The Level 3 route into data and AI runs on the Data Technician standard. The level only works if your day job can hold the practice, so check that before you apply or sponsor.
Level 3 AI Apprenticeship: Standard, Funding & Fit
By James Cotton · Last updated · 15 min read
By James Cotton, Founder, iO-Sphere
The question most people arrive with is "which level can I get onto?" Entry is rarely the hard part. The question that decides whether a funded programme works is "which level's work genuinely exists in my job?" Get that wrong and the apprenticeship fails without ever looking like a failure: the learning detaches from the day job and becomes classroom theory with extra admin. Get it right and it compounds, because you are practising the actual work while you are paid to do it.
What a Level 3 AI apprenticeship actually is
A Level 3 AI apprenticeship is the entry-level, work-based route into data and AI. In England it runs on the Data Technician standard (ST0795), approved by Skills England and currently at version 1.1. There is no apprenticeship standard formally titled "AI apprenticeship" at Level 3, so if you search for one, ST0795 is the standard you will actually enrol on: the AI-titled standards sit higher, starting with the AI and Automation Practitioner at Level 4 (ST1512). Level 3 itself sits at roughly A-level standard on the national qualification scale.
The standard covers the foundations of working with data: sourcing and collecting it, cleaning and preparing it, storing it safely, analysing it, and presenting it so a non-technical audience can act on it, with data protection, data quality and the responsible use of AI and automation running through all of it. Everything more advanced in this field builds on those competencies. We deliver the route as Data & AI Essentials.
Is it the same as an academic AI course?
No, and the difference is the point. An academic course teaches AI as a subject, front-loading the theory and testing it by recall in an exam. An apprenticeship is built around the work. You learn by doing the job, coached by people who have done it, and the end-point assessment judges what you actually produced: typically a project or portfolio review and a professional discussion, run by an independent assessment organisation. A knowledge test appears only where a standard's assessment plan specifically requires one.
The test that picks the level
Every apprenticeship carries protected off-the-job training time. For ST0795, the published minimum is 348 hours, all of which must happen inside normal, paid working hours; catching up at home in the evening is not permitted. The funding rules guarantee the hours. What they cannot guarantee is that the hours get spent on meaningful work, because that depends entirely on the match between the level and the role.
So run the test: can your role fill 348 protected hours with real data work at Level 3? Ask the employer plainly what work they will give you at that level. The syllabus cannot supply what the role does not contain.
The same test tells you when Level 3 is too low. If your day job already routinely includes building data pipelines, writing SQL that stakeholders rely on, or producing analysis people act on, the Level 3 competencies sit beneath your current practice, and enrolling would be box-ticking. That is the sign to enter at Level 4: our Advanced Data & AI programme (the Level 4 Data Analyst standard, ST0118) or AI Transformation (ST0117; the AI-titled Level 4 standard, the AI and Automation Practitioner ST1512, is a genuine current route, and not one we run). Our Level 4 guide walks the same decision from the other side, and weighs the two Level 4 standards honestly.
Levels map to the complexity of the work, not to seniority or age. An experienced professional whose role already holds Level 4 work enters at Level 4; someone whose role holds foundational data work starts at Level 3 because that is where the practice lives. The wrong level is the main way a funded programme quietly fails, so this test comes before the funding, before the syllabus, before everything.
Who can apply: entry requirements and the role you need
You do not need a degree for any apprenticeship, and the standard itself sets no entry bar: Skills England publishes no entry requirements for the standards we deliver, so entry criteria are set by each provider. The one requirement that does come from the funding rules is English and maths at Level 2, meaning a GCSE at grade 9 to 4 (a standard pass, the old grade C) or an Ofqual-regulated Functional Skills Level 2. For apprentices aged 16 to 18, achieving it is a condition of completing.
