For an add-on that includes a flow, a small team is ready to widen its scope when the current flow has been tested with values for the input variables that would be used, and the team has a defined way to monitor deployed AI components in production. These checks support a decision; they do not create a complete approval standard. The team must still make the final call from its own evidence and constraints.
Check the current flow before expanding it
The flow-testing guidance says that flow creators can test a flow by entering values for the input variables that would be used and then running the flow. That gives the team a concrete way to inspect current behavior before adding more use cases.
A test result speaks to the values used in that test, not automatically to every future input. The cited guidance does not state a required test count, a fixed input list or a pass threshold, so the team must decide what evidence is sufficient for the proposed scope.
Make production monitoring part of the decision
The production-monitoring guidance says that the functionality and behavior of the AI system and its components—as identified in the map function—should be monitored when in production.
Production monitoring should therefore be considered before the scope expands, rather than treated as a task to postpone until after deployment. The team should be able to identify which components are covered and how observations will be reviewed. The cited guidance does not specify a monitoring frequency, escalation rule or service-level target; those choices remain with the team.
What the team must still confirm
Before widening the add-on, the team still needs to confirm:
- Which input values represent the intended use.
- What behavior would be acceptable in testing and what would trigger concern in production.
- Which deployed components are covered by monitoring.
- Who reviews the evidence and who authorizes the expansion.
- What action should follow if a concern appears, including whether to pause or reverse the expansion.
- Whether the wider use remains compatible with existing work and constraints without requiring replacement of core tools.
The cited guidance supports testing and monitoring, but it does not establish a universal approval chain, deadline, threshold or guarantee of outcomes. Any wider scope should be supported by the team's evidence, not merely by the completion of a test or the start of monitoring.