If you commission L&D for a UK business, the most consequential shift happening to your job in 2026 isn’t the AI one. It’s the shift from role-based to skills-based workforce models. 79% of UK HR managers told TalentLMS their organisation is moving toward skills-based working. That’s not a future trend; it’s already changed the brief most L&D teams operate under, whether the buyer has registered it yet or not.
The implication that’s easy to miss is operational rather than conceptual. In a role-based model, you commission training for “our sales managers” or “our finance team”. In a skills-based model, you commission training for “people who need AI-governance capability” or “people who need workflow automation skill”. The unit of demand changes. And once the unit of demand changes, the formats of training that fit changes with it.
This piece unpacks what a skills-based organisation actually is, why the shift is happening now in the UK, the L&D commissioning model that fits it, and the funded training formats that map cleanly to a skill rather than to a role. Heavy emphasis on the apprenticeship unit format because it’s the funded route that’s structurally aligned with skills-based commissioning, and most UK employers still aren’t using it.
The skills-based organisation in one paragraph
An organisation that defines work, develops people, and deploys capability in terms of discrete skills (workflow automation, AI governance, financial modelling, data literacy) rather than in terms of fixed job titles. People hold portfolios of skills; work is assembled from skill requirements; development happens at the skill level, not the role level. The operating-model shift that follows: dynamic team composition, skill-based pay differentials, talent-marketplace tooling, and a fundamentally different L&D commissioning model.
Why this is happening now in the UK
Four forces are converging on the same answer:
AI compression of role definitions. AI is dissolving the work that defined the bottom half of many traditional roles, while extending the work that defined the top half. Junior analyst, mid paralegal, ops coordinator, marketing executive, in each case the role boundary is blurring. The clean response is to stop defining the role and start defining the skills that compose the work.
Talent shortages in narrow skill categories. The UK doesn’t have enough “senior data engineers” in the role sense; it has even fewer if you list the specific skill set (Snowflake + dbt + Python + governance + stakeholder management). The harder the talent market gets in a narrow skill, the more attractive it becomes to build the skill internally rather than hire the role externally.
Apprenticeship policy alignment. The Skills England portfolio rationalisation since 2023, plus the Growth & Skills Levy reforms, plus the apprenticeship unit format being designed around discrete skills rather than entire occupations, all three explicitly push the funded training infrastructure toward skill-level commissioning.
HR platform maturation. Workday, SAP SuccessFactors, Eightfold, Gloat, the talent-marketplace and skills-graph tooling is now mature enough to actually run a skills-based operating model. Five years ago this wasn’t true.
The honest question for L&D commissioners isn’t whether to adopt a skills-based model, it’s whether your commissioning process is currently buying training that fits one. Most aren’t. Most are buying role-based programmes against a skills-based future, then wondering why the ROI conversation gets harder year on year. , Rod Doyle, Director, TESS Group
What changes when you commission L&D by skill, not role
Four operational shifts. None of them are theoretical.
1. The unit of demand changes
Role-based: “train our 14 senior leaders”. Skills-based: “train 14 people across the org who need AI governance capability and who, between them, span the three populations that will own AI risk decisions for the next 24 months”. The second cohort isn’t a clone of the first; it’s a different shape entirely.
2. The cohort composition changes
Skills-based cohorts cut across departments. A typical AU0010 cohort we run includes compliance leads, COLPs/COFAs, IT security, operations leadership, internal audit, and a head of AI or chief data officer. Six different reporting lines, one shared skill outcome. That cohort composition is materially impossible to run inside a traditional role-based commissioning model.
3. The funded format changes
Apprenticeship units (AU0009, AU0010, AU0011, plus the wider unit catalogue) are explicitly funded at the skill level. Each unit funds a discrete skill outcome rather than a whole apprenticeship occupation. That’s the funding structure that matches the commissioning model. Full apprenticeships still have a role here, for the deep capability owner, but the workforce-wide skill build moves to units.
4. The success metric changes
Role-based L&D measures completion, role progression, retention against the role. Skills-based L&D measures skill acquisition, skill deployment to work that needs it, and skill currency over time. None of those are well-served by traditional LMS reporting; all of them require a skills graph and a way to evidence skill acquisition.
Why apprenticeship units are structurally aligned
The apprenticeship unit format, introduced by Skills England in 2024 and expanded through 2025/26, is the only funded training format that’s natively skill-shaped rather than role-shaped. Three structural reasons:
Reason 1: One unit, one skill outcome. Each unit is built around a discrete skill (e.g. AU0010: AI Adoption & Governance). The skill outcome is the specification, not the role of the learner.
Reason 2: Skill-level funding. Each unit attracts £750 of levy spend under Funding Model 99, drawn against the skill outcome. The funding doesn’t require the learner to be on a full apprenticeship or to hold a specific role.
