Training your managers to manage in an AI-augmented workplace: where to start?
The deployment of AI in organisations creates a blind spot in most training plans: employees are trained to use AI tools, but managers are not trained to manage teams that use AI. Yet AI fundamentally changes the nature of some managerial tasks: assessing individual performance, allocating work, detecting errors, developing team skills. Managers who have not updated their frame of reference on these subjects make decisions on bases that have become inaccurate.
What AI changes in managers' work
AI increases some employees' productivity asymmetrically. Two employees at the same seniority level and with identical job descriptions can have very different productivity levels depending on their mastery of AI tools. This asymmetry creates equity and performance management challenges that managers have not yet learned to handle.
AI makes some tasks almost invisible. When a task that took 3 hours is reduced to 20 minutes thanks to AI, the freed time is not automatically reallocated to higher-value tasks. Without a manager who questions how freed time is used and creates the conditions for it to be reinvested productively, the potential productivity gain remains theoretical.
AI raises new accountability questions. When an employee makes a decision based on a recommendation from an AI system, who is responsible for the decision? The employee, the manager, or the organisation that deployed the tool without training its teams to critically evaluate AI outputs? These questions are not settled in law in 2026, and managers must take clear positions on this within their teams before an incident forces them to improvise.
AI raises equity questions in work allocation. If an employee masters AI and produces in 4 hours what colleagues produce in a day, the manager must decide: do I assign them more work (and how do I recognise this productivity gap)? Do I train other colleagues to narrow the gap? Do I reconfigure roles?
The specific managerial skills to develop
Assessing the quality of AI outputs. A manager who cannot spot a hallucination in an AI-generated report, or who cannot evaluate the reliability of an AI-produced analysis, cannot effectively supervise an AI-augmented team.
Managing performance augmented by AI. Revising objectives, evaluation criteria and recognition to account for AI-augmented productivity without creating perverse incentives (rewarding AI use that hides shallow work).
Detecting and addressing over-reliance. Spotting when a team member delegates too much to AI and loses critical judgement, and coaching them back to appropriate reliance.
Leading learning. Creating space for the team to share AI practices, experiment safely and learn from errors, rather than leaving AI adoption to individual initiative.
Where to start
Train managers before scaling AI tools more widely. Start with a half-day workshop using real team cases, then embed AI-related objectives into the professional development review cycle.