An Overview of the SUVA Chat Window

This article details the structure and elements of the SUVA chat window. The appearance of the chat window is configured from Theme Editor.

NOTE.

To request customizations beyond those available in the Theme Editor, please reach out to the SUVA team.

The key elements of the chat window are:

1. User Input. User queries are entered in this section.

2. Send. Pressing this button posts the user query.

3. Minimize. Clicking on it minimizes the chat window.

   

4. More Options. Clicking shows more options including:

   • Settings. Selecting this opens the languages section where you can select the language(s) of your choice.

   • Email Chat. Clicking this list option allows users to send the email transcript to their email addresses.

   • Clear Chat. Clicking the third option from the list clears the chat history while the chat session continues.

   • Close Chat. Selecting this option closes the chat window and ends the chat session.

   

5. Citations in LLM-Generated Responses. If a user wants to know about the source of the LLM response, they can use the citation icon in the chatbot response. Hovering over the icon shows the references used to generate the response.

   

6. Auto-Populate Case Details. When a user moves to create a case (case creation is enabled via Salesforce), a form including the Subject and the Description fields shows up. Based on the conversation, the subject and the description fields are auto-populated. All you have to do is verify the details, make edits, if required, and Submit the case.

   NOTE. This feature works only when LLM integration is activated and the LLM is assigned to the respective virtual agent.

   An illustration is as below:

   

7. User Feedback. A mechanism to analyze the perceived usefulness of chatbot responses and conversations.

   When a user ends a session by closing chat window, an emoticon-based feedback mechanism pops up to capture the overall experience of the user.

   

   For traditional and LLM-powered chatbots, a thumbs-up and thumbs down feedback pops up along with the chatbot responses (as shown in the image below). In LLM-powered chatbots, this feedback option is available for each bot response. For traditional bots, it appears at the end of a story (configured in the Story Board).

   

8. Rich Snippet Responses. Rich Snippet responses–text, image, video, or URL responses–are enabled from search clients and are visible only if they are active. Given below are some examples of how rich snippet responses appear in the SUVA chat window.

   A) Text with References

   A text response along with the citation () icon is shown in the image below. Hovering over the icon shows the references used to generate the response.

   

   B) Image Responses

   

   C) Video Response

   

   D) Text (With Steps) Response