Reason 3: Stackable into role-shaped portfolios. A learner who needs three related skills can stack three units. A learner who needs one specific skill can do one unit. The format flexes to the skill demand rather than imposing a fixed role-shape.
AU0009/10/11: a worked skills-based commissioning example
Skill demand: 18 people across the business need AI governance capability inside 6 months. Population: 4 compliance leads, 3 IT security, 3 ops directors, 4 commercial leads, 2 internal audit, 2 risk committee members.
Role-based commissioning answer: impossible. No single role-based programme fits all 18.
Skills-based answer: 18-seat AU0010 cohort. £13,500 of levy spend. Six-week intensive. All 18 learners hit the same skill outcome with cohort experience inside their own population.
The new commissioning playbook
Five practical steps that take an L&D function from role-based to skills-based commissioning:
1. Inventory the skill demand, not the role demand
List the skills the business needs people to hold over the next 12–24 months. AI workflow design, AI governance, data fluency, financial modelling, supplier negotiation, project ownership. Each one a skill, not a role.
2. Map skills to populations
For each skill, identify the population that needs to hold it. Most skills are held by populations that span 3–6 different role categories. That cross-role composition is the signal you’re commissioning correctly.
3. Choose the funded format per skill
Deep, long-horizon capability with formal qualification: apprenticeship. Targeted, skill-level outcome for wide cohort: unit. Immediate applied output for small focused team: short course. Most skills end up on units or short courses; the apprenticeship is reserved for the long-term capability owner per skill area.
4. Run cohort recruitment cross-functionally
The cohort recruitment process changes. Instead of asking line managers to nominate from within their team, you publish the skill outcome and let any employee opt in (with line-manager sign-off). The cohort that lands is naturally cross-role.
5. Evidence skill acquisition, not completion
Set up a skill-attestation process at the back of every cohort. The skill graph in your HR platform gets the new entry; the L&D dashboard reports skill acquisition rather than course completion. That’s the metric your CHRO will start asking for inside 12 months if they haven’t already.
How TESS delivers skills-based cohort programmes
We run AU0009/10/11 cohorts that are explicitly cross-functional, the typical cohort spans 5–7 different reporting lines, all on the same skill outcome. Cohort recruitment is supported by a skill-fit screening process so the learners who arrive are the ones with the genuine need. Skill-acquisition evidence packs at the end of each unit feed directly into your HR platform skill graph. Most clients run two skills-based cohorts in parallel rather than one large role-based programme.
Want a skills-based commissioning frame for your next L&D cycle?
Bring us your skill demand list (we’ll help build it if you don’t have one), your headcount split across functions, and the funded training budget you’re already paying into. We’ll lay out the cohort plan and the funded format per skill.
Where to read next
Three pieces that round out the skills-based picture: what are apprenticeship units?, the full unit format explained; the AU0009/10/11 unit guide, the three AI-specific Level 5 units; and our apprenticeship vs unit vs short course decision guide, the three-route decision matrix that fits skills-based commissioning.
Frequently asked questions.
What is a skills-based organisation?
An organisation that defines work, develops people, and deploys capability in terms of discrete skills rather than fixed job titles. People hold portfolios of skills; work is assembled from skill requirements; development happens at the skill level. 79% of UK HR managers say their organisation is moving toward this model.
Why is it happening in 2026 specifically?
Four converging forces: AI compression of role definitions, talent shortages in narrow skill categories, apprenticeship policy alignment with skill-level funding (apprenticeship units, Funding Model 99), and HR platform maturation (Workday, SuccessFactors, Eightfold, Gloat) making the operational model feasible.
What changes for L&D commissioning?
Four shifts. The unit of demand changes from role to skill. The cohort composition becomes cross-functional rather than within a single department. The funded training format shifts toward apprenticeship units rather than role-shaped apprenticeships. The success metric becomes skill acquisition and deployment rather than course completion.
Why apprenticeship units specifically?
Each apprenticeship unit is built around a discrete skill outcome (e.g. AU0010: AI Adoption & Governance), funded at £750 of levy spend per learner under Funding Model 99, drawn against the skill outcome rather than the learner’s role. Units stack flexibly into role-shaped portfolios. That structural alignment makes them the natural funded format for skills-based commissioning.
Do full apprenticeships still have a place?
Yes, for the deep, long-horizon capability owner per skill area. ST1512 (AI & Automation Practitioner) and ST1398 (Machine Learning Engineer) are the right call for the people who will own a specific capability domain long-term. Apprenticeship units handle the wider workforce-skill build.
How do we evidence skill acquisition rather than course completion?
Set up a skill-attestation process at the back of every cohort. The HR platform skill graph gets the new entry; the L&D dashboard reports skill acquisition rather than completion. TESS provides skill-acquisition evidence packs at the end of each unit that feed directly into the major HR platform skill graphs.