Once in a Generation: Building an Inclusive Skills System for the AI Transition
Written by Vicky Murray & Kate Wolfenden
Systems rarely change because someone finally writes the right report. They change when several separate forces move at once, and somebody is ready.
That is where the UK skills system sits today. AI is set to reshape the labour market faster than any transformation in living memory.
Andy Burnham arrived in Downing Street with an explicit governing philosophy of devolution and investment in place, or as he puts it, taking power out of Whitehall and carrying it into every postcode. And the skills system is midway through its own reform moment, with the October 2025 White Paper on education and skills after 16 now being translated into new qualification pathways and a reformed funding model.
Within days of taking office the government announced technical routes from age 14, tied some £90bn of annual procurement to employers' records on local jobs and apprenticeships, and said Britain must value the hard hat as much as the graduation cap.
Whatever you make of the rhetoric, the direction now matches what we have spent the past year mapping. Working for the last year,, we set out to understand what is really holding the UK workforce skills system back, and what smart responses to those deep systemic challenges would look like.
Understanding the skills system today
Looking inwards
We began by mapping the skills system rather than cataloging its symptoms i.e. looking at how problems connect and reinforce each other, not just listing them one by one, to bring deeper insight. This process included desk top research, around 20 interviews across government, employers, unions, education and local authorities, then a workshop to define the problem state (the ‘as is’ reality), together with further interviews to test what we found.
The result is a systems map of 48 dynamics, distilled into nine interconnected themes. Six describe how the system locks itself into underperformance, each deepening the next.
Interventionism, where low growth erodes the funding and mandate for coherent policy, producing reactive policymaking, more young people out of work and training, and more shortages.
Investor confidence, where perceived incoherence dampens employer investment, leaving those shortages unaddressed.
Human impact and social costs, where declining skills and financial security erode health and confidence, deepening a low skills poverty trap.
Talent leakage, where stagnating productivity weakens our ability to keep highly skilled people.
Buying in, where employers are rationally responding to the skills shortage by acquiring ready made talent rather than building talent pipelines.
And AI or skills, where such shortages continue to nudge employers toward automation as a substitute for training.
Three further dynamics show the system trying to respond: the Industrial Strategy, place based initiatives, and devolution and coordination. Here the tension is live. At its worst, local leaders hold responsibility for outcomes without full authority over how to deliver them.
The Integrated Settlement, (merging skills, transport and housing funding into one flexible pot), went live in Greater Manchester and the West Midlands in April 2025 and reached five more authorities this year. Where it exists, it lets a place connect training and jobs to housing and transport, which is where inclusion is actually won: a great course or apprenticeship is worthless to someone who can't physically get there, or afford the transport to attend. But the capacity to use the Integrated Settlement is still catching up, and elsewhere the architecture does not exist. Whether further devolution resolves the coordination problem or relocates it is one of the more consequential open questions in the system today.
We have published the full map as a community asset, with the evidence behind each dynamic, so anyone can find themselves in it: what acts on them, what they act on, and where they could intervene. Explore the map here.
Looking outwards
Then we looked outwards, at what is happening to the market and society shaping influences of work itself. Two things are moving in parallel, at very different speeds.
AI adoption is racing. 97% of UK organisations report a significant AI skills gap, while only 21% of adults can explain AI in any meaningful detail. National training targets measure success on courses completed rather than on who is reached, above a digital exclusion floor of roughly ten million adults without basic digital skills. To compound the problem, access is sharply stratified: people earning over £100k are two and a half times more likely than those under £30k to have had 20 hours or more of AI training in a year.
And the impacts, while complex and hard to pin down to one specific cause, are arriving faster than expected. Graduate roles are down 45% year on year, with some 707,000 graduates are out of work and claiming benefits: a 46% rise since 2019. Firms most exposed to AI are 16.3% less likely to post new vacancies, and the cuts concentrate in the entry level roles firms have always used to grow senior talent. In short: The career ladder is losing its crucial bottom rungs.
