UK workforce outlook, autumn 2026 to spring 2028: the entry-pipeline productivity test

The thesis: productivity pressure is becoming an entry-pipeline test

The UK workforce question for autumn 2026 to spring 2028 is not simply whether employers will hire less. It is whether cost-conscious employers can improve productivity without hollowing out the routes through which people become capable technicians, managers, advisers and business owners.

That distinction matters now because several forces are arriving together. The macroeconomic environment continues to make firms cautious about costs and demand; employers report decisions in a setting where pay, prices, investment and recruitment cannot be considered separately. At the same time, Skills England’s evidence continues to frame skills need as a structural issue, not merely a consequence of the current business cycle. ↗ ↗ ↗

Our base case is therefore a more selective labour market, not a simple no-hiring market. Roles that can show an immediate operational return may still be approved, while roles framed as general capacity are more exposed to delay or redesign. That will make the quality of workforce planning more consequential: organisations will need to decide which work should be automated, which should be redesigned, and which must remain a deliberate learning route.

This is an analytical forecast, not a numerical projection. Monetary and employer-intentions evidence can indicate the direction of pressure; it cannot, by itself, determine the volume or occupational composition of future hiring. The practical judgement in this article is that the interaction between entry-level recruitment, management practice and AI adoption will be more important than any single headline indicator. ↗ ↗

Why cutting junior hiring can create a delayed capability problem

Entry-level hiring is often treated as the flexible part of the workforce: easier to pause than experienced recruitment and easier to justify only when demand is visibly expanding. But junior roles do more than add capacity. Properly designed, they convert education, prior experience and informal potential into workplace judgement. They give organisations a route to develop people who understand their systems, customers, risks and standards.

The Government’s snapshot of entry-level hiring provides a necessary warning against assuming that the market offers straightforward routes into work for people at the start of their careers. It should be read alongside the Skills England annual report and the digital and technologies sector assessment: employers’ stated demand for capability does not automatically produce accessible pathways into that capability. ↗ ↗ ↗

The mechanism is important. When routine administrative, research or first-line customer tasks are reduced, centralised or supported by AI, an employer may conclude that fewer junior roles are needed. Yet those tasks have often been the supervised practice through which people learn quality control, commercial judgement, escalation and communication. If the task disappears without a replacement learning design, the firm may save on immediate wage cost while weakening its future supply of experienced staff.

This does not mean employers should preserve low-value work for its own sake. Some junior roles have offered poor progression and little genuine development. The better response is to distinguish between routine work that should be removed and foundational work that should be redesigned. A junior operations role, for example, can combine AI-assisted preparation with source checking, exception handling, customer communication and process improvement. The value lies in the supervised judgement around the output, not in reproducing every manual step.

For local economies, the stakes are wider. High-street businesses, small professional firms and growing local enterprises often depend on accessible first jobs to build retention, succession and customer relationships. If entry routes become restricted to applicants who already possess experience or informal networks, employers may narrow their own talent pool while reducing employability opportunities in their communities. That is a business resilience issue as well as a labour-market one.

AI changes the composition of work before it determines the number of jobs

Claims that AI will either eliminate most work or solve the productivity problem are both premature. The Office for National Statistics evidence on business use of AI shows why: adoption needs to be understood as a changing business practice, not as a single technology event. A firm using one tool for a limited task is in a materially different position from a firm that has redesigned processes, data governance, assurance and management routines around it. ↗

The more plausible near-term effect is a shift in task composition and in what employers ask candidates to demonstrate. Work involving drafting, retrieval, scheduling, basic analysis or standardised communication may be accelerated. In response, employers may place greater value on verification, domain knowledge, problem definition, customer handling, data awareness and the ability to identify when an automated output is not reliable. The AI foundation skills benchmark is useful here because it frames capability as work-relevant understanding, rather than fluency with a particular interface. ↗

That creates both an opportunity and a risk. The opportunity is to make entry-level roles more productive and more developmental: an employee can complete useful work while learning how to check evidence, manage exceptions and improve a workflow. The risk is that firms use AI to raise recruitment requirements without providing the experience through which applicants acquire them. An expectation that candidates arrive already proficient can turn technology adoption into another barrier to entry.

Employers should resist using a single tool or training-completion measure as proof of transformation. A more meaningful question is whether a process now produces better quality, faster resolution, reduced rework or better customer outcomes—and whether staff can explain and challenge the technology-assisted output. This is where productivity claims become testable rather than rhetorical.

