Skip to content

Conversation Details (Analytics)

The Conversation Details page is where you review a single conversation end-to-end: - metadata (time, participants, channel, agent/team), - transcript or message thread, - AI insights (values + explanations).

What you’ll typically see

Conversation metadata

Common metadata fields include: - date/time, - agent / owner, - channel (call/chat/email/ticket), - duration (for calls), - queue or routing information.

Transcript / thread

For calls, this is the transcript (and sometimes audio playback). For text channels, this is the message thread.

Analytics / Insights section

This section displays AI insights extracted from the conversation, often grouped into categories such as: - CX metrics (CSAT, NPS, churn risk), - Conversation understanding (summary, sentiment, topics), - Sales metrics (lead score, objections, next actions), - QA (Auto QA scores and details).

The UI may also display threshold labels (for example “Dissatisfied”) alongside numeric values.

  1. Start with the summary (if enabled). It gives context quickly.
  2. Review key metrics (for example CSAT, sentiment, outcome).
  3. Read explanations for any metrics you plan to act on.
  4. Validate with the transcript/thread when the outcome is important (coaching, QA, escalations).
  5. Take action using your org’s workflow (tag, note, coaching follow-up, escalation).

Call Details view

The Call Details page includes several tabs:

  • Call Details – Basic call information and metadata
  • Analytics – AI-generated insights and explanations
  • Transcript – Full conversation transcript
  • QA – Quality assurance scores (if Auto QA is enabled)
  • Shared access – Sharing settings
  • Notes – User notes

Call Details - Call Summary

Figure: Call Details showing AI-generated call summary with key highlights.

Analytics tab

The Analytics tab displays all AI-generated insights organized by display groups (e.g., "CX Metrics", "Sales Metrics"):

Call Details - Analytics tab

Figure: Analytics tab showing insight values with explanations.

Each insight shows:

  • The metric name
  • The extracted value
  • An explanation citing evidence from the conversation