Anticipating the jobs of tomorrow: why we keep looking in the rear-view mirror

---

Nearly every organisation says it wants to “anticipate” – to get ahead of changes to their roles, the critical skills of tomorrow, the impact of AI or the green transition on their workforce.

Look at what’s actually happening, though, and the pattern is usually the same: organisations document the present, project the past forward, and rework yesterday’s frameworks with a few tweaks. And they call this anticipation.

Looking in the rear-view mirror isn’t a lapse in judgement — it’s a logical response to a real difficulty. Anticipating means working where the data is thin, where the answers aren’t sitting in a benchmark report, where uncertainty simply won’t go away. That uncertainty needs to become part of the analysis, not something to be brushed aside.

What we want to explore in this article is what makes anticipating jobs and skills so difficult and why organisations so often sidestep it. Not to criticise current practice, but to identify what would allow them to do things differently.

---

---

Why we don’t really anticipate the jobs of tomorrow

Most HR forecasting exercises rely on three familiar moves: analysing sector trends, projecting headcount three years out, and identifying “critical” skills. These exercises have value — but they share a common flaw that limits what they can deliver: they all start from what already exists.

Current roles get projected into a slightly different future. Today’s skills gaps get listed. Competitors get benchmarked — competitors who are often doing exactly the same thing at exactly the same time.

The result is a vision of the future that looks a lot like the present. Reassuring, well-documented, and defensible in front of the board. Rarely enough to inform the decisions that actually matter.

The deep transformations — the ones that genuinely reshape jobs — don’t come from within. They come from disruption: technological, regulatory, societal, economic. And these disruptions don’t show up in existing HR data. They show up in weak signals, sector dynamics, and shifts that aren’t yet visible on the org chart.

The confusion between mapping and anticipating

There’s a common confusion between two very different activities: mapping skills and anticipating jobs.

Mapping is taking a snapshot of what exists today. It’s useful for managing resources, spotting gaps, and structuring internal mobility. But a snapshot, however precise, tells you nothing about what those jobs will become in three years.

Anticipating is something else entirely. It means working with scenarios of change, combining internal and external signals, and building hypotheses about what jobs will become – not what they are today. It means accepting that conclusions will be provisional, that certainty will be rare, and that the value of the exercise lies not in the document it produces but in the quality of the decisions it helps inform.

This confusion matters. It explains why so many organisations invest in highly detailed skills frameworks… without it changing much about their ability to anticipate. The tool is there, but the strategic thinking hasn’t really happened.

You see this same confusion in the maturity models used by consultancies. Skills mapping is described there as a foundation — a descriptive base layer. Scenario planning belongs to a different, rarer level that few organisations actually reach. Recent HR maturity studies place most large European companies in a transitional stage, with forecasting capability still underdeveloped. In other words, many organisations have done the mapping work and believe they’ve done the anticipation work too. They’ve built the descriptive foundation and mistaken it for genuine forecasting capability.

What AI reveals about our blind spots on jobs

The arrival of generative AI in organisations is a good stress test for this difficulty.

Many HR teams have tried to identify “jobs at risk” or “skills to develop” in response to AI. A legitimate exercise — but often carried out with the same old tools: frameworks, matrices, percentages of tasks that could be automated.

The trouble is, AI’s impact on jobs isn’t linear. Task-automation matrices measure potential exposure, not actual impact. What could be automated isn’t necessarily what will be. Adoption depends on cost, regulatory constraints, and organisational choices. Plenty of technically automatable tasks simply aren’t worth automating in a given context.

These methods also assume tasks stay stable. But jobs are being reshaped as AI gets integrated into them. New activities are emerging — oversight, supervision, integration — that no framework saw coming. Some roles judged “at risk” are actually becoming more complex. Others, thought to be stable, are changing faster than anyone expected.

What AI really reveals is less a skills problem than a reading problem. The question isn’t just “which skills should we develop?” but “how will our roles be reshaped, and what does that mean for how we organise, recruit and develop people?”

Asking the question differently changes what you go looking for — and who you go looking with.

Anticipating with jobs, not for them

One of the most common limitations of HR forecasting exercises is that they’re built behind closed doors. An HR team or a consultancy produces an analysis, presents it to leadership, and rolls out an action plan. Operational managers receive the conclusions without having taken part in the thinking.

The problem isn’t methodological — it’s more fundamental than that: trying to anticipate how a job will evolve without the people who actually do it means missing half the relevant signals.

People on the ground see things the data doesn’t show yet. They notice how their work is shifting, the tensions that are emerging, and the skills quietly becoming important before anyone has formally flagged them.

Research on involving operational staff in forecasting exercises backs this up: bringing in the people who actually do the work corrects biases that expert analysis tends to reproduce — underestimated roles, transformations spotted too late. Incorporating these perspectives doesn’t just enrich the thinking; it also creates the conditions for genuine buy-in.

