Hi everyone!
I've been testing the new enhanced conversational understanding for knowledge searches in Agent Copilot, and the difference was really clear in a simple scenario.
The customer first said, "I want to rent a car" and then asked, "But how long does it take?"
Without the new configuration, Copilot kept the search focused on the previous message, "I want to rent a car", and the result was more generic and not really related to the latest question.
After enabling Detect knowledge questions and Refine search queries, Copilot used the conversation context and reformulated the query to something like "How long does it take to rent a car?"
The result was much more relevant to what the customer was actually asking.
I think this can help a lot in real conversations because customers rarely ask complete and standalone questions. Most of the time they say things like "And how long does it take?", "What about this case?" or "How do I do that?"
Using the conversation context to refine the knowledge search query can make a big difference in the relevance of Copilot suggestions.
I also liked that the refined query is visible to the agent. It makes it easier to understand how Copilot interpreted the conversation and can also help when troubleshooting knowledge search results.
Has anyone else tested this feature yet? I'd be curious to hear how it is performing in other use cases and with larger knowledge bases.
#AICopilot(Agent,Supervisor,Admin)------------------------------
Mateus Nunes
CX Manager at Solve4ME
mateus.nunes@solve4me.com.brBrazil
------------------------------