Question 28 — AB-620 Case Study Set
Based on the scenario, which integration pattern should Paideffort use for the current carrier tracking requirement? (Select one answer)
Use a Send HTTP Request node directly in the shipment-tracking topic.
Build an agent flow containing the single carrier GET request.
Create a connector tool solely for the carrier tracking operation.
Invoke the Foundry support agent to perform the REST request.
Correct answers
Use a Send HTTP Request node directly in the shipment-tracking topic.
Explanations
Correct Answer:
Use a Send HTTP Request node directly in the shipment-tracking topic.
Explanation:
Paideffort needs one self-contained REST call inside one topic, and the carrier provides no Power Platform connector, OpenAPI specification, or MCP server. There is also no reuse or multi-step orchestration requirement. The Send HTTP Request node therefore provides the smallest suitable integration surface without introducing unnecessary reusable infrastructure.
Wrong Answers:
Build an agent flow containing the single carrier GET request.
An agent flow is better suited to multi-step processing, conditional workflows, or reusable logic. Paideffort currently has one request used by one topic, so adding a flow would increase maintenance without addressing a stated requirement.
Create a connector tool solely for the carrier tracking operation.
A connector could provide reuse if Paideffort had a broader integration requirement, but the carrier currently has no connector and the operation is isolated to one topic. Building connector infrastructure isn't necessary for this scope.
Invoke the Foundry support agent to perform the REST request.
The tracking operation is deterministic and doesn't require model reasoning. Routing the request through the Foundry agent would introduce avoidable latency and complexity.