Theme
Agent Guide
Agent Assist provides real-time AI suggestions directly within your agent desktop while you handle customer conversations. As customers speak or type, the system captures the transcript, runs it through a RAG (Retrieval-Augmented Generation) pipeline, and surfaces relevant answers, sources, and follow-up questions so you can respond faster and more accurately.
Capabilities
| Capability | Description |
|---|---|
| AI Suggestions | Context-aware answer cards generated in real time from your organization's knowledge base. |
| Source Citations | Every suggestion includes ranked source documents with page numbers, snippets, and relevance scores so you can verify accuracy before responding. |
| Knowledge Search | A search bar that lets you manually query the knowledge base when you need information beyond what auto-suggestions provide. |
| Quick Replies | Pre-generated short responses and follow-up questions you can send with a single click. |
| Feedback | Thumbs-up and thumbs-down controls on answers and individual sources that feed back into the system to improve future results. |
Where to Go Next
| Page | What You Will Learn |
|---|---|
| Getting Started | How to access the Assist panel, understand the layout, and work with your first suggestions. |
| AI Suggestions | How suggestions are generated, what they contain, and how to use them effectively. |
| Suggestion Cards | Anatomy of a suggestion card, streaming behavior, and timing indicators. |
| Source Citations | How to read and evaluate the source documents attached to each suggestion. |
| Knowledge Search | How to manually search the knowledge base for answers. |
| Quick Replies | How to use pre-generated reply chips and follow-up questions. |
| Feedback | Why feedback matters and how it improves the system over time. |
| Answer Feedback | How to rate overall answer quality. |
| Source Feedback | How to rate individual source documents. |