None of this is unprecedented in kind, only in speed. The UK was historically the first to industrialise but has been reshaped by every wave since, without ever arriving with a credible plan for the people affected. In addition, each transformation has run at roughly half the duration of the last: 80 to 100 years from agrarian to industrial Britain, 50 to 60 through electrification, 30 to 40 into the knowledge economy. If historical precedents are anything to go by, we may be seeing the deeper societal impacts of automation in as little as 10 to 20 years.
However, the good news is that in each and every historic case we studied, it was stakeholders acting together, rather than government alone, that made it work on the ground.
What can we do about it?
That’s exactly the question we asked ourselves, putting the UK workforce skills system stakeholders at the centre of the question. So we convened a second workshop, with Skills England, DfE, DWP, the TUC and partners across business, local government, education and civil society, to align on a guiding star and design how to reach it.
Out of it came 21 interventions across four leverage areas: foresight on workforce and societal disruption, devolution coordination and capacity, corporate buy not build, and outdated career paths. Within those, four clusters of coordinated action carried the greatest potential.
The Future Skills Amplifier. Harnessing the opportunity to increase the popularity of green and sustainable jobs, to direct people into as the displacement impact of AI adoption grows.
Hopeful Futures. Changing the hearts and minds of the UK public on what counts as an aspirational career, and on their own likelihood of achieving one.
Inclusion by Design. Building a centre of gravity around inclusion at the design and insight stage of the AI transition.
Responsible Business. Accelerating recognition and reward for responsible AI adopters, while ensuring displacement becomes coordinated and well supported.
Our golden opportunity
So, if you want to meaningfully take action on this important confluence of issues, where should an organisation start?
1: Build responsible foundations. Credibility here needs to be earned. No organisation can convincingly call for sector transition plans or workforce impact disclosure while its own AI adoption is unmanaged and invisible to its supply chain. Across many stakeholders across the system, that means starting at home: supporting equitable adoption across your teams, extending that support into SME supplier relationships, and an internal policy for forecasting and managing AI's workforce impact. None of it needs finishing before acting externally, but we recommend that earnestly taking these steps to prove your credibility does matter.
2: Seize the future skills opportunity with a future skills amplifier. This is our golden nugget of opportunity to be seized now. Around four million UK workers need retraining by 2035 to close the green skills gap, that’s just one of our critical future skills needs. And the reward is real: PwC finds a 23% average pay premium for entry level green roles across 60% of occupations, including where no degree is required. Yet only 27% of people in their twenties can say what a green job is, and just one in eight workers holds green skills. Meanwhile the number of 16 to 24 year olds not in education, employment or training has passed a million for the first time since 2013. Two transitions are failing separately. For this reason, we recommend a Future Skills Amplifier that maps at risk or displaced workers together with locked out young people to roles the country already needs filled, supported by a career transition programme and run as a nationally connected network of local programmes led by Combined Strategic Authorities.
3: Call for a fair transition. Finally, while 1 and 2 focuses on what we can do to be responsible and find opportunity within this seismic workforce skills system shift, the wider societal responsibility piece is left unanswered. For this, we recommend gathering our voices together to call for an independent, inclusive coalition for a just AI transition, jointly owned by workers, learners, SMEs, local authorities and affected communities alongside government, employers and frontier labs. The recent government shifts e.g. a new dedicated minister, and TUCs call for a national AI workforce strategy are positive; but the prevailing focus for the UK right now is growth, and its dedicated advisory committee remains predominantly big international tech. This cannot, on its own, be the forum in which a fair UK transition is designed. This is why we recommend a small founding alliance to incite and steward that coalition into fruition, seeded this year and inaugurated no later than early 2027.
The window will not stay open
If you found this interesting, we will be launching a deeper dive or the full systemic action insights at the Blue Earth Summit in London this October.
The timing is not incidental. Three windows that usually move independently have opened together for the first time in a decade: AI, devolution, and skills reform. Each sits at a different level of the system, and for once they reinforce rather than compete. That creates a rare permission to act at scale, to attempt what would ordinarily meet resistance, and to build inclusion into the AI transition from the start.
We welcome you to join us at the Blue Earth Summit, or get in touch directly on vicky@103.ventures if you are inspired to act.
