Description

Create a new Rovo Skill to support continuous improvement of Rovo agents by analyzing user conversations and generating privacy-preserving, agent-focused improvement logs. These logs must not expose or store verbatim user messages or identifying information. Although the skill need not apply improvements automatically, it must reliably produce structured insights from conversation patterns, enabling developers to review and iteratively enhance agent performance.

This feature helps teams improve their Rovo agents by highlighting patterns, sentiment trends, and structural improvements from user interactions, while complying with strict organizational data protection requirements.

Problem Statement
Currently, Rovo agents lack a built-in system for continuous, conversation-driven improvement. Conversation logs contain sensitive or personal information, so they require manual anonymization before use in analytics or training.
Teams need a way to:

• Understand user struggles
• Identify gaps in agent knowledge or behavior
• Detect negative sentiment or repeated friction points
• Highlight opportunities to improve intents, summaries, or tool usage

Without reviewing raw conversation transcripts or asking employees to reveal detailed user dialogue, which may breach privacy or compliance policies.