For anyone starting at 19 or over, the English and maths qualification stopped being compulsory on 11 February 2025: the employer decides whether to ask for it, and if you lack it you can still be funded to study up to Level 2 where your employer agrees. At iO-Sphere we assess English, maths and digital footing at the start of the programme, not as a hurdle but as groundwork that finds small gaps before they slow you down.
What you do need is a job whose real work contains data. That role rarely has "data" in its title: people apply from finance, marketing, operations, customer service and plenty of other functions, because data now runs through most jobs. There is no requirement for a data team or a data job title. The test is whether the role's day-to-day work lets you practise and evidence the standard's duties. If it does, the apprenticeship can be laid over the job you already hold: your employment and salary carry on as they are, and the training wraps around the work.
How long it takes, and where the hours come from
The legal minimum for any apprenticeship is 8 months or 187 off-the-job hours, whichever the standard requires (reduced from 12 months on 1 August 2025). In practice that means no apprenticeship can run shorter than 8 months, and no standard's training-hours minimum can fall below 187. The floor is not a typical duration, and an individual standard's assessment plan can require longer: ST0795's published typical duration is 24 months (Skills England). We deliver our apprenticeship programmes in 15 months of training plus a 3-month end-point assessment, with accelerated options a few times a year.
The off-the-job hours come from the standard, not from a rule of thumb. ST0795's published minimum is 348 hours, and we smooth those across the 15 months of training, which works out at roughly 5.4 hours a week. If you already hold some of what the standard teaches, the programme shrinks to fit through recognition of prior learning: the content you can already evidence comes out, and the duration and the price come down with it. No reduction can go below the 8-month, 187-hour floor.
What it costs the employer in 2026-27
Measured by what the training is worth, a full Level 3 apprenticeship is a five-figure investment. Measured by what the employer hands over, it is usually little or nothing, because government funding covers the price. Each standard carries a funding band that caps what the levy pays: for ST0795 the band is up to £13,000 (Skills England), and bands are revised periodically. The money comes through the Growth & Skills Levy (formerly the Apprenticeship Levy), charged at 0.5% of an employer's pay bill above £3 million; from April 2026 the levy also began funding shorter, more flexible training alongside full apprenticeships.
What the employer actually pays depends on two things: pay bill and the apprentice's age. Any start you plan now lands in the 2026-27 funding year (starts from 1 August 2026):
- Apprentice aged 16 to 24 at the start: £0. The government funds 100% whether or not the employer pays the levy, and even where a levy payer's funds are already spent (2026-27 funding rules v3, 29 July 2026).
- Apprentice aged 25 or over, non-levy employer: the employer pays 5% of the training price and the government funds 95%.
- Apprentice aged 25 or over, levy payer with insufficient funds: the employer pays 25% and the government funds 75%.
"Non-levy" means a pay bill under £3 million; headcount is irrelevant. And not paying in does not mean missing out. For an SME the costs are the ones listed above, nothing for a 16 to 24 start and 5% of the price for an apprentice 25 or over, and a levy transfer can remove even that 5%. A larger employer can pass on unused levy funds to another business, up to half of them since April 2024 (raised from a quarter), and a transfer often covers an SME's entire training cost. The full mechanics are on our funding options page.
Since 28 April 2026 the levy can also fund short apprenticeship units: 30 to 140 hours of training carrying no regulated qualification. Our view is plain. The funding is small and the format constrains depth, so a unit suits a team that already has data practitioners and needs one bounded skill topped up. Building foundational capability from scratch is a full apprenticeship's job, and no unit substitutes for it.
One ordering rule saves employers the most money of all: decide the capability first, then fit the funding to it. A funded programme at the wrong level changes nothing about what anyone can do, and it spends the dearest resource a team has, which is working time. Answer the level question first and the funding arithmetic takes two minutes.
What you're paid during it, and what it opens after
The training costs you nothing as the apprentice: its price is settled between the government and your employer, and no part of it passes to you. What you earn depends on how you enter. If you take the apprenticeship in the job you already hold, your salary simply continues: adding the training changes nothing about your pay.
