The work that is coming, not the work that was

Roles are usually described through their current tasks. As tools change, tasks change faster than the descriptions, and organisations continue to recruit, assess and promote against a profile that no longer matches the work. The mismatch shows up as unexplained hiring difficulty and as capable people underperforming in roles that quietly became something else.

The World Economic Forum's Future of Jobs research continues to report substantial expected skill disruption and reskilling need over the current decade (World Economic Forum, 2025). The specific percentages matter less than the planning implication: the unit of analysis should be tasks and judgement, not job titles.

Where tools take on drafting, summarising, retrieving and first-pass analysis, the remaining human contribution concentrates in framing the question, evaluating the output, deciding under ambiguity and taking responsibility for the consequence. Those are precisely the capabilities most organisations have historically developed informally, through exposure and apprenticeship.

That creates an underexamined risk: the entry-level work that used to build judgement is often the work most readily automated. Organisations that remove it without replacing the development pathway will find the shortage appearing five years later, in the layer that matters most.

Employees increasingly evaluate employers on flexibility, development, purpose and the quality of management. Employers increasingly need adaptability, cross-boundary working and comfort with ambiguity. These are not in conflict, but they are frequently negotiated as if they were — through policy positions rather than through an honest description of what the work now requires.

Deloitte's Human Capital Trends work continues to document the widening gap between the way organisations are designed and the way people expect to work (Deloitte, 2026).

Precise workforce forecasts in a period of technological change tend to be confidently wrong. A more durable approach is to identify the assumptions the current model depends on, and to define what evidence would show those assumptions failing. That converts an unanswerable forecasting question into a monitoring discipline.