Fully Funded AI Training for Non-Technical Staff (UK)

By James Cotton, Founder, iO-Sphere

Key figures at a glance

Levy rate: 0.5% of an employer's annual pay bill over £3m (HMRC, unchanged since 2017)
Maximum levy transfer to another employer: 50% of unused funds — raised from 25% on 22 April 2024 (gov.uk / CIPP)
Employers rating their AI knowledge "beginner" or "novice": 56% of those using or planning to use AI (gov.uk, 2026-01-28)
Employers with no staff currently working with AI: 61% (gov.uk, 2026-01-28)
UK businesses using AI: ~23%, late September 2025 (ONS BICS, 2 October 2025)
Who owns apprenticeship funding policy now: Department for Work and Pensions (DWP), from 16 September 2025

Policy correct as of July 2026 — apprenticeship funding rules are changing through 2026; check the latest DWP funding rules before you commit.

What "fully funded AI training" actually means in the UK — and what it doesn't

In the UK, "fully funded AI training for staff" almost always means one specific thing: training paid for out of the apprenticeship levy, the pot every employer with an annual pay bill over £3m already contributes 0.5% of that pay bill to (HMRC, unchanged since 2017). It isn't a grant you apply for, and it isn't a free online course you sign up to and walk away from. It's your own money — money the government already took — coming back out to pay for structured training.

That distinction matters because "free" is doing a lot of hidden work in most search results. A genuinely funded programme comes with real commitments: eligibility rules, employer backing, and a working-time promise your staff have to actually keep. The small print is where plans die, so it belongs up front.

Here's the reframe we'd argue you should start from. Don't ask "is it free?" first. Ask "what do I want these people to be able to do afterwards, and how will I know it worked?" Funding determines the route, not the result. A funded course that doesn't change what someone can do is the most expensive training you can buy, because it spends the one thing you can't get back — your team's working time. Decide the outcome, judge the programme as if you were paying cash, then fit the funding to it.

Who counts as "non-technical staff" — and why they're the priority for AI upskilling now

Non-technical staff are the people who run the business without writing code: operations, marketing, finance, admin, HR, customer service. They read a business problem faster than most specialists because they live inside it. What they've usually never had is a structured way to apply AI to their own recurring work.

That's exactly where the demand sits. As of January 2026, 56% of employers whose businesses are using or planning to use AI rate their organisation's overall knowledge as "beginner" or "novice", and 61% of all employers have no staff currently working with AI (gov.uk, 2026-01-28). The gap isn't a shortage of data scientists. It's a shortage of ordinary teams who can use AI on the work in front of them.

The market has already picked a response. Among UK businesses using or unsure whether they use AI, the most common workforce move is upskilling existing staff — around 33% train or retrain, against roughly 10% automating roles and 4% recruiting new AI-skilled people (ONS Business Insights and Conditions Survey, late September 2025). Employers are choosing to build the skill inside the people they already have. That's the case for funding your existing team rather than hiring around them.

Our view: AI fluency is becoming one of the new basics, like spreadsheets before it. There's no academic gate on the door. "No degree needed" doesn't mean "no learning needed" — it means the barrier was never a qualification, it was never having done the work.

Route one: using your apprenticeship levy to fund AI training

If your organisation pays the apprenticeship levy, this is your route: use your levy pot to fund an apprenticeship whose standard develops AI and data skills your non-technical staff can apply. The levy is 0.5% of your pay bill over £3m, collected monthly through PAYE, sitting in your digital apprenticeship service account waiting to be spent (HMRC). Existing employees are eligible — an apprenticeship isn't only for new hires; your current ops, marketing or finance staff can be the learners.

From April 2026 the first Growth & Skills Levy products went live — the levy is being transformed from the Apprenticeship Levy, with short, flexible apprenticeship units fundable from 28 April 2026. The underlying mechanic — 0.5% of a pay bill over £3m — is unchanged. Apprenticeship funding policy now sits with the Department for Work and Pensions (DWP), which took it over on 16 September 2025; higher and under-19 education stays with the Department for Education. The apprenticeship standards themselves are maintained by Skills England, which replaced the Institute for Apprenticeships and Technical Education (IfATE) on 2 June 2025.

A note on the trap here. "Use it or lose it" levy thinking buys shelf-ware — training chosen because the budget existed, not because an outcome was named. Levy funds expire on a rolling clock (for contributions from 1 August 2026, that clock shortens to 12 months), which tempts employers to spend fast. Spend on the outcome, not the deadline.

