Configuring an endpoint for GenAI

The endpoint must be deployed successfully and have an active status before you configure the endpoint for GenAI integration. For optimization models, you can integrate fast-track deployment endpoints and design-and-deployment endpoints.

  1. Open the endpoint details.
  2. Select the GenAI Tool tab.
  3. Select the Integrate this model with GenAI option.
  4. Specify a description in the Model Description field. The description must explain the business purpose of the model, the type of result that the model returns, and the scenario in which GenAI uses the model.
  5. Specify column descriptions in the Input Schema Context field. This information is required.
  6. Review the endpoint schema. Ensure that descriptions exist for all required input parameters, sets, and constants in the optimization model.
  7. Click the Save.
    Note: 

    When you save the configuration, the endpoint becomes available for GenAI-assisted invocation, subject to the GenAI feature enablement and tenant configuration.

    If the changes made to the real-time endpoint schema after the GenAI integration is configured, review and update the GenAI input schema context before saving the endpoint again. This action ensures that the schema descriptions remain aligned with the endpoint input requirements.