Get grounded answers from your knowledge base via Co-Pilot
Combine AI understanding with knowledge-base retrieval to recommend accurate answers and supporting information. The feature reads the current conversation context, identifies the customer’s topic or information need, searches approved knowledge sources, and returns relevant suggestions that can be used to respond more accurately.
This capability is based on (RAG) Retrieval-Augmented Generation. The AI should not depend only on general model knowledge. It first retrieves trusted organizational content such as documentation, knowledge articles, PDFs, web pages, troubleshooting guides, and policy content. The LLM then uses that retrieved content to prepare grounded suggestions.
How to enable the feature
AI Knowledge Base Suggestions Configuration Guide
How it works
-
The customer asks a question.
-
AI understands the conversation context and identifies the topic or required information.
-
Searches indexed content using semantic search.
-
Relevant snippets, article titles, and source links are returned.
-
AI uses the retrieved content and conversation context to generate a useful suggestion.
-
The suggestion is presented for review and manual use.
Value
AI KB Suggestions improve response accuracy and reduce search time. They help agents find approved information quickly, maintain consistency across conversations, and reduce the risk of providing unsupported or outdated answers. The same knowledge-backed approach can also improve summaries, refined responses, and quality outcomes because conversations are handled with more accurate information.