For non-technical AI upskilling, iO-Sphere delivers AI Transformation at Level 4 (on the IS Business Analyst standard, ST0117) and Data & AI Essentials at Level 3 (on the Data Technician standard, ST0795) — both built for people from non-technical backgrounds.

Route two: levy transfer — fully funded AI training if your employer doesn't pay the levy

If your organisation doesn't pay the levy, you're not shut out. Under levy transfer, a levy-paying employer can move up to 50% of its unused levy funds to another UK business to fund that business's apprentices — a share raised from 25% to 50% on 22 April 2024 (gov.uk / CIPP). Larger firms routinely do this for supply-chain partners, smaller businesses and charities. For a non-levy-paying SME, a transfer can make an AI apprenticeship genuinely free at the point of use.

There's a distinction that changes who counts as "fully funded". An employer's own levy-funded apprentices are paid for from their own pot. A levy-transfer recipient is a different business having its apprentices funded by someone else's unused levy. Both can be fully funded — but as a recipient you depend on a transferring employer having spare funds and choosing to send them your way, so it's a route to line up early, not assume.

Where a transfer isn't available, government co-funding can still cover most of the cost, with the split depending on employer type and learner age. One dated fact worth planning around: for starts from 1 August 2026, apprentices aged 16–24 at non-levy employers are 100% government-funded (DWP apprenticeship funding rules, 2026-07-12). Beyond that, check the current DWP funding rules rather than relying on an older percentage.

What a genuinely non-technical AI training programme should cover

A programme built for non-technical staff should teach them to run the full loop on their own work: spot a problem, point AI at it, and judge whether what comes back is any good. The applied layer — using AI on your recurring tasks and validating the output — is the one most jobs need first and the one most training skips in favour of a tool tour.

Concretely, that means covering: using generative-AI tools (ChatGPT, Claude, Copilot) as a co-pilot on real tasks; enough data fluency to read and question what's in front of you; a working sense of what AI shouldn't be trusted with and where governance rules bite; and the judgment to frame a problem well and interrogate the answer. Not a lecture on machine learning. The test of a good syllabus is simple — at the end, can the learner point at work that changed? A report that didn't exist, a process automated, a weekly task cut from hours to minutes.

Keep the compounding in view. Small automations and small analysis wins look trivial one at a time and add up to serious hours over a year. Training that only measures big-bang outcomes misses where most of the return actually lands. Our guide to AI skills for non-technical professionals goes deeper on the specific skills to build first.

What to look for in a provider: applied practice, not theory-first courses

Judge a provider on one thing: how much of the programme is your staff doing the work versus watching someone explain it. In our experience people get good at AI and data by doing real tasks, coached by people who've done the job — not by sitting through theory-first lectures. That's the whole difference between capability and a completion certificate.

So ask the awkward questions. Who coaches the sessions, and have they actually done this work in a business, or do they only teach it? What will a learner have built by week six? Is there a real environment to practise in, or just slides? On our programmes that have it, learners work in Prism — a simulated e-commerce company built on 500M+ rows of real data — a sandboxed company, so the practice is real and the risk isn't.

The one route we don't deliver: data scientist

If what you actually need is someone building and training machine-learning models from scratch, that's a data scientist — a technical, code-heavy role, and not what these programmes produce. Our AI upskilling routes make non-technical staff fluent at applying AI and reading data, not at engineering the models underneath. If it's a modelling specialist you're after, look to a dedicated data science route elsewhere; we'd rather say so than sell you the wrong fit.

How do I tell a genuinely funded route from a paid one marketed as "free"?

On a government-funded apprenticeship, the learner is legally barred from paying towards their own training — that legal bar is what makes it free to them. A course advertised as "free" that asks the individual to contribute, or that turns out to be a lead-gen funnel for a paid product, isn't the same thing. If the working-time commitment, eligibility and employer backing aren't stated plainly and early, treat "free" as a headline, not a fact.

Eligibility, rules and disclosures to check before you enrol a team

Three things decide whether a funded plan survives contact with reality: eligibility, employer backing, and the working-time commitment. Say them plainly to your team before you enrol anyone — a reader who discovers them late walks away.

Working time is real. Apprenticeship funding requires protected off-the-job training hours, set per standard by the DfE rather than the old flat 20% rule (for new starts from 1 August 2025). Your team needs that time genuinely protected, not squeezed around the day job.
Existing staff qualify — but prior knowledge is netted off. Current employees can be apprentices. Recognition of Prior Learning means funding won't pay for skills the learner already holds; it reduces content, duration and price rather than adding to them.
Minimum duration. Any apprenticeship now runs a minimum of 8 months (reduced from 12 months for new starts from 1 August 2025), or the standard's published off-the-job hours, whichever applies.
This is England. The apprenticeship levy is reserved policy — Scotland, Wales and Northern Ireland run different systems.
The rules are in flux. Fund expiry, co-investment splits and the Growth & Skills Levy flexibilities are changing across 2026. Check the current DWP funding rules before you commit; our Growth & Skills Levy explainer tracks the mechanics.

