Filtering and Reviewing Query Analytics
Summary
The Analytics page includes powerful tools that allow you to filter and review user questions and the corresponding AI-generated answers. You can narrow down the data displayed in your dashboard based on timeframes and specific data attributes, and then drill into individual threads to audit answer quality.
CleanShot 2026-01-05 at 10.34.37.mp4 3024x1964
Reviewing Query Threads
Use the Threads tab in Analytics to review questions asked by users and the answers generated by the AI.
Steps
- From the main navigation menu on the left, click on Analytics.
- Make sure the Threads tab is selected (this is selected by default), which displays a list of recent user queries.
- Click on any query in the list to expand it and view more details about that interaction.
- A Generated answer section will appear directly below the query, showing the exact response the user received.
- Review the answer for accuracy and completeness. The system can provide detailed responses, including lists, breakdowns of information, and step-by-step instructions.
- To view another thread, click the current query again to collapse it, and then select a new one from the list.
Why this matters
Regularly reviewing queries and generated answers helps you:
- Monitor your AI’s performance
- Verify that responses are accurate, complete, and on-brand
- Identify gaps or outdated content in your knowledge base
Date Range & Time Zone Logic
The date picker on the analytics page will display the customer’s local time zone. This is achieved using the code below:
const timezone = Intl.DateTimeFormat().resolvedOptions().timeZone;
The value returned by Intl.DateTimeFormat().resolvedOptions().timeZone comes from the system time zone settings of the user’s device. It uses the browser or the runtime environment’s Internationalization API (Intl) to retrieve the current time zone information.
However, for consistency and alignment, all analytics values are stored and filtered using the UTC time zone.
Example
from: "2025-02-24T15:00:00.000Z"
timezone: "Asia/Tokyo"
to: "2025-03-04T14:59:59.999Z"
stand for Tokyo time:
02/04/2025, 00:00:00.000
03/05/2025, 11:59:59.999
Step-by-Step: Filtering Analytics Data
- Navigate to Analytics: Open the Analytics section from the main sidebar.
- Set Date Range:
- Click the date picker in the top right corner.
- Choose a preset range (e.g., Last 7 days) or select Custom to define a specific start and end date.
- Apply Filters:
- Locate the filter bar above the search box.
- Click Select a value (or the + Add Filter button) to see available categories such as Source, Action, Status, or Feedback.
- Select a category (e.g., Source) and then choose a specific attribute (e.g., Agent Assist) to refine the list of threads displayed.
Key Filter Options
- Source: Filter by the origin of the query (e.g., API, Slack, Agent Assist).
- Status: View threads based on their current state (e.g., Answered, Unanswered).
- Action: Filter by specific user or system actions taken during the interaction.
How Date Filters Affect Analytics
- The selected date range is interpreted in the user’s local time zone in the UI, but queries are executed against UTC timestamps.
- This ensures that all analytics aggregations (counts, trends, etc.) are consistent across different regions while still appearing in the user’s local time.
- When debugging or integrating with APIs, use the UTC
fromandtovalues shown in the request, along with thetimezonevalue, to understand how the UI range maps to stored data.
Summary
In summary, the Analytics section provides a straightforward way to:
- Audit your AI’s performance by reviewing user queries and generated answers in the Threads tab
- Narrow down threads by date, source, status, and actions using powerful filters
- Understand how local time zone selection maps to UTC-based analytics storage
By regularly reviewing and filtering query analytics, you can gain insights into customer needs and verify the effectiveness of your knowledge base and AI configuration.