The upgraded system introduces three core components designed to automate deep research and trend identification. The enhanced AI Analyst now features auto-execution, where users provide a natural language prompt, and the system independently determines the necessary research parameters, data filters, and multi-modal outputs. This agentic approach builds upon a foundation that has already processed over 4 million hours of research.

Alongside the analyst tool, the company has overhauled its Topic Discovery and Real-Time Trends modules. Topic Discovery uses large language models to generate a hierarchical taxonomy of customer calls, allowing administrators to refine categories through a human-in-the-loop process. Meanwhile, the Real-Time Trends system monitors for minute-level spikes in conversation volume. It is specifically built to capture events outside of established categories—such as unexpected product outages or regional disruptions—providing executives with auto-generated root-cause explanations and proactive email alerts. CEO Ping Wu noted that these enhancements aim to eliminate the burden of manual discovery, enabling systems to flag critical issues that businesses might not even know to look for.