Genesys Cloud - Main

I am using programs currently to score all interactions, however I noticed that it's evaluating all interactions rather than the ones specifically tied to the topic. Is there anyway to have the virtual supervisor AI score all interactions within a program for ONLY the topic selected?

An example, we would like to evaluate only the calls where the client doesn't have routing & account number and we created a topic for EFT phrases only, but when we enable to program it evaluates calls even without that topic in the interaction.

Hello Caitlyn, 

As of right now, Genesys Cloud doesn't support restricting automated scoring in programs to only interactions where specific topics are detected. When you turn on automated scoring, it evaluates all interactions within the assigned queues or flows, regardless of whether a topic is present. Programs can still use topics to help identify intent and organize analytics, but topic detection itself doesn't act as a filter for scoring. So even if you're only interested in certain topics, the system will still score every eligible interaction.

As a workaround, you can use topic-based evaluation assistance in your evaluation forms to automatically populate answers when certain topics are found, but this still happens after all interactions are scored. You can also filter interactions by topic afterward to focus your review, though again, scoring will have already been applied across the board. Quality Management policies may help narrow things down using criteria like queues or wrap-up codes, but they don't support topic-based filtering either. So for now, there isn't a built-in way to pre-filter automated scoring by topic, but these approaches can help you get closer to what you're looking for.

Hope this helps!

I made a trigger which fires on specific topicId's.

v2.speechandtextanalytics.conversation.{id}.topics

$.topics[*].topicId contains  <TOPICID>

In this workflow I then gather the agentID and the conversationID as input to assign one of the AI QM forms to the interactions where the topics matched.