For new entrants there is a legal wage floor: £8.00/hour from 1 April 2026 for under-19s or those in their first year, but that's a floor, not the going rate (gov.uk minimum wage rates). Once past the first year, anyone aged 19 or over moves to the minimum wage for their age under the same rates: £10.85 an hour at 19 and 20 (the statutory 18 to 20 rate), and £12.71 from 21.
A completed Level 3 opens the entry-level data and AI roles: data technician, junior analyst, reporting and insight work. It is also the funded step to Level 4, where the Advanced Data & AI and AI Transformation routes sit, and to Level 5 Data Engineering beyond that. The ladder keeps going above the levels we deliver: Level 6 holds both the Machine Learning Engineer standard (ST1398) and the Data Scientist integrated degree (ST0585), and the AI Data Specialist (ST0763) sits at Level 7. If that is your destination, Levels 3 and 4 build the foundation, and we will point you onward.
The demand underneath those roles is measured, not assumed: in the Department for Education's Employer Skills Survey 2024, employers linked 50% of their skill-shortage vacancies to a lack of complex analytical skills, up from 43% in 2022 (UK-wide figures). On starting salaries after completion, ranges vary widely by sector and region, and they belong beside their sources: our entry-level data analyst salary guide holds the current measured and advertised figures with the basis of each.
How iO-Sphere delivers the Level 3 route
We think you learn this work by doing it, coached by people who have done the job rather than only taught it. Data & AI Essentials is delivered by practitioner coaches in small cohorts built around live discussion and peer challenge, not a video library. We consistently see work-embedded cohorts sustain motivation better than self-paced study, because a cohort and a coach create accountability that self-study never will. The programme line is also new by design, built after generative AI changed what data work looks like, so there is no pre-AI curriculum underneath it and nothing was bolted on.
We have trained 900+ learners in data and AI since 2022 across our bootcamps and funded qualifications, we are rated 4.8 out of 5 on Google from 78 reviews, and a growing cohort of apprenticeship learners now sits across UK employers. The level test protects the model: pitch the level above the work the role can hold and there is nothing for a coach to work with, so the programme slides back into classroom theory. Match the level to the work and the coaching has something to attach to.
Who should not start at Level 3
Level 3 is the wrong choice in three situations.
- Your role already routinely demands Level 4 work. You would be re-learning things you can do. Enter at Level 4 through Advanced Data & AI instead.
- Your destination is Level 6 or 7 work, such as machine learning engineering or data science. A Level 3 or 4 builds the foundation and the progression route towards it, not an equivalent, and we will say so and point you onward.
- Your role cannot supply data work at all. No apprenticeship at any level fixes an absent job. The routes that need no employer are our fee-paying Applied Diploma in Data Analytics, 14 weeks full-time ending in a regulated NCFE Level 4 qualification, and a funded Skills Bootcamp, where cohorts run in windows.
How to start: apply first, conversation second
If you are the learner, apply before you ask. Apply on the Data & AI Essentials page, and our admissions team will help you build the business case and prepare you for the conversation with your employer. Managers mostly turn down a vague request made on the spot; a prepared case, with the funding worked out, is far harder to turn down. A softer opening works too: raise it in a development conversation to feel it out, and speak to us in parallel.
The economics do most of the persuading, especially for younger apprentices. An employer sponsoring an apprentice who starts aged 16 to 24 pays nothing for the training, and employers pay no employer National Insurance on apprentices under 25 earning up to £50,270 a year (per gov.uk National Insurance guidance).
If you are the employer, start with a scoping conversation rather than a product pick. What does the role actually contain, and what should the person be able to do at the end? Answering that is how the level gets matched to the work before anyone signs anything.
One or two people are enough: a small team joins an open cohort alongside peers from other companies, and there is no minimum on your side. Explore sponsoring on our employers page. If your team's need sits above Level 3, or a short course would serve it better than a full apprenticeship, we will say so.
