Filtering
The Filtering feature in Sessions lets you quickly find and analyze specific types of conversations — whether you’re tracking conversions, testing new flows, or identifying areas for improvement.
It helps you zero in on what matters most: performance trends, missed goals, or patterns in user sentiment.
Where to Find It
You can find the filtering panel at the top-right corner of your Sessions dashboard — right beside the search bar.
Simply click Filter to open the filtering options.
Available Filters
Filter Type | Description |
|---|---|
Favorites | View sessions you’ve bookmarked or marked for review. |
Goal Conversions | Show sessions where users completed your defined goals (e.g. “Book a Demo”). |
Principles Fixed | Displays sessions where your AI adjusted behavior based on feedback or moderation principles. |
Missing Knowledge | Find sessions where your AI didn’t have enough knowledge to respond confidently — perfect for improving your Knowledge Base. |
Negative Sentiment | Isolate sessions where user tone or feedback was detected as negative, so you can improve tone or content. |
How to Use Filters
Navigate to Sessions.
Click the Sessions button once again and the filters appear.
Select one or multiple filters (e.g. “Goal Conversions” + “Negative Sentiment”).
The filtered sessions will appear in the left-hand list.
Click any session to review its full transcript and insights.
Pro Tip:
Combine “Goal Conversions” with “Version” or “Date Range” filters to compare performance across time or Agent versions.
Search and Filter Together
In addition to filters, you can search directly by:
Message text
Session ID
User name
Use this when you need to locate a specific conversation or debug a particular user case.
Why It Matters
Filtering transforms your session list into a real-time performance analysis tool.
It helps you:
- Identify broken flows quickly
- Understand where your AI struggles (Missing Knowledge)
- Measure tone and sentiment over time
- Find best-performing conversation styles
Best Practice Tip
Always review “Missing Knowledge” sessions weekly — they’re the fastest way to identify new Q&As or articles your AI needs to learn from