Skip to content

Topic Analysis

Topic Analysis identifies what the conversation is about (topics/themes) to help you understand demand drivers and operational issues.

  • Dropdown (primary topic) or multi-select (multiple topics)

MiaRec supports multi-label topic classification—a conversation can be tagged with multiple topics. This provides richer analysis but requires careful taxonomy design to avoid overlap.

Recommendation: Start with a primary topic field for stable dashboards. If needed, add multi-topic support later as your taxonomy matures.

EDITOR NOTE: multi-topic support

Ask product/engineering:

1) Can AI Tasks populate multi-select fields? - A) Yes - B) No

2) Preferred model for most customers: - A) Primary topic only - B) Multiple topics - C) Primary + "Other topics" text field

Best-guess recommendation:

  • Start with primary topic for stable dashboards; add an optional "Other topics" text field later.

Configuration overview

  1. Define topic taxonomy (start small; expand)
  2. Create Custom Field(s)
  3. Enable or create Topic Analysis AI Task
  4. Build dashboards and saved searches:
  5. topic distribution over time
  6. drilldowns to conversations per topic

Prompt recommendations

  • Provide allowed topic list
  • Require selection only from list
  • Provide “Unknown”
  • Provide explanation based on evidence