Configure the AI provider for the organization associated with the API credentials. Choose one provider for LLM interactions, and optionally a separate provider for embeddings.

Usage notes

• To perform this operation, you must use API credentials owned by a user assigned the Admin account type.
• Providers differ in which models and regions they support. Review Supported AI models and Set up an AI provider first.
• Retrieve the connectionId by calling the /v2/connections endpoint.
• embeddedModel is optional for warehouse providers, OpenAI, Gemini, and Bedrock. If you omit it, the embedding model for the LLM provider is used. Required for Anthropic and Azure OpenAI.

Usage scenarios

• Automated organization setup: Configure the AI provider as part of onboarding or infrastructure-as-code, rather than clicking through the administration portal.
• Sigma Tenants: Set the AI provider for each tenant organization programmatically, using impersonation to target a specific tenant.