問題文
Three requirements arrive for the same agent. R1: ground answers on a large third-party knowledge base with citations. R2: read the current status of an order from a line-of-business system without copying the data. R3: create a case in that system. Which allocation is correct?
選択肢
- All three to Copilot connectors, because indexing covers reading and writing once the schema is defined.
- R1 and R2 to Copilot connectors, R3 to an agent flow with no connector, because flows can write directly to any system.
- R1 to Copilot connectors, R2 to a Power Platform connector used as real-time knowledge, R3 to the same connector used as a tool or action.
- All three to Power Platform connectors, because runtime calls can also provide citations at scale.