After every resolved conversation, customers can rate their experience. These ratings flow into your Reviews dashboard where you can track satisfaction, read feedback, and spot patterns. This guide covers how ratings work from the customer's perspective and how to use the dashboard.
How Customers Rate a Conversation
The rating prompt appears automatically when an agent resolves a conversation. The customer sees a full-screen overlay inside the chat widget with three steps:
1. Star Rating (Required)
Five stars with labels that change as the customer hovers:
| Stars | English | Spanish | French |
|---|---|---|---|
| 1 | Poor | Malo | Mauvais |
| 2 | Fair | Regular | Passable |
| 3 | Good | Bueno | Bien |
| 4 | Great | Muy bueno | Tres bien |
| 5 | Excellent | Excelente | Excellent |
The labels are automatically translated based on the widget's language setting. Stars animate with a scale effect and glow when hovered.
2. Category Chips (Optional)
After selecting a star rating, category chips appear. Customers can tap one or more to describe why they gave that rating:
- Helpful — the response solved their problem
- Knowledgeable — the agent/AI knew the subject well
- Fast response — quick reply times
- Friendly — good tone and personality
- Accurate — the information provided was correct
These categories are also translated for Spanish, French, and Russian speakers.
3. Written Comment (Optional)
A text field lets the customer leave a free-form message (up to 500 characters). This is where you get the most actionable feedback — specific praise or complaints.
Skip Option
Customers can skip the entire rating by clicking Skip. The conversation still resolves — they just don't leave feedback. No data is recorded.
When the Rating Prompt Appears
The prompt is triggered when:
- An agent resolves the conversation from the dashboard — the widget receives a real-time event and immediately shows the rating overlay
- The customer closes the widget while a resolved conversation is active
The rating overlay covers the entire widget with a subtle blur backdrop. The customer must either rate or skip before they can start a new conversation.
The Reviews Dashboard
Navigate to Reviews in the sidebar to see all customer feedback. This page requires Admin or Owner access.
Summary Cards
Two cards at the top give you the big picture:
- Average Rating — your overall score out of 5 stars (rounded to one decimal). Shows "—" if no ratings exist yet.
- Total Reviews — how many conversations have feedback (either a star rating or a written comment). Also shows how many have a star rating specifically.
Recent Feedback List
Below the cards, a paginated list of all reviews. Each entry shows:
- Star rating — 1–5 filled stars (gold/amber colored)
- Category badges — the chips the customer selected (e.g., "Helpful", "Fast response") shown as indigo-colored pills
- Timestamp — when the rating was submitted
- Written comment — the customer's feedback message (if they wrote one)
- View conversation — a link that takes you directly to the conversation where this rating came from, so you can read the full context
Pagination
Reviews are shown 20 per page with Previous/Next navigation. The most recent feedback appears first.
Ratings in Other Places
Ratings also appear in:
- Conversations list — the Rating column shows the star count for each conversation
- Conversation detail panel — when you open a rated conversation, the rating, categories, and comment are displayed in the panel header
- Overview dashboard — the "Average Rating" stat card shows your overall score
Tips for Better Ratings
- Resolve conversations promptly — the faster you resolve, the fresher the experience is when the customer rates. A quick resolution while the positive feeling is still fresh leads to higher ratings.
- Read the written comments — star ratings tell you how customers feel; comments tell you why. A 3-star rating with "Had to repeat my question three times" is more actionable than just seeing "3 stars".
- Look for category patterns — if "Fast response" is rarely selected but "Helpful" is common, your answers are good but speed could improve. If "Accurate" drops, check your knowledge base.
- Click "View conversation" — when you see a low rating, always read the full conversation. It's the fastest way to identify what went wrong and whether it's an AI training issue or a process problem.
- Train your AI with feedback — if customers consistently say the AI gave wrong answers about a topic, update your documents or add specific Q&A content to your knowledge base.
- Don't fear low ratings — they're a gift. A 2-star review with a comment is more valuable than no feedback at all. It tells you exactly where to improve.