Skip to content

Getting Started

If you’re new to Conversation Analytics, this page shows a quick “first session” path to get value right away.

10-minute quick start

1) Start with a dashboard

  1. Open the Dashboards area.
  2. Choose a dashboard your organization uses regularly (for example: CSAT, Top Issues, Auto QA).
  3. Set a time range (for example: last 7 days).

What to look for - Trends over time (up/down changes). - Distribution “buckets” (for example: satisfied vs dissatisfied). - Spikes in negative sentiment, escalations, or specific topics.

2) Drill down to the underlying conversations

Many dashboards support drilldowns. 1. Click a bucket (for example: Dissatisfied). 2. You should land on a filtered list of conversations that match that bucket.

3) Open a conversation and review the insights

  1. Open one conversation from the list.
  2. Go to the Analytics / Insights section.
  3. Review:
  4. the score/value (for example CSAT = 2),
  5. the explanation (why the AI chose that score),
  6. supporting context in the transcript/thread.

Use Search to narrow down patterns. Examples: - “CSAT < 3 AND Topic = Billing” - “Competitor Mentioned AND Deal Stage = Negotiation” - “Auto QA score < 80 AND Escalation Reason present”

5) Save or share your view (if enabled)

If your organization uses saved views: - save the filter as a Saved Search / Saved View for recurring review, - share the dashboard or saved view with your team.

  • Read the explanation before you judge the score. It often points to specific moments in the transcript/thread.
  • Use drilldowns instead of random sampling. You’ll find patterns faster.
  • Look for consistency. One bad call is an outlier; a bucket trend is a process problem.