Hi @Ursula Correa Babeto,
Agent Copilot's summarization feature generates summaries based on the agent-customer conversation content. When the interaction ends, Copilot analyzes the conversation transcript to create a summary, along with suggestions for wrap-up codes and fields like "Reason contacted" and "Resolution."
The knowledge base is used separately during the interaction for a different purpose - to surface relevant knowledge articles in real-time based on customer queries and conversation context. This helps agents find answers while they're talking to customers but it's not integrated into the summary generation process.
With personalized (custom) prompts, you can customize how the summary is created by defining the tone, length, structure and focus - but the summary is still generated from the conversation content itself. The summary automation runs after the call concludes, processing what was said during the interaction to create a qualitative history of what happened and how situations were resolved.
Hope this helps you.
------------------------------
Josh Coyle
Senior Professional Services Consultant
------------------------------
Original Message:
Sent: 03-05-2026 10:22
From: Ursula Correa Babeto
Subject: Knowledge base and summaries
Hey everyone! Does anyone know if, when generating a personalized summary, Copilot also checks the knowledge base for context, or if it relies only on the conversation itself?
#AIConfiguration
#AICopilot(Agent,SupervisorAdmin)
------------------------------
Ursula Correa Babeto
ursula.babeto@capgemini.com
------------------------------