I just wanted to check peoples opinion on using the Participant Data as a mechanism for exposing business metrics in flows/bots? i.e:
- how many customers dropped out after being asked question X
- how many customers answered the question with Y
- how many customers found themselves in Z part of the flow
The flows add an attribute of a particular format to the Participant Data. A Lambda will then extract these attributes from the AttributeUpdated events and push them to our observability platform. We can then make graphs/alerts from these metrics in realtime.
We're piggy backing on the Participant Data, but the solution is simple (cuts out using DAs all over the place), reliable (there are retry mechanisms built into EventBridge/Lambdas) and the metric attributes in the Participant Data offer information when trying to understand a conversation in the Interaction View.
Does anyone see any potential problems to this solution? I recall someone saying Participant Data should only be used for routing, but I can't see anything in the documentation stressing that this is the only use-case.