What happened
Cloverleaf Labs ran five leading large language models through a set of workplace conflict scenarios and found that, across 75 conversations and 638 separate recommendations, just three urged the employee to make a real effort to repair the relationship, Contxto reported. All three came up in cases where a manager, not an employee, was the one complaining about a direct report.
To count as relational repair, advice had to meet four tests set by Cloverleaf: prompting curiosity about the other person, reflection on one's own conduct, stating personal needs and openness to feedback. Kirsten Moorefield, the company's co-founder and Chief Strategy Officer, said what surprised her most was that every model suggested preparing an exit, such as considering quitting or seeking a different team.
Moorefield said she did not expect a resulting wave of resignations. Instead she warned of rising disengagement and fear that would quietly wear down collaboration. She cited McKinsey research indicating that heavy AI users are 10% likelier to quit within three to six months. She also argued that the tools give better guidance when supplied with personality data about the people involved.
Context
Employees already turn to chatbots for help with difficult bosses and unreliable colleagues, while employers are building AI into workflows, staff support and decision-making. Cloverleaf Labs pairs personality science with AI-driven workplace coaching. By Moorefield's count, manager engagement has fallen sharply, from one in three engaged a few years ago to four in five now disengaged, and some staff worry about training AI tools that could replace them.
The briefing
The week in tech, in one email.
The stories worth your time, every week. Unsubscribe anytime.
