If you’re an employer trying to plan your 2026 AI hiring, or a learner deciding between two routes, you’re looking at the same fundamental question: do you need a practitioner who can configure existing AI tools into business workflows, or do you need an engineer who can design, build and maintain machine learning systems from the ground up?
This guide compares the two main UK apprenticeship standards that answer this question, then gives you an interactive tool that uses the official Knowledge, Skills and Behaviours (KSBs) from each standard to score which one fits.
The Two Standards at a Glance
| Dimension | AI & Automation Practitioner (L4) | Machine Learning Engineer (L6) |
|---|---|---|
| Reference | ST1512 v2.0 | ST1398 v1.0 |
| Level | Level 4 (HNC equivalent) | Level 6 (Bachelor’s degree equivalent) |
| Typical duration | 18 months | 24 months + 4 month EPA |
| Min off-the-job hours | 420 | 418 |
| Max funding | £18,000 | £22,000 |
| Approved for delivery | 10 December 2025 | 18 December 2024 |
| Lars code | 828 | 795 |
| Route | Digital | Digital |
| EQA provider | Ofqual | Ofqual |
| Professional recognition | Multiple embedded qualifications (BCS, Microsoft, NCFE) | BCS Advanced RITTech |
| Progression routes | Data engineer, Data scientist (integrated degree), AI data specialist | AI data specialist |
“The biggest mistake employers make is treating these as alternatives. They’re not. Level 4 trains the people who’ll use AI to change how the business works. Level 6 trains the people who’ll build the AI itself. Most UK businesses need both, in different proportions.”
What the Standards Actually Cover
AI & Automation Practitioner (Level 4) — ST1512
The Level 4 apprenticeship is a practitioner standard. The official six Assessment Outcomes (AOs) are:
- AO1: Strategic and Ethical Adoption of AI Automation — aligning solutions with organisational values, ethics and workforce needs; change management; impact assessment
- AO2: Solution Design and Development — designing and configuring AI/automation solutions using appropriate tools, platforms and integration methods
- AO3: Testing, Evaluation and Iteration — testing solution viability through user feedback and feasibility analysis; iterating to optimise performance
- AO4: Governance, Assurance and Ongoing Risk Management — assurance activities; algorithmic impact assessments; supporting compliance, transparency and accountability
- AO5: Stakeholder Engagement and Workforce Enablement — engaging diverse stakeholders; delivering training; communicating technical concepts in accessible formats
- AO6: Continuous Improvement and Change Delivery — identifying productivity opportunities; supporting sustainable change initiatives
The performance descriptors expect an apprentice to apply AI knowledge to “well-defined but complex and non-routine problems” with autonomy “within set parameters.” Distinction-grade requires applying skills with “flexibility and operational fluency.”
In practice, the L4 apprentice uses tools like Microsoft Copilot, Anthropic Claude, Google Gemini, n8n, Make, Zapier and Power Automate to ship working business automation. They configure rather than code. Most successful L4 apprentices come from operations, finance, project management, marketing operations or customer experience backgrounds.
Machine Learning Engineer (Level 6) — ST1398
The Level 6 apprenticeship is an engineering standard. The standard’s nine duties cover the full ML engineering lifecycle:
- Implementing ML/AI solutions in a safe, trusted, responsible manner
- Planning the engineering development of ML applications and frameworks
- Developing, testing, staging and building in a pre-production environment, prototyping ML products including experiment tracking
- Monitoring and supporting ML models through operational deployment in the live environment
- Monitoring operating resource implications and developing scalable, environmentally sustainable systems
- Delivering responsive technical engineering support services to mitigate operational impact
- Maintaining collaborative stakeholder relationships and providing auditable records at each decision point
- Ensuring compliance with data governance, ethics and cyber security
- Keeping up to date with technological engineering developments
The Knowledge requirements (K1–K31) explicitly include programming languages and modern ML libraries (K16), the relationship between mathematical principles and core ML techniques (K18), feature engineering (K9), software development best practices including version control, CI/CD (K31), and the stages of the ML lifecycle (K2). Behaviour B5 says the apprentice “Operates in settings of technical complexity and uncertainty.”
This is a different kind of apprentice. They typically come in with a software development, data analysis, science or engineering background and a strong technical aptitude. They’ll spend the apprenticeship building, deploying and operating production ML systems.
Who Each Apprenticeship Is For
Pick Level 4 if…
- Your candidate is a mid-career operations, finance, project management, customer experience or marketing professional
- They are comfortable with technology but don’t code professionally
- The business need is configuring AI tools (Copilot, Claude, Gemini, n8n, Zapier) into workflows that solve real business problems
- You want measurable productivity uplift in 6–12 months
- You don’t need bespoke ML models built from scratch
- You want governance, compliance and stakeholder engagement skills alongside technical configuration ability
Pick Level 6 if…
- Your candidate already has a software development, data analytics, science or engineering background
- They’re comfortable with mathematical concepts (statistics, linear algebra, calculus) and want to go deeper
- The business need is to build, deploy, monitor and maintain custom ML/AI models — not just use someone else’s
- You can support a 24-month + 4-month EPA timeline
- You have data infrastructure (data engineering, MLOps tooling, CI/CD) or are committed to building it
- You operate in technical complexity and want depth, not breadth
🧠 Which AI Apprenticeship Is Right for You?
Built from the official Skills England KSBs for both standards. Six questions, takes 90 seconds.
Salary and Career Path
| Stage | Level 4 trajectory | Level 6 trajectory |
|---|---|---|
| During apprenticeship (UK avg) | £28,000–£38,000 | £32,000–£48,000 |
| Post-completion (year 1) | £42,000–£58,000 | £55,000–£78,000 |
| Mid-career (5 years post) | £65,000–£90,000 (Lead/Head of AI Operations) | £90,000–£130,000 (Senior ML Engineer) |
| Senior (Director / Principal) | £100,000–£160,000 (Head of AI / Director of AI Strategy) | £130,000–£200,000+ (Principal ML Engineer / Architect) |
| Most common next role | AI Operations Lead, Head of AI, AI Programme Manager | Senior ML Engineer, ML Architect, AI Specialist |
Salary ranges are approximate UK norms in mid-2026; actual ranges vary by region (London materially higher), sector and individual progression. Both apprenticeships have strong post-qualification earning potential, with the L6 ceiling materially higher reflecting the deeper technical specialism.
How TESS Group Delivers Each
| Need | Programme | Length |
|---|---|---|
| L4 Practitioner training | AI & Automation Specialist Level 4 apprenticeship | 15–18 months, levy-funded |
| L4 same content faster | AI & Automation L4 Accelerated | 8 months, levy-funded |
| L4 Microsoft 365-heavy environments | AI Copilot Apprenticeship L4 | 15–18 months, levy-funded |
| L4 Google Workspace-heavy environments | AI Gemini Apprenticeship L4 | 15–18 months, levy-funded |
| L6 Machine Learning Engineer | AI & Machine Learning Fellowship Level 6 | 24 months + 4-month EPA, levy-funded |
| L5 leadership for execs accountable for AI | AI Leadership Pathway (AU0009/AU0010/AU0011) | Levy-funded |
| Quick uplift on prompt engineering | AI Prompting Accelerator | Short programme |
| Whole-team AI literacy | Building AI-Ready Teams short course | 1 day |
| Microsoft Copilot end-user mastery | Microsoft Copilot short course | 1 day |
| Google Gemini end-user mastery | Google Gemini short course | 1 day |
| Cross-tool prompt engineering depth | Ultimate Prompt Engineering | 1–2 days |
| AI risk & governance leads (single unit) | AI Adoption & Governance Unit | Levy-funded modular unit |
The most common combined pattern at UK enterprise clients: 3–5 mid-career staff on the Level 4 apprenticeship shipping workflow automation, plus 1–2 software-engineering-background staff on the Level 6 ML Fellowship building the underlying capability, with the senior leadership team taking the Level 5 AI Leadership Pathway for governance accountability.
“If you treat L4 vs L6 as ‘cheaper or pricier’ you’ve missed the point. They produce different roles. The right answer for most UK enterprises is ‘both, in the right ratio for our business.’”
Frequently Asked Questions
Which is harder, Level 4 or Level 6?
Level 6 is more technically demanding because it requires writing code, applying mathematical principles to ML model design, and operating with minimal supervision in technical complexity. Level 4 is configuration, integration and governance work using existing AI tools — challenging but at a different level of abstraction. Level 6 is roughly equivalent to a Bachelor's degree level of technical depth; Level 4 is roughly equivalent to a Higher National Certificate.
How much do they each cost the employer?
Maximum funding is £18,000 for the AI & Automation Practitioner Level 4 and £22,000 for the Machine Learning Engineer Level 6. Both are funded through the Apprenticeship Levy / Growth and Skills Levy for levy-payers, or 100% government-funded for SMEs (employers under £3m payroll).
How long does each take?
Level 4 takes 18 months minimum (ST1512 v2.0). Level 6 takes 24 months plus a 4-month End-Point Assessment window (ST1398 v1.0). Both require 20% off-the-job learning time, equivalent to roughly one day per week.
Can someone do Level 4 first then Level 6?
Progression from L4 to L6 isn't a single named pathway in Skills England's standards, but it's the natural career direction for an apprentice who completes ST1512 and wants to specialise in deeper ML engineering. The ST1512 explicitly lists 'Data engineer', 'Data scientist (integrated degree)' and 'AI data specialist' as example progression routes. ST1398 lists 'AI data specialist' as its own progression route.
Which one teaches Python and machine learning model building?
Level 6 (Machine Learning Engineer ST1398) explicitly covers programming languages, modern ML libraries, model training, feature engineering, and deployment. Level 4 (AI & Automation Practitioner ST1512) is configuration-focused — using LLMs and no-code workflow tools rather than building models from scratch.
What if my team needs both — practitioners AND ML engineers?
Run cohorts at both levels. Many UK employers do exactly this: 3-5 mid-career operations and finance staff on the Level 4 to ship workflow automation, plus 1-2 software-engineering-background staff on the Level 6 to build and maintain the underlying ML capability. The two roles are complementary, not competitive.
Is there an entry requirement difference?
Both require Level 2 (GCSE) English and maths to be completed during the apprenticeship if not already held. Level 6 in practice usually requires existing technical foundation — software development, data analysis, or science/engineering background — to land successfully within the 24-month timeframe. Level 4 is genuinely accessible to mid-career staff from any function.
Which one has more embedded qualifications?
Level 6 includes BCS Advanced RITTech professional recognition. Level 4 (as TESS Group delivers it) embeds up to 5 industry qualifications including Microsoft AI-900, BCS, NCFE Prompt Engineering and AI Governance. Quantity favours Level 4; depth and seniority of the BCS Advanced RITTech recognition favours Level 6.
Not sure which programme fits your team?
Our directors will walk you through it on a 30-minute call — bringing your team profile and levy budget. We’ll show you which apprenticeship (or which combination) is the right fit, even if that’s a multi-provider mix. No hard sell.
Book a Free Discovery Call