An open-data portal web app must reach the correct Foundry project that hosts the generative assistant. What should administrators configure?

Configuring the web app connection to the correct Foundry project endpoint is what open-data administrators must do so the portal reaches the intended generative assistant. Foundry organizes models and resources by project; mis-pointed endpoints send traffic to the wrong models, quotas, or environments. Explicit project endpoint configuration in the app settings binds the portal to the assistant that was deployed and evaluated for open data.

Configuring the web app connection to the correct Foundry hub management endpoint points the portal at hub administration rather than the project hosting the generative assistant, so it cannot reach the deployed open-data assistant. Configuring the web app connection to the correct Foundry search index endpoint routes requests to a search index for data lookup instead of the project endpoint, so the portal never reaches the assistant itself. Configuring the web app connection to the correct Foundry model publish endpoint sends traffic to the model publishing workflow rather than the production project hosting the open-data assistant.

Exam caveat: connection configuration to the correct Foundry project is a first-class deploy step for generative web apps. Operational check: resolve the configured endpoint against the intended project in the Azure/Foundry portal, send a canary prompt, and verify response metadata (deployment name/project) matches the open-data assistant—not another department's project.