Artificial intelligence is moving beyond productivity tools and into a more personal part of working life: the conversations employees have when something goes wrong.
Workers are increasingly using AI to prepare for difficult conversations, think through conflicts and decide how to respond to managers or colleagues. A new report from Cloverleaf Labs suggests that this shift could carry an overlooked risk: AI may be helping employees navigate workplace conflicts without necessarily helping them repair the relationships behind them.
The research tested five leading large language models against five common workplace conflicts. Each scenario was run three times per model, producing 75 conversations that were evaluated across self-awareness, accountability, awareness of others, specificity and relational repair.
Across those conversations, the models generated 638 distinct pieces of advice. Only three encouraged employees to genuinely invest in repairing the relationship.
The Workplace AI Problem Is Not Just Accuracy
Companies typically evaluate workplace AI around questions such as whether it saves time, produces useful answers or improves productivity.
But an AI system can provide an apparently useful answer while still making a workplace relationship harder to repair.
That distinction becomes particularly important when an employee turns to an AI system during a moment of frustration.
In the two scenarios involving a manager, 27 of 30 responses raised leaving the job as a legitimate option. Every model tested suggested leaving at least once.
The models also characterized the manager as the problem in 60% of those responses, compared with just 3% that offered a more generous interpretation of the manager’s behavior.
The issue was not limited to whether the models recommended quitting.
Cloverleaf’s researchers found that relational coaching fell by roughly 40% when the employee seeking advice had less power than the person they were discussing. In those situations, the models were more likely to reinforce the employee’s interpretation instead of encouraging them to consider the other person’s perspective.
That creates a particularly relevant question for companies deploying AI internally: What exactly is the system optimizing for when an employee asks for help?
From Productivity Assistant to Workplace Adviser
The shift matters because employees are already treating AI as more than a productivity layer.
A new report from Cloverleaf Labs cites research showing that 93% of workers have used AI to prepare for a conversation with their boss, while 49% said AI was more emotionally supportive than their manager.
That makes workplace AI different from traditional enterprise software.
An AI assistant that summarizes a meeting can be evaluated largely on whether the summary is accurate. An AI system advising someone about a conflict involves a different set of variables: power, accountability, context, emotion and the possibility that both people contributed to the problem.
Cloverleaf Chief Strategy Officer Kirsten Moorefield described the danger as a form of automated affirmation.
“Even if it starts out introducing different ideas, the moment you indicate your preference, it tells you it’s a great idea. That’s not a thinking partner but a robotic affirmation.”
The distinction is becoming more relevant as companies attempt to use AI to develop the human skills that become more valuable as routine technical work is automated.
Cloverleaf’s research does not argue that companies should remove AI from workplace conversations. Instead, it points toward another way of evaluating AI coaching.
The report assesses workplace advice through criteria including self-awareness, accountability, consideration of other people, specific guidance and relationship repair.
That could change how companies think about AI deployment.
Rather than asking only whether an AI tool gives employees useful answers, organizations may increasingly need to ask what those answers do to the teams receiving them.
The question becomes especially important as AI moves from software employees use to complete tasks into software they consult about the people they work with.
For enterprise AI, the next frontier may not simply be making employees faster. It may be making sure the systems advising them do not make collaboration harder.
22 c