AI-Powered People Analytics: From Descriptive to Prescriptive HR

Most HR teams today are fluent in "descriptive analytics." They can tell you what the turnover rate was last quarter. Some are moving into "diagnostic analytics," figuring out why it happened (e.g., exit interviews pointed to manager issues). But the frontier the space where HR becomes truly strategic is Prescriptive Analytics.

By 2025, advanced machine learning models are ingesting thousands of data points that were previously ignored: collaboration patterns in Slack, the sentiment of language used in emails, the frequency of 1-on-1 meetings, and even calendar density. These models don't just report history; they predict the future. They can identify an "at-risk" employee months before they quit, based on subtle changes in their digital behavior; perhaps they stop participating in team chats or their response times slow down.

Vendors like Eightfold, Visier, and Glint (now part of Microsoft) are pioneering this space. The AI doesn't just flag the risk; it prescribes the intervention. It might suggest to a manager: "Based on engagement patterns, your direct report, Sarah, is showing signs of burnout. We recommend she take a mental health day, or you could reassign one of her current projects to free up her bandwidth."

However, this power comes with immense responsibility. In 2026, the biggest debate in HR will be about the ethics of this "bossware." Employees are wary of being monitored. The successful HR teams will be those that are radically transparent about what data is collected, how it is used, and how it benefits the employee (by preventing burnout) rather than just surveilling them. Prescriptive analytics is a tool for care, but only if wielded with consent and trust.