What iO-Sphere believes about this

Funding is plumbing. It decides how training gets paid for; it says nothing about whether people get good. So we don't lead with "free". We lead with the work: name the capability you're buying, judge the programme on what learners will actually be doing and who coaches it, then let the funding follow.

Used that way, the levy is genuinely generous — the same plumbing can fund graduate schemes, retrain people at risk of redundancy, and lift the productivity of teams you already have. A structured funded programme is the grad scheme most employers couldn't afford to build, and it widens who you can hire: degree and non-degree, younger and older, a far broader range of backgrounds, all brought to real competency. The mechanics keep improving, and the direction of travel is generous — but the decision always runs outcomes-first.

How to get started: assessing your organisation's funded route

Start with the outcome, not the funding form. Write down what you want a specific team to be able to do in ninety days — the report they'll build, the process they'll automate, the decision they'll make faster — and how you'll see it in their work. Then work out the route: if you pay the levy, it's your own pot; if you don't, it's a levy transfer or government co-funding. Then judge the programme as if the money were coming out of your own budget.

The fastest way to know which route fits is to check your organisation's eligibility with someone who runs these programmes. We'll tell you plainly whether an apprenticeship suits your team or whether a paid short course or team training is the better fit — and if it isn't right for you, we'll say so.

Talk to us about your organisation's funded AI route →

FAQ

Is AI training really free for non-technical staff in the UK?

It can be genuinely free to the learner and the employer, but "free" means funded, not no-strings. The usual mechanism is the apprenticeship levy — your own money, if you pay the levy — or a levy transfer of up to 50% of another employer's unused funds if you don't (raised from 25% on 22 April 2024). Both come with eligibility rules, employer backing and a protected working-time commitment.

Can we fund AI training if my company doesn't pay the apprenticeship levy?

Yes. A levy-paying employer can transfer up to 50% of its unused levy funds to fund your apprentices, which can make an AI apprenticeship free at the point of use for a non-levy SME. Where a transfer isn't available, government co-funding covers most of the cost, with the split depending on employer type and learner age — check the current DWP funding rules for the exact figure, as they're changing through 2026.

Which apprenticeship covers AI skills for non-technical roles?

For non-technical staff, iO-Sphere delivers AI Transformation at Level 4 (on the IS Business Analyst standard, ST0117) and Data & AI Essentials at Level 3 (on the Data Technician standard, ST0795). Both are built for people without a technical background and both are fundable through the levy or a levy transfer. Apprenticeship standards are maintained by Skills England, which replaced IfATE on 2 June 2025.

Do non-technical staff need a degree or a coding background?

No. These programmes are designed for people from operations, marketing, finance and admin who've never written code. The barrier to AI fluency was never a qualification — it was never having applied the tools to real work. "No degree needed" doesn't mean "no learning needed"; it means the learning happens by doing the job.

How much working time will my team need to commit?

Apprenticeship funding requires protected off-the-job training hours, now set per standard by the DfE (for new starts from 1 August 2025) rather than a flat 20%. Your staff need that time genuinely protected. Say the commitment plainly before you enrol anyone — it's the thing most likely to derail a plan if it surprises people later.

What's the difference between the apprenticeship levy and the Growth & Skills Levy?

They're the same pot. From April 2026 the Apprenticeship Levy began its transformation into the Growth & Skills Levy, with the first products — including short, flexible apprenticeship units — live from that month. The core mechanic — 0.5% of a pay bill over £3m — is unchanged. Apprenticeship funding policy now sits with the DWP, from 16 September 2025.

Does this training make someone a data scientist?

No — and that's the wrong goal for most teams. A data scientist builds and trains machine-learning models; that's a technical, code-heavy role. These programmes make non-technical staff fluent at applying AI and reading data on their own work. If you genuinely need a modelling specialist, that's a separate, dedicated route elsewhere.

How do I know the training actually worked?

Judge it by changed work, not a completion rate. Completion measures attendance; the only return that matters is what a person can now do in the business that they couldn't before. The practical test: can you point at work that changed — a report that didn't exist, a task automated, a decision made faster? If the training can't survive that question, it was shelf-ware, however it was funded.