snowflake account is hosted in a different region, enable the `CORTEX_ENABLED_CROSS_REGION` account parameter. For more details, see [CORTEX\_ENABLED\_CROSS\_REGION](https://docs.snowflake.com/en/sql-reference/parameters#label-cortex-enable-cross-region) in the Account parameters page in the Snowflake documentation.

* A large language model (LLM) used for reasoning, called through the Cortex REST API. Specifically, Sigma requires the following:

  * If your Snowflake account is hosted in AWS, your AWS region must also support the model that Sigma uses for AWS-hosted accounts.
  * If your Snowflake account is hosted in Azure, your Azure region must also support the model that Sigma uses for Azure-hosted accounts.
  * If your Snowflake account is hosted in a region that does not support either of those models, you must enable the `CORTEX_ENABLED_CROSS_REGION` account parameter to allow access to a region that supports the relevant model. If you enable this parameter, the `openai-gpt-5.1` model is used. For more details about cross-region inference, see [Cross-region inference](https://docs.snowflake.com/en/user-guide/snowflake-cortex/cross-region-inference) and [CORTEX\_ENABLED\_CROSS\_REGION](https://docs.snowflake.com/en/sql-reference/parameters#label-cortex-enable-cross-region) in the Account parameters page in the Snowflake documentation.

  For details, see the section [Model availability](https://docs.snowflake.com/en/user-guide/snowflake-cortex/cortex-rest-api#model-availability) on the Cortex REST API page in the Snowflake documentation.

## Configure a warehouse-hosted model as your AI provider

After setting up your connection and confirming the availability of supported models, set up your AI provider to use a warehouse-hosted model:

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. In the **AI provider** section, for **Provider hosting**, select the **Data warehouse hosted model** option.

3. In the **Connection** dropdown menu, select the connection.

4. (Optional, BigQuery connections only) In the **Region** dropdown menu, select the region to access the Vertex AI API from. By default, Sigma uses `us-central1`.

5. Click **Save**.

After adding the AI provider, [choose data sources for Sigma Assistant to use](/docs/configure-ai-features-for-your-organization#choose-data-sources-for-sigma-assistant).

## Configure an embeddings provider

Sigma uses an embeddings model to index your data sources for [semantic search](/docs/search-in-sigma). By default, Sigma uses your AI provider to generate embeddings. If your AI provider does not support embeddings, or you want to use a different provider for embeddings, you can configure a separate embeddings provider.

> **Note**
>
> Anthropic does not provide an embeddings model that Sigma can use. If you use Anthropic as your AI provider, you must configure a separate embeddings provider.

To configure a separate embeddings provider:

1. Go to **Administration** > **AI settings**.

2. In the **Embeddings provider** section, turn off the **Use the configured AI provider** toggle. If your AI provider does not support embeddings, this option is unavailable and you must configure a separate embeddings provider.

3. For **Provider hosting**, select one of the following, then configure your selected provider:

   * **External models**: For **AI provider**, select **OpenAI** or **Gemini**, then enter the **API key**.
   * **Data warehouse hosted model**: In the **Connection** dropdown menu, select a connection. For BigQuery connections, you can also select a **Region**. For connection and model requirements, see [Configure a warehouse-hosted model as AI provider](/docs/configure-warehouse-ai-model-integration).

4. Click **Save**.

> **Info**
>
> When you add or change the embeddings provider, Sigma reindexes searchable sources in the background.

## Manage a warehouse-hosted AI provider

You can edit the AI provider integration at any time. For example, to update your credentials:

1. Go to **Administration** > **AI settings**.
2. In the **AI provider** section, click **Edit**.
3. Make your changes. For example, change the AI provider.
4. Click **Save**.

### Remove an AI provider

To disable AI functionality within Sigma, you can remove the AI provider:

1. Go to **Administration** > **AI settings**.
2. In the **AI provider** section, click **Edit**.
3. Click **Remove**.
4. When prompted to confirm, click **Remove**.
5. Click **Save**.

   The selected AI provider, credentials, and any configured data sources are removed.

After you remove the integration, the **AI provider** section displays a **Save** button, and AI functionality is unavailable for your organization.

## 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)
* [Configure an external AI provider](/docs/manage-external-ai-integrations)
* [Supported AI models](/docs/supported-ai-models)