The real challenge for management: redesigning work in the wake of AI

And what is needed to achieve the so-called ‘amplification effect’?

This effect occurs when AI enhances human thinking rather than replacing it. With clear guidance on when to rely on AI, when to question its results and when not to use it, it is possible to broaden perspectives and improve decision-making, whilst preserving people’s judgement, problem-solving skills and professional autonomy.

If the move towards more generalist roles is perceived as a loss of expertise or a devaluation of the skills built up over the years, what should the company do?

Competence is not merely a set of skills. It is often also a source of recognition, self-confidence and professional development. Our research has found that more generalist roles can offer employees a 21 per cent greater variety of skills and tasks. However, there are conflicting indicators: 20 per cent of employees say their company expects them to become more generalist, whilst 23 per cent state that the aim is greater specialisation. If taking on a wider range of tasks is perceived as a devaluation of one’s skills, the company should not dismiss this reaction as mere resistance to change. Instead, it should demonstrate how the skills already acquired continue to form the foundation upon which to expand their contribution, clarify the boundaries of roles and the expected quality standards, and ensure that recognition, incentive and professional development systems value both the breadth and depth of skills. The ultimate aim is to ensure that employees experience this evolution as an expansion of their skills, not as a loss.

When AI agents carry out tasks and interact with business systems, how do trust, accountability and governance change? 

When AI agents move from providing suggestions to taking concrete action, trust can no longer be based solely on the apparent usefulness of their results. Organisations must clearly define the limits of the decisions an agent can make, ensure transparency regarding the actions it has taken, and identify a person responsible for the results achieved. Coordination must also be rethought, so that everyone knows when an agent is acting, when human verification is required, and how to intervene.

Is there a risk that excessive collaboration with officers might weaken, rather than strengthen, relationships between colleagues?

Our research has found that 40 per cent of employees already turn to AI for emotional support on work-related issues, whilst 80 per cent believe that agents should support their work, rather than supervise it. The impact on relationships therefore also warrants attention. The goal should be a form of collaboration tailored to needs: using agents to reduce administrative and coordination burdens, whilst at the same time deliberately preserving the human interactions through which trust, judgement, learning and relationships are built. Agents should create more opportunities for colleagues to collaborate with one another, rather than becoming the intermediary through which they interact.

How will the professionals of tomorrow be trained if AI takes over some of the routine tasks that are often the first stepping stone for young people learning a trade? 

The professionals of tomorrow will need to learn through intentionally redesigned early-career roles, rather than simply relying on the routine tasks through which previous generations were trained. Organisations, with HR at the forefront and in collaboration with business unit managers, have a responsibility to safeguard career development pathways by rethinking recruitment, internal mobility and promotion processes, so that young talent can develop their skills as the nature of work evolves. The solution is not to preserve routine tasks, but to replace the learning opportunities of the past with meaningful assignments, structured exercises, coaching and opportunities to observe how more experienced colleagues exercise their judgement.

What is becoming more important in human leadership today? 

As AI takes on an increasingly important role in coordination, routine feedback and coaching, managers should spend less time supervising tasks and more time developing people, building trust and fostering a sense of belonging. Many current leaders were selected for their technical skills and track record, not for the motivation or abilities required to lead people. It is therefore necessary to redefine the profile of the ideal manager, and this need is consistent with two broader shifts in the organisation of work: a move from managing people to empowering individuals to lead themselves, and from a pyramid-shaped organisational structure to a diamond-shaped one, with fewer levels of coordination and more people closer to operational activities. In this model, managers are not merely supervisors, but become, above all, facilitators of others’ success.

Final question: in your opinion, what is the most serious mistake a company can make today when introducing AI into its processes? 

Viewing artificial intelligence primarily as an automation initiative, rather than as an opportunity to redesign work. In the absence of a compelling vision for the transformation of work – one that is shared by senior management and clearly communicated to staff – organisations will end up using AI mainly to improve efficiency. This will produce mixed results and limit the potential that can emerge from collaboration between people and technology. The most urgent decision leaders must make concerns the commitment to allocate the time, effort and resources needed to redesign work and the talent management and development systems that support it.