Skip to main content

Intent and Topic Analysis

Understand what your users are trying to do — and whether your agents are meeting those needs.

What This Section Covers​

Optimly classifies each user message by intent and topic, giving you clarity on what users are asking for, how often, and how well your agents respond to those needs.

This level of insight is essential for improving onboarding, documentation, support flows, and product design.

Key Concepts​

Intent Detection​

Each message is analyzed to determine the purpose behind it. Common intents include:

  • Question – The user is asking for information
  • Feedback – Positive or negative comments about the product or service
  • Complaint – An issue or frustration is being expressed
  • Action Request – The user wants the agent to do something (book, generate, send, etc.)
  • Clarification – The user is unsure or wants the agent to repeat something
  • Chitchat – Casual or off-topic interaction

Intent detection is automated and updated in real time as conversations evolve.

Topic Classification​

Messages are also categorized into custom topics based on their content and your business context.

Common categories include:

  • Pricing
  • Technical Support
  • Product Features
  • Billing
  • Onboarding
  • Policies and Terms
  • Account Management

Topics can be customized per account or vertical. One message may belong to multiple topics.

Visualizations and Metrics​

  • Top Intents by Frequency
  • Topic Distribution per Agent or Time Period
  • Unanswered Intents
  • New or Emerging Topics
  • Conversation Heatmaps by topic over time

Use Cases​

  • Identify high-volume intent categories for better documentation
  • Understand when and where confusion occurs (e.g., onboarding vs. pricing)
  • Detect new support needs that aren't covered in current knowledge
  • Track how well agents handle specific types of requests
  • Use topic data to segment user behavior

Next: Response Quality