React warns AI must not replace management training

React has argued that businesses investing heavily in artificial intelligence must not neglect face-to-face training in management and communication skills. The workplace learning specialist said the spread of AI has made human judgement and empathy more important.

React, which has used professional actors in workplace training for more than three decades, said managers still need opportunities to practise difficult conversations with real people rather than relying solely on theory, online courses or AI-generated advice.

Its intervention comes as employers adopt AI tools across customer service, data analysis, recruitment, administration and decision-making, while also using the technology in staff learning and development.

That shift risks creating a gap between knowing what to say and responding effectively in a live exchange with an employee, colleague, or senior executive. React's training model places managers in simulated workplace situations where another person may challenge them, disagree, become upset or react defensively.

Practise gap

The argument centres on the idea that many managers are asked to handle some of the most difficult conversations of their careers without rehearsal. These may include discussions about poor performance, inappropriate behaviour, workplace conflict, restructuring or disputes over management decisions.

Such interactions can affect careers, staff wellbeing, productivity and retention. React argues that leaders build or lose trust through how they listen, communicate difficult decisions, respond under pressure and behave when challenged.

Adrian Cook, Chief Executive Officer of React and founder of the Trust Pyramid Ecosystem, said human interaction remains central to developing those behaviours.

"AI can give a manager information, suggest what they should say and even help them prepare for a difficult conversation, but knowing the answer is not the same as being able to handle the person sitting opposite you. People don't follow a script. They become defensive, emotional or frustrated. They interrupt, challenge you, misunderstand what you meant or simply disagree. Those moments require judgement, empathy and the ability to listen and adapt. As we embrace AI, we need to be careful that we don't lose the opportunities people have to develop those skills through genuine human interaction," Cook said.

Trust framework

Alongside that message, React has introduced The Trust Pyramid, a framework it describes as helping organisations assess and strengthen trust at different levels of a business. The model is intended to shift trust from a broad cultural idea to a set of observable behaviours and leadership practices.

The framework is designed to help leaders identify the relationships and behaviours that influence commercial performance, employee confidence and customer loyalty, according to React. The wider offering includes leadership development programmes, behavioural diagnostics and an AI-focused masterclass.

Cook's comments reflect a broader debate in corporate learning over whether AI can replace parts of management development or should remain a support tool. AI systems can generate scenarios, provide coaching prompts and offer feedback at scale, but critics of an AI-only approach say they cannot reproduce the pressure and unpredictability of a tense workplace exchange.

React argues that experiential learning still fills that gap. In its model, a workplace conversation can be paused, reviewed and repeated, allowing a manager to test how changes in tone, wording or approach affect the other person's response.

The company draws a parallel with other professions where rehearsal is routine. Actors practise before performances, athletes train repeatedly before competition, and pilots spend extensive time in simulators before facing high-stakes situations. React argues that management is unusual in expecting people to handle emotionally charged or career-defining conversations for the first time in real conditions.

That stance is likely to resonate with employers under pressure to deploy AI quickly while maintaining workforce trust. As more routine tasks are handed to software, the relative value of skills such as listening, persuasion, judgement and empathy may rise rather than fall.

Cook said those skills cannot be learned solely through written values or formal instruction. "Those behaviours cannot simply be written into corporate values or learned as theory. They must be demonstrated consistently in real interactions," he said.