Skip to main content
Skip to article
  • Qualtrics Platform
    Qualtrics Platform
  • Customer Journey Optimizer
    Customer Journey Optimizer
  • XM Discover
    XM Discover
  • Qualtrics Social Connect
    Qualtrics Social Connect

Conversational Data in Document Explorer (Studio)

Was this helpful?


This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

The feedback you submit here is used only to help improve this page.

That’s great! Thank you for your feedback!

Thank you for your feedback!

About Conversational Data

Conversational data includes transcribed conversations between representatives (or bots) and customers on audio and digital channels. When opened in the document explorer, you’ll see visualizations and analysis specialized for conversational data.

Feature Availability

Voice interactions (including audio playback) are only available for accounts that use XM Discover’s Voice Transcription Service to blend Contact Center interactions with other customer feedback.

Chat interactions are available for accounts that upload digital interaction data through XM Discover’s chat and messaging connectors (including Qualtrics Social Connect) or generic interactions connectors.

Contact your Discover representative to learn more about these options.

Conversational Feedback in Document Explorer

When you view conversational feedback in the document explorer, you’ll have a few additional options.

conversational data in document explorer

  1. Choose which conversation participants to display. Here, we’ve selected “All Participants.”
  2. For voice interactions only, you’ll see the following metrics:
    • Silence: The percentage of silence in the call. By default, silences less than 2 seconds are not counted.
      Qtip: For documents that use the Qualtrics conversational format, this metric is based on the CB % Silence enrichment attribute. If a document is not in the Qualtrics conversational format, silence percentage is not calculated this way and is not available as an attribute.
    • Longest: The longest uninterrupted silence in the call. Click this value to jump to the sentence spoken right before the silence.
    • Agent: The percentage of the call that the representative was speaking. This is calculated when you view the document.
  3. Play the audio, pause it, or jump to any part of the conversation.
  4. The graph to the left is called a spine graph. It gives a visual summary of the conversation, which you can click to jump to key points.
Qtip: You can customize the way interactions display in Account Settings with Conversation Display.

Spine Graphs

Spine graphs provide a visual summary of an entire conversation. The conversation starts at the top of the graph and ends at the bottom. Click a bar to jump to any point in the interaction. Clients are on the left, and agents are on the right.

spine graph

Qtip: Use red bars to pinpoint problem areas.
Qtip: Hover over the bars or enrichment icons to skim sentences, enrichments, and voice interaction events like Overtalk, Silence, and Response gaps.

These are the columns displayed:

  • Sentiment: Each bar represents a sentence, either spoken by the client or an agent. Each client and agent bar is colored according to the sentiment, with height determined by sentence’s length.
    Qtip: Each bar may also represent multiple sentences instead of just one, if the sentences fall within the same enrichment metric band.
    Qtip: Gray is neutral, but blue is no sentiment detected.
  • Effort:
    • Orange circle: A sentence with Hard or Very Hard Effort score.
    • Blue circle: A sentence with Easy or Very Easy Effort score.
    • No circle: A sentence with a Neutral Effort score.
  • Emotional Intensity: A circle represents a sentence with emotional intensity, with the shade of purple getting darker for higher intensity.
  • System default: This column shows key sentence types. A gray diamond will appear if any of the following are detected:
    • Client mentions Tenure, Churn, Cry for Help, Not Recommend.
    • Representative mentions Don’t Know, Disclosure, Mini-Miranda, Transfer, Hold.
    • Anybody mentions Request.
  • Intelligent Scoring: Shows criteria pass (green checkmarks), fails (red X), and automatic failures (crossed out red circle).
    Qtip: Use these to quickly jump to coaching opportunities.

Automatic and Manual Scrolling

For voice interactions, you can scroll freely through the dialogue. If you scroll while the audio is playing, the Resume Automatic Scrolling button will let you get back to synchronized scrolling.

How Participant Types are Identified

XM Discover identifies conversation participants using a combination of these attributes:

  • CB Participant Type: Identifies participants as representatives (agents), clients, or unknown.
  • CB Kind of Participant: Further identifies representatives as humans, chatbots, or IVR (Interactive Voice Response) bots.
  • Participant ID: Identifies individual participants within a conversation.

This is how participants are visualized in document explorer:

Icon Participant Side of conversation
human silhouette wearing headset Agent Right
robot Bot Right
human head silhouette without headset Client Left
question mark icon Unknown Left

 

Qtip: You can customize participants’ names in Account Settings with Conversation Display.

Definition Conversation Participants with Attributes

You can define participant attribute values when loading conversational data into XM Discover via file-based connectors (Ad-Hoc File Upload, Files, XM Discover Link).

Data Format/Attribute CB Participant Type CB Kind of Participant
Digital interactions All files: use the participantType field. All files: use the is_bot field.
Call transcripts Excel and CSV: use the participantType field.

JSON: use the type field inside the participants object.

 All files: use the is_ivr field.
Qtip: You can visualize more than 2 participants within an interaction. For example, you can have agent 1, agent 2, bot 1, bot 2, client 1, and client 2. Participants with unique IDs of the same type are each broken out in the interface on their side of the conversation, and you can display or hide dialogue from different participants.
all participants dropdown

FAQs