Talent Sentiment Agent Insights
Overview
The Talent Sentiment Agent uses a large language model (LLM) to generate insight cards. The agent bases these insights on the scores and comments that employees provide during the current survey round.
Insights adhere to these key standards:
- All summaries and facts derive from employee comments and scores.
- The agent ranks insights so the most critical information appears first.
- The agent delivers content in a tone suitable for a corporate environment.
Insight Generation Process
The agent follows these steps to generate insights:
- Analyzes the emotional tone (positive, negative, or neutral) of every employee comment and score.
- Identifies critical themes, entities, and specific pain points within the feedback.
- Combines these elements into a ranked list of insights based on urgency and impact.
- Displays insights as cards containing a title, a concise summary, and supporting data.
Conversational Capability
Managers who have access to
Chat with Talent Sentiment Agent
can use the Explore
button to start a conversation with the agent. In the conversation, they can:- Ask the agent in natural language why it generated a specific insight.
- Request deeper analysis on specific trends or categories.
- Ask how the agent ranked the importance of an insight.
Data Archival
Workday Peakon Employee Voice automatically archives conversations with the agent into a read-only state 24 hours after the last interaction. Archived conversations remain accessible for future reference, and managers can start new conversations on the same insight at any time.