Anticipating with the people in these roles also means accepting some uncomfortable conversations — about what might disappear, about what certain roles will become, and about the trade-offs the organisation will have to own. These conversations aren’t always easy to start. But they’re exactly what makes forecasting worthwhile.

Changing the starting question to genuinely anticipate

Most forecasting exercises start with the same question: “Do we have the right skills?” It’s a fair question, but it points towards stock-taking, comparison against a framework, and hunting for a gap to close.

A different question opens up a different space: “What will we decide, armed with this knowledge?”

Reallocate resources? Step back from certain activities? Invest in emerging profiles rather than retrain existing ones? Outsource what won’t be core business in three years?

If the forecasting exercise doesn’t lead to decisions — not necessarily immediate ones, but ones that are identified and owned — it remains an intellectual exercise. Useful for reflection, perhaps, but rarely leading to an actual trade-off.

This shift — from mapping to deciding, from stock-taking to trade-offs — is what gives HR forecasting its real strategic value. It also requires a clear mandate for the HR function: not just to document change, but to help leadership and the business make sense of it together.

A limitation that deserves its own article: SWP governance

Asking the right starting question isn’t enough on its own. We regularly see serious forecasting work — built on scenarios, informed by external signals — that still ends up producing documents rather than decisions. The scenarios exist; they circulate. People call them “interesting”, but they never trigger a single reallocation of resources.

The reason isn’t analytical — it’s organisational. These exercises aren’t connected to the places and moments where decisions actually get made: budget cycles, investment committees, and trade-offs between business lines. Without that connection, even the best forecasting has no effect.

This deserves its own treatment. We’ll be dedicating a separate article to the governance of Strategic Workforce Planning, and the conditions that turn forecasting into an actual decision-making input.

For now, the takeaway is simple: changing the question isn’t enough if the organisation hasn’t worked out where that question will actually be settled.

What organisations that genuinely anticipate do differently

Without claiming there’s a universal model, a few common habits stand out among organisations that manage to turn their forecasting into a genuine decision-making lever.

They work with scenarios rather than forecasts. They’re not trying to predict a single future. They build several hypotheses about how things might unfold and identify what would hold true across each of them. This changes the nature of the decisions being made – the goal is no longer finding the right answer but making choices that stand up across multiple possible futures.

They don’t confine themselves to internal data. Regulatory changes, sector reshuffles, new market entrants, shifting employee expectations: forecasting draws first on what’s happening outside, in places HR data can’t yet see.

They separate their time horizons. What’s urgent over the next six months and what’s structurally important over the next three years call for different responses and different people. Conflating the two usually means sacrificing forecasting to day-to-day firefighting.

Finally, they accept the discomfort of provisional conclusions. Good forecasting rarely produces certainties. It produces a shared understanding, explicit hypotheses, and better-framed questions. That’s already a great deal — provided you accept that it’s enough to move forward.

Anticipating the jobs of tomorrow: what makes the difference

Anticipating tomorrow’s jobs doesn’t require more sophisticated tools or more precise data. Above all, it requires changing what you’re looking for and accepting that you’re working in a space where definitive answers are rare.

The organisations that get this right aren’t the ones with the best frameworks. They’re the ones that have learnt to have the right conversations, ask the right questions, and make owned decisions under uncertainty – rather than waiting for the uncertainty to disappear.

A few questions to guide your own thinking:

→ Does our forecasting process produce decisions — or documents?

→ Do operational managers take part in thinking through how their roles will evolve, or do they simply receive the conclusions?

→ Are we working with scenarios of change, or projecting the present forward?

→ What mandate have we actually given HR on these issues?

→ Are our scenarios connected to the moments when the organisation truly makes decisions — or do they exist somewhere off to the side?

Sources

KPMG, Strategic Workforce Planning, 2025.

McKinsey & Company, The critical role of strategic workforce planning in the age of AI, 2025.

Deloitte, Human Capital Trends 2023 — The Future of Workforce Management, 2023.

McKinsey & Company, How banks can build their future workforce today, 2021.

5 IDC, Indice de maturité des RH — Étude menée auprès de 740 décideurs RH d’entreprises de plus de 500 salariés, 2024.

Commission européenne, Strategic Workforce Planning in the European Commission, 2025.

Organisation internationale du Travail (OIT), Generative AI and Jobs: A Global Analysis of Potential Effects on Job Quantity and Quality, 2023.

Ada Lovelace Institute, Evaluating claims about AI and productivity in the UK public sector, 2026.

Organisation internationale du Travail (OIT), Developing Skills Foresights, Scenarios and Forecasts, 2016.

Deloitte, High-Impact Workforce Research: In Brief, 2020.