Management quality is the transmission mechanism

Technology adoption does not automatically improve performance because managers determine how work is allocated, how targets are set, how problems are surfaced and whether staff have room to improve a process. ONS research on management practices, and its linked work on management, technology and AI adoption, makes management capability central to the interpretation of adoption evidence. ↗ ↗

In practice, weak management can produce two unhelpful outcomes. One is uncontrolled experimentation: teams use different tools, without clear standards for data, quality or escalation. The other is defensive prohibition: leaders ban or constrain use so broadly that employees receive neither productivity gains nor guided practice. Both can leave a business with the cost and confusion of technological change but little operational benefit.

The more demanding alternative is managerial design. Managers need to identify tasks suitable for assistance, set boundaries for sensitive or consequential work, establish review points, and revise role expectations as processes change. They also need to protect time for coaching. If AI reduces routine volume but managers do not reinvest some of that capacity in feedback and progression, the organisation may gain speed while losing its ability to develop people.

This matters particularly for SMEs. Smaller firms may not have specialist learning, data-governance or transformation teams, but they can often make decisions and redesign workflows faster than larger organisations. The constraint is managerial bandwidth. Shared local provision, sector partnerships and practical peer learning may therefore be more valuable to many small businesses than a large catalogue of generic courses.

A practical workforce response: protect progression, redesign work, measure outcomes

Sanctuary’s recommendation is not to ring-fence every existing junior vacancy. It is to make an explicit decision about which entry routes are strategically necessary and what productive work they will contain. That calls for four connected actions.

First, create a role-level workforce map. Separate roles that provide immediate scarce capability from roles that are primarily capacity, and identify the entry roles that feed future technical, supervisory and commercial positions. A recruitment pause may be justified in one area, but leaders should see the downstream capability consequence before applying it across the organisation.

Second, redesign junior roles around supervised contribution. Replace avoidable manual repetition with assignments that require checking, exception management, customer insight, documentation and process improvement. Define what good work looks like, what must be reviewed by a manager, and what evidence demonstrates progression. This approach makes an entry role more credible both to the employer and to the candidate.

Third, build AI foundation skills into normal work rather than treating them as a standalone compliance exercise. Staff should understand appropriate use, limitations, verification and escalation in the context of their role. Training should be followed by opportunities to use those skills on real, bounded tasks. ↗

Fourth, hold managers accountable for adoption quality. Useful measures include time released from low-value work, rework or error rates, service outcomes, internal progression, retention and the proportion of entry-level employees receiving structured feedback. These measures are not a universal scorecard; they are a way to test whether a claimed productivity gain is creating sustainable capability.

For employers with limited resources, the first version can be modest: redesign one entry route, select one repeatable process and train the manager responsible for both. The essential point is to connect hiring, learning and workflow improvement, rather than treating them as separate initiatives.

Scenarios and indicators: what could change this view

The base case could prove too pessimistic if demand, investment and employer confidence recover more strongly than expected. In that case, firms that have preserved credible entry routes may be better placed to recruit and promote quickly. It could also prove too optimistic if cost pressure intensifies and businesses make broader reductions in recruitment or training. Monetary policy reporting and the Decision Maker Panel remain useful signals of the macroeconomic and employer decision environment, but neither should be treated as a substitute for organisation-level workforce intelligence. ↗ ↗

Leaders should monitor five signals. First, whether recruitment approvals are declining across all roles or concentrating in particular functions. Second, whether pay and retention pressure is easing evenly or simply shifting towards scarce occupations. Third, whether entry-level hiring is recovering, narrowing further or changing its skill requirements. Fourth, whether AI use is moving from isolated experimentation into core processes. Finally, whether managers can show improvements in quality and throughput alongside credible progression for staff.

The counterargument is that, in a weak market, employers cannot afford to invest in development. In some cases that will be true. But indiscriminate retrenchment has a cost: it makes later recruitment harder, places more strain on experienced staff and can leave technology investments unsupported by the people needed to use them well. The strategic choice is not between austerity and development. It is between passive cost reduction and a selective productivity strategy.

The judgement

Over the next 6–18 months, the strongest workforce advantage is unlikely to come from hiring the fewest people or from adopting the most AI tools. It will come from redesigning work so that technology removes avoidable friction while people continue to gain the judgement, confidence and experience that organisations need.

For UK employers, that makes entry-level hiring a board-level productivity issue rather than a peripheral talent initiative. Firms that protect and improve credible routes into work can build capability, widen access and strengthen local economic value at the same time. Those that remove the pipeline without redesigning it may discover that a short-term efficiency decision has created a longer-term management and skills constraint.