A Copilot Studio topic calls a Power Automate cloud flow that processes an expense report. The topic has collected the employee's name, department, and total amount as topic variables. The developer needs to pass these values to the flow and receive back an approval status. How is this data exchange between the topic and the flow configured?

Power Automate flows that are called from Copilot Studio must be built with the 'Run a flow from Copilot' (formerly 'When Copilot calls a flow') trigger. This trigger defines the flow's input and output contract.

Option B is correct. When the developer builds the Power Automate cloud flow, they add the 'Run a flow from Copilot' trigger and configure three text input parameters: EmployeeName, Department, and TotalAmount. At the end of the flow, a 'Return value(s) to Copilot' action sends back the ApprovalStatus output. In the Copilot Studio topic's flow action node, the developer maps Topic.EmployeeName to the EmployeeName input, Topic.Department to Department, Topic.Amount to TotalAmount, and maps the ApprovalStatus output to a new topic variable Topic.ApprovalStatus.

Option A is incorrect. Using SharePoint as an intermediary data store for real-time flow input/output exchange is unnecessarily complex, adds latency, and creates potential timing issues. Direct parameter passing through the trigger is the correct approach.

Option C is incorrect. Power Automate flows do not have automatic access to Copilot Studio global variables. Data must be explicitly passed as flow input parameters. The flow runs in a separate execution context.

Option D is incorrect. The topic can absolutely pass data to a flow through input parameters defined in the 'Run a flow from Copilot' trigger. This is the primary and intended mechanism for topic-to-flow data exchange.