Level 4 Apprenticeship / ST1512

UK Government approved Skills England standard ST1512 Level 4 / LARS 828 Quality assured by Ofqual Funding band up to £18,000

AI and Automation Practitioner

A UK government approved qualification in AI and automation.
Up to 100% funded.

It is a nationally recognised Level 4 standard, not a private course with a certificate at the end. And it does not just teach your staff about AI. It trains someone already on your team to identify where your business is leaking hours, then build the thing that fixes it, on a real process, inside your own systems.

Most employers pay just 5%. If your apprentice is under 25 and your payroll is under £3m, you pay nothing at all.

Book a 20 minute call See how it works

15 months · around 7 hrs a week · delivered live online

How much time is one process taking? Live

People doing the repetitive task

Hours a week each, on that task

Hours a year on this one task932

That is working days124

Working days back if automation removes 40%50

Worked on 46 working weeks and 7.5 hour days. Deliberately no pound figures here, because that depends on who is doing the work. Putting a defensible cost against these hours is the first thing your apprentice is taught to do, using your real numbers.

The difference

Two skills, and almost every course only teaches the second one.

Knowing how to use an AI tool is not the same as knowing which problem is worth pointing it at. This programme deliberately spends its first five months on the first skill, before anyone is allowed to build anything.

Step one / Identify

Find where the time actually goes

Your apprentice audits real working time in 30 minute blocks, maps how the work genuinely happens rather than how the process document says it does, and puts a cost against it.

Duplicated data entry between systems that never got connected
Work sitting in an inbox for two days before anyone picks it up
Reports rebuilt by hand every month
Information trapped in a spreadsheet only one person understands
Processes where AI is genuinely the wrong answer, and saying so

Step two / Solve

Build the fix, then prove it worked

They design it around the people who actually use it, build it in low-code and no-code tools, test it properly, get your team using it, and report what it saved in real numbers.

A working build by month three, not month fifteen
The main automation live in your systems by month ten
Documented and handed over, so it survives them leaving
Governance, security and monitoring built in, not bolted on
A costed report by month twelve showing what it actually delivered

Orientation

The role, and what this technology genuinely does and does not do.

Months 1 to 2

Discovery

Mapping real processes, auditing data, designing around people.

Months 3 to 5

Build

Prompting, automation platforms, integrations, testing.

Months 6 to 9

Embed

Adoption, change, governance and life after go-live.

Months 10 to 11

Value

Measuring it honestly, and assessing who it affected.

Months 12 to 13

The funding

Up to 100% of the cost is paid by the UK government.

This is not a grant you apply for and might not get. It is how apprenticeship funding already works, and this qualification sits inside it. Which of the three applies to you depends on your payroll and the age of the person you put forward.

Apprentice under 25, payroll under £3m 100% You pay nothing

Fully funded by the government. For most small and medium businesses putting forward a younger member of staff, there is no fee to you at all.

Payroll under £3m 95% You cover 5%

The standard position for most SMEs. The government funds 95% of the cost and you contribute the remaining 5%.

Levy payer, payroll over £3m Levy Money you already pay

This comes out of a pot you already pay into every month. Unspent levy funds now expire after 12 months, so if you do not spend it on training like this, you lose it.

Funding band for this standard is up to £18,000. If you are a levy payer whose pot for the year is already exhausted, co-investment on anything beyond it is 25% rather than 5%. Worth checking before you assume either figure.

What you get, and when

You do not wait fifteen months to find out whether it worked.

Most apprenticeships hand you a certificate at the end. This one is sequenced so something real exists in your business early: a working build in month three, a second in month five, and the main automation in month seven. This is the same month by month plan we issue to your training provider, so nothing on this page is a sales version of it.

Thirteen taught sessions across fifteen months, each applied to your process that month. One live online group session, four hours, plus around five hours a week of their own working time. Not evenings, and not a person removed from your team. Across the fifteen months it totals 420 off-the-job training hours, against a 420 hour minