The measurable gap in AI adoption between large UK employers and small-to-medium enterprises is no longer a macroeconomic abstraction. YouGov polling, cited in a joint report by the Learning and Work Institute and The Rigby Foundation, finds that workers at small businesses are less than half as likely to use AI extensively as their counterparts at large enterprises. For HR Directors and People leads at SMEs, the implication is direct: a skills deficit of this scale concentrates retention risk, depresses internal mobility, and weakens the employer-value proposition at precisely the moment the labour market is pricing AI fluency into offer decisions.

The same report models a route to closing that gap with measurable economic consequence. Comprehensive AI upskilling could generate a 2.3% economy-wide productivity uplift by 2035, worth approximately £80 billion, based on the Office for Budget Responsibility's central scenario. The mechanism proposed is a Skills Tax Credit administered through HMRC (His Majesty's Revenue and Customs): 20p back for every £1 an employer spends on accredited training, with a 30p AI super-rate running until 2030. The policy signal is clear even before Royal Assent. Government intends to co-fund employer training spend. The investment decision, and the internal capability to deploy it, still sits on the employer's balance sheet.

Where the adoption gap actually sits

The headline figure from the YouGov polling is that 85% of UK firms use AI in some capacity, but only 18% use it extensively. Segmented by employer size, the picture sharpens considerably: 25% of large firms report extensive AI use, against 11% of small businesses. A quarter of small businesses have not adopted AI at all, compared with 3% of large enterprises. These are not rounding errors. They represent a structural divergence that compounds quarter by quarter as large-firm workforces accumulate AI-assisted workflow experience.

Translated into headcount terms, approximately 5.2 million SME employees, around 26% of the UK workforce, depend on employer decisions about AI tooling and training access that are being made, or deferred, by organisations with thinner learning-and-development budgets than large-cap peers. A further 1.3 million workers sit inside firms that have not engaged with AI at all. The Office for National Statistics (ONS) Annual Survey of Hours and Earnings consistently shows a pay premium in roles requiring advanced digital competence; as AI fluency becomes a component of that competence, the earnings and mobility gap between AI-exposed and AI-absent workers will widen. HR functions that do not model this internally are already operating with incomplete workforce-planning data.

What the Skills Tax Credit would and would not solve

The proposed 20p-per-£1 credit on accredited training, rising to 30p for AI-relevant courses until 2030, is calibrated to shift the marginal economics of employer training investment. For an SME running a £50,000 annual learning-and-development budget, the standard rate reduces the net cost of accredited spend by a fifth. The AI super-rate, for a defined window, almost doubles that incentive on qualifying programmes. Whether this is enough to move the needle in practice depends on two factors that the policy does not address directly.

First, the credit is a legislative proposal, not enacted law. The report calls on government to establish a dedicated AI Skills Intelligence Unit and to develop quality kitemarks for non-accredited training provision. HMRC would administer any enacted credit, but the accreditation infrastructure it relies on does not yet exist in the form required. Second, a tax credit defrays cost; it does not build the internal HR capability to scope training needs, select credible providers, or measure whether completed programmes translate into measurable productivity change. The Chartered Institute of Personnel and Development (CIPD), whose annual L&D survey tracks employer training investment across the UK, has consistently found that SME L&D functions are thinner in analytical resource than those at large enterprises. A tax incentive that arrives before that capacity is built risks being claimed without being deployed strategically.

The comparison with the Apprenticeship Levy is instructive. Introduced in 2017, the Levy was designed to shift employer training spend upward, but ONS and CIPD data subsequently showed that smaller employers, who do not pay the Levy directly, were slower to access funded apprenticeship routes than large levy-paying firms. A Skills Tax Credit faces a structurally similar adoption curve at the SME end.

Stephen Evans, chief executive of the Learning and Work Institute, set out the stakes in the report: the UK's AI advantage depends on ensuring that leaders and workers across the economy have the skills to make the most of these new technologies, and that AI's transformative potential and its gains are within reach.

Retention pressure concentrates on the mid-skill layer

The workforce segment most exposed to the SME AI gap is not the senior leadership cohort, where AI strategy conversations are already common. It is the mid-skill operational and administrative population: roles where AI-assisted workflows produce the fastest measurable productivity delta and where structured training access creates a visible performance differential between employees at different types of employer.

When a mid-skill employee at a large firm receives structured AI training and earns a recognised credential, and their SME-employed counterpart does not, the CV gap between them compounds. For HR, the retention implications sit across three distinct decision points. Retention conversations with mid-skill employees now require a substantive L&D answer that goes beyond a pay review. Candidate expectations at the offer stage increasingly include direct questions about AI tooling availability and who funds the training. And workforce planning models need to distinguish between AI-fluent and AI-adjacent roles, because the two are diverging on both internal mobility pathways and external market rates. CIPD reward benchmarking data should be the reference point for that rate divergence in the UK context.

What HR teams at SMEs can do before the policy lands

Deferring action until the Skills Tax Credit reaches the statute book is not a workforce strategy. The Learning and Work Institute report frames government action as necessary but insufficient. The government's stated commitment to train 10 million people in AI by 2030 sets a population-level ambition but does not specify the internal mechanism for any individual employer. The Department for Science, Innovation and Technology (DSIT), which published the UK's AI Opportunities Action Plan, has signalled that employer-led training will be central to that target. HR functions that move now build the operational infrastructure required to claim and convert any credit that follows. Three concrete steps are available before legislation.

• Baseline the current state. Audit which roles use AI tools, at what intensity, and with what formal training behind them. This audit becomes the denominator against which any future credit is claimed and reported to the board. Without it, claims are unverifiable and investment cases are anecdotal.
• Pre-qualify accredited providers. If the 30p AI super-rate is enacted, the SMEs that already hold a shortlist of accredited training partners will move quickly. Those still scoping vendors at that point will lose at least two procurement cycles before meaningful training begins. The CIPD's provider directory and the Institute for Apprenticeships and Technical Education (IfATE) register are the practical starting points for that shortlist.
• Tie training to role progression. Accredited AI training funded by the employer generates a retention benefit only when it maps to a visible internal outcome: a band change, a remit expansion, or a formally recognised capability credential. Training that carries no internal pathway pays for the employee's next external move rather than the employer's next business cycle.

The employer-branding read

The AI adoption gap is reshaping the SME employer proposition in a way that sits alongside, and in some sectors now outweighs, the traditional axes of comparison with large employers. Historically, SMEs competed on role breadth, decision-making proximity, and cultural character against large-firm compensation structures and formal career architecture. AI training access has emerged as a distinct axis of that comparison, particularly among candidates with between three and ten years of working experience, the cohort most likely to have formed expectations from previous large-employer or consulting exposure.

Employer reviews on platforms where current and former staff describe day-to-day tooling and development access are becoming a leading indicator of this shift. HR functions monitoring their employer reputation through review data will typically see the AI question surface in review commentary before it appears in exit interview data. That lag between signal and survey is an intelligence gap. Review-level commentary on AI tooling and training access should be treated as structured workforce intelligence rather than anecdotal noise, and fed into L&D planning cycles rather than communications responses.

Steve Rigby, CEO of Rigby Group and chair of the Rigby Foundation, framed the scale of the opportunity in the report: an £80 billion return on investing in the workforce is within reach, and SMEs and government together can make this AI transition a success story for the whole economy, with the current administration holding a genuine opportunity to reform the skills system and drive immediate productive investment in human capital.

Bottom line for HR leaders at SMEs

The Learning and Work Institute and Rigby Foundation report is a policy paper, but the decision it surfaces for HR is operational. The 11%-versus-25% extensive-use gap between small-business and large-enterprise AI adoption is the number that reward committees, People Directors, and CFOs at SMEs should place alongside their retention and hiring cost data. It is a gap large enough to reshape which SMEs win the mid-skill talent market over the next 24 months, and small enough to close with focused, sequenced investment.

If the Skills Tax Credit is enacted and administered through HMRC in the form proposed, the organisations that convert it into workforce advantage will be those that have already completed the baseline audit, established accredited provider relationships, and built the internal L&D capability to measure training outcomes against role performance. Those that treat the credit as a procurement trigger rather than a planning input will spend the first year of availability in vendor selection rather than skills delivery. The CIPD's annual L&D survey provides a ready benchmark for where UK SME training investment currently sits relative to sector peers. That benchmark, not the policy timetable, is the appropriate planning anchor for 2026-27 budgeting decisions.