Previously using Einstein bot with picklist questions, it was easy to report on the customers intent when starting a messaging session.
Moving to Agentforce only, we no longer have this picklist option so do not have the same solution.
Has anyone found a good way to report on contact reasons/customer intent with Agentforce only conversations please?
#Agentforce #Reporting
Hi Vicky,
Since Agentforce replaces the picklist-menu style intent capture with free-text + AI classification, the reporting side moves to Agentforce Observability/Analytics rather than a Case/session field.
Two native options to look at:
1. Agentforce Observability — lets you segment and cluster conversations by "Intent tags," drilling into which Topic/subagent handled each session. This is the closest native replacement to your old picklist-based intent reporting, but it's session-level analytics, not a field on the record itself.
2. Utterance Analysis dashboard (in the Agentforce Default folder) — groups similar customer utterances into Clusters, then into Categories, so you can see patterns in what customers are actually asking for, even without a picklist. Reports refresh weekly.
If you specifically need intent captured as a reportable field on the Case/Messaging Session (e.g., for existing dashboards), the practical workaround most teams use is: have the agent invoke a Flow/Apex action early in the conversation that writes the classified Topic name into a custom field on the Case or MessagingSession record — since Topic classification already happens under the hood, you're just persisting it instead of relying on native analytics.
References:
-
https://www.salesforce.com/agentforce/observability/-
https://trailhead.salesforce.com/content/learn/modules/agentforce-analytics-and-monitoring/monitor-agent-conversations-with-utterance-analysisWorth checking if Observability's Intent tags alone cover your reporting need before building the custom field workaround — it may save you the extra flow logic.
