A company uses a Dataverse table to manage support cases.

The maker should define a prompt that references the support-case fields and create the destination column with the Prompt data type. The prompt contains the instructions for producing the case summary and declares the record inputs needed to ground that output. The prompt column then makes the generated summary available in the Dataverse/model-driven app context for the current row. A formula column performs deterministic Power Fx computation and cannot generate a natural-language summary from multiple fields. A business rule can set or validate values, but it does not invoke a generative model. The prompt should limit inputs to the approved columns, specify the required structure, and handle empty or contradictory case data so the generated content remains useful. Security on the source fields must still be respected, because an AI-generated summary must not become a route around restricted data. The AB-410 Study Guide lists “configure prompt columns” under Build data models and “add inputs to a prompt” under AI Hub. Options A and D supply the instruction definition and the Dataverse column behavior required for record-context generation.

Study Guide reference/topic: “Build data models — Configure prompt columns.” AB-410 Study Guide | Microsoft Learn technical reference