Swifty AI’s Insights dashboard helps you understand how users interact with your knowledge base through search and chat. By reviewing resolved and unresolved interactions, you can identify gaps, optimize your content, and improve your customers’ experience.

In this article, you’ll learn:

• How to access Swifty AI Insights
• The difference between search and chat interactions
• How to review resolved vs. unresolved interactions
• Where to find detailed interaction history
• Best practices for analyzing interaction data

Accessing Swifty AI Insights

1. From your Helpjuice dashboard, navigate to Analytics in the left sidebar
2. Select Swifty AI (Beta).

Understanding the Interactions Table

The table displays all user interactions currently tracked with Swifty AI.

• Source: Whether the interaction came from Search or Chat.
• User: Identifies who asked the question (internal users, or external users).
• Analysis: Shows if the interaction was Resolved (green) or Unresolved (red).
• Time: Displays when the interaction occurred.

You can also:

• Sort interactions by source or user.
• Use the pagination at the bottom to move through larger lists.

Filtering Data

You can filter results to focus on a specific time range:

• Last 7 Days
• Last 30 Days
• All Time
• Custom Period

This helps you spot short-term trends or analyze long-term patterns in how users interact with Swifty AI.

Viewing Interaction Details

1. Click on any interaction from the list.
2. The Interaction Overview panel (on the right side) will display the chat or search history.
3. Review the full context of the question and Swifty AI’s response.

Use Cases

Here are a few ways teams use Insights data in practice:

• Identifying missing articles: If users frequently search for “reset password” and interactions are unresolved, you can create or update a password reset guide.
• Improving article clarity: If a question shows as unresolved but an article exists, review the wording -  the content may need simpler phrasing or better keywords.
• Spotting training needs: If many unresolved questions come from internal users, it could signal that your team needs guidance on where to find resources.
• Tracking content impact over time: By filtering to “Last 7 Days” vs. “Last 30 Days,” you can see whether recent updates reduced unresolved interactions.

Customizations

This section is not managed by frontend customizations. If there's any option you'd like to see on that page, feel free to submit a Feature Request on our Community Forum.

Best Practices

• Review Unresolved interactions regularly to identify missing or unclear content.
• Use date filters to track both recent issues and long-term trends.
• Update knowledge base articles linked to recurring unresolved questions.
• Monitor both search and chat activity for a complete picture of user needs.