FAQ
What Level 3 apprenticeship covers AI in the UK?
The Skills England-approved Data Technician standard (ST0795) is the Level 3 route that covers foundational data and AI work in England: there's no standard formally titled "AI apprenticeship" at Level 3. It builds the entry-level competencies of collecting, cleaning, analysing and presenting data and applying AI and automation responsibly. It's assessed on real work through an end-point assessment, the statutory minimum duration is 8 months, and the formal entry requirement is Level 2 English and maths. iO-Sphere delivers this route as Data & AI Essentials over 15 months plus a 3-month assessment.
Are there other Level 3 routes covering AI besides ST0795?
For data-first AI work at Level 3, ST0795 is the standard. The AI-titled apprenticeship standards all sit higher: the AI and Automation Practitioner (ST1512) at Level 4, the Machine Learning Engineer (ST1398) at Level 6 and the AI Data Specialist (ST0763) at Level 7. So a Level 3 entry into AI starts with the data foundations, and that order is no accident: data is the substrate AI runs on.
Do you need a degree for a Level 3 AI apprenticeship?
No, you don't need a degree for any apprenticeship. The formal requirement covers English and maths at Level 2 (GCSE grade 9 to 4 or an Ofqual-regulated Functional Skills Level 2), and for anyone starting at 19 or over even that stopped being compulsory on 11 February 2025; the employer decides whether to ask for it. What you do need is a job whose real work contains data tasks, so the protected off-the-job hours can be filled with genuine work at that level.
Can I do the apprenticeship in the job I already have?
Yes, and it is the main route we run: existing employees can be apprentices, and the programme is laid over the role you already hold, so your job and salary continue unchanged while the training wraps around the work. The role does not need a data title; it needs enough real data work to practise and evidence the standard's duties. Apply on our website and our admissions team helps you build the business case and prepares you for the conversation with your employer.
How much does a Level 3 apprenticeship cost the employer?
For starts in the 2026-27 funding year (from 1 August 2026), an apprentice aged 16 to 24 carries no employer co-investment: the government funds 100% where levy funds don't cover the cost, and where they do, the training is paid from funds the employer has already set aside. For apprentices 25 and over, a non-levy employer pays 5% of the training price and a levy payer whose funds are already spent pays 25% (2026-27 rules v3, 29 July 2026). Non-levy means a pay bill under £3 million, never a headcount, and SMEs can also fund fully through a levy transfer. The funding band for ST0795 is up to £13,000 (Skills England).
How long does a Level 3 AI apprenticeship take?
The legal minimum for any apprenticeship is 8 months or 187 off-the-job hours, whichever the standard requires (reduced from 12 months on 1 August 2025), but that's a statutory floor, not a typical length, and a standard's assessment plan can require longer. iO-Sphere delivers its apprenticeship programmes in 15 months of training plus a 3-month end-point assessment, with accelerated options a few times a year. Recognition of prior learning can shorten a programme but never below the 8-month floor.
How is a Level 3 AI apprenticeship assessed?
On real work, not exam recall. The standard's assessment plan sets the method: typically an end-point assessment built on a project or portfolio and a professional discussion, delivered by an independent assessment organisation. A knowledge test appears only where a standard's plan specifically requires one; it's the exception, not the default.
Is a Level 3 apprenticeship the right level for me?
It's the right level if your current role genuinely contains foundational data work that can fill the 348 protected off-the-job hours. It's the wrong level if your work already demands Level 4 tasks such as building pipelines or writing SQL for stakeholders, in which case the Level 4 Advanced Data & AI route fits better, and it's equally wrong if the role can't supply data work at all. The practical test comes before the syllabus: ask your employer what work they'll actually give you at that level.
Ready to check the match? Apprentices can apply through Data & AI Essentials, and employers can start a scoping conversation.
Build a data-literate workforce
Data apprenticeships funded through the Growth & Skills Levy: £0 for any apprentice under 25, and most of the cost covered for older starts.