This article explains how to measure the performance of your conversational agents using Novus. Tracking agent interactions helps you explore adoption, view real prompts and responses, and understand how visitors engage with your application.
You can use Novus to automatically detect and set up tracking for conversational agents in your application.
Note: Novus automatically scans your application to find conversational interfaces. During this scan, Novus locates the text input fields, identifies the submit buttons, and determines the underlying model your agent uses. 3. Review the detected agents and choose the ones you want to track. 4. Complete the setup to generate the tracking code.
Note: Novus automatically writes the exact tracking code needed to capture visitor prompts and agent responses. Novus then creates a pull request in your repository so your team can review and merge the changes.
After your tracking code is merged, you can monitor how visitors interact with your conversational agents.
Note: Novus continuously processes incoming interaction data. Novus calculates the total number of conversations, determines your visitor adoption rate, and tracks active visitors over specific time periods.