MINNEAPOLIS — Staffing solutions corporation Meridian Group has replaced its 340-page employee handbook with a ChatGPT prompt that begins "You are the HR department of Meridian Group" and ends with "be supportive but noncommittal," according to documents obtained by this publication.

The transition occurred without formal announcement. Employees discovered the change when the company intranet's "Policies & Procedures" section began redirecting to a chat interface that cheerfully answered questions about benefits, conduct standards, and termination procedures—often incorrectly.

"I asked about our parental leave policy and it told me we offer 16 weeks," said account manager Diane Prescott. "Our actual policy is six weeks. When I pointed this out, it apologized and said it was 'aspirational guidance.'"

Human resources director Kevin Mallory defended the system. "Employees never read the handbook anyway," he said. "Now they have an interactive experience that provides personalized answers. The fact that some answers are wrong is no different from when human HR gave wrong answers. At least the AI is polite about it."

The AI handbook has generated particular confusion around the company's dress code, which it describes differently to each employee based on conversational context. One engineer was told business casual was required; another was informed that "creative expression through fashion is encouraged." A third received a 400-word essay on the history of workplace attire that concluded with no actionable guidance.

Legal counsel has expressed concern that an AI-generated handbook may not constitute a binding policy document, a distinction that could prove relevant in future employment disputes. Mallory dismissed this concern, noting that "no one reads legal disclaimers either."

The company projects annual savings of $180,000 from eliminating the handbook's maintenance and distribution costs. Those savings will fund the ChatGPT Enterprise subscription.