Skip to main content
At the end of this page, resolved conversations in your chosen inbox ask end users for a rating, and you know where to read the results.

​
Before you start

Before you start, sign in to the dashboard and select the organization you want to work in.
  • CSAT is an inbox setting, and inbox settings are Admin-only.
  • The inbox needs live chat traffic with at least one human agent reply. See Set up inboxes if the inbox is new.

​
Turn on the CSAT survey for an inbox

CSAT is off by default. Each agent has its own inbox, so you turn the survey on inbox by inbox.
1

Open Conversations, then select Settings at the bottom of the views rail

The Conversations settings area opens with Inboxes in the settings nav.
2

Select the inbox row, for example the Support inbox

The inbox detail page opens with the tabs General, Members, Routing, Availability, and Analytics.
3

Open the Analytics tab and find the Customer experience card

The card holds three toggles. The first one is CSAT survey.
4

Turn on CSAT survey

Its description reads “Send a 5-star satisfaction survey after an agent resolves the conversation.”
5

Select Save changes on the Unsaved changes bar

A toast confirms Inbox settings updated. Select Discard instead to throw the change away.

​
The Edit inbox shortcut

Admins can also flip the survey without leaving the inbox. Hover an inbox in the views rail, open its menu, and select Edit inbox. The Edit inbox dialog includes a CSAT Survey switch with the subtitle “Send satisfaction survey after resolving”. Select Save to apply it.

​
When the survey sends

The inbox sends one survey per conversation, at the moment the conversation is resolved. Two conditions must both be true:
  • The CSAT survey toggle is on for the inbox.
  • A human agent took part in the conversation at some point.
The resolver can be the human agent or the AI agent. A conversation the AI agent handled alone never gets a survey, even in an enabled inbox. If a human agent resolves and then reopens the conversation before the end user answers, the unanswered survey disappears from the thread. For the Resolve conversation dialog and its optional labels and reason, see Resolve a conversation.

​
What the end user sees

The survey appears as a card inside the same conversation, right after the conversation closes. It reads “This conversation has been closed” and asks how the conversation experience was. Below the question sit five faces, from Very poor to Excellent, which map to a rating from 1 to 5. Selecting a face reveals an optional feedback box with the placeholder “Let us know how we can improve…”, plus Submit Feedback and Cancel buttons. After submitting, the card shows “Thanks for your feedback!” and the note “Your response helps us improve.” below it. A footer reads “Still have an issue?” with a Chat with us link, so the end user can keep talking.

​
Where ratings appear in the inbox

Human agents see the survey and its result inside the conversation:
  • In the thread, a system chip reads CSAT survey sent when the survey goes out.
  • Once the end user answers, a Customer Satisfaction Survey card appears in the thread with the face, the rating such as (4/5), and any feedback text.
  • In the conversation list, the card shows an amber star with the rating next to the last activity time. Hover it to read the rating and the feedback.
The rating is not shown in the details rail on the right. Open the thread or the list card to read it.

​
Where ratings appear in analytics

Open Analytics, then the Team tab. The headline card leads with the average CSAT rating, the number of Responses, and Low ratings (1–2). The CSAT by resolver card splits responses between AI resolved and Human resolved and draws the rating distribution. The per-agent table shows CSAT for each human agent. See Team performance for the full walkthrough. CSAT also appears as a column in the Overview tab channel table, and as the CSAT responses and CSAT average measures in the Explore tab. The command palette entry CSAT reports jumps to the Team tab.

​
Verify

Ask another member to start a conversation as an end user, reply to it as a human agent, and resolve it. The end user sees the survey card in the conversation. After they submit a rating, the thread shows the Customer Satisfaction Survey card and the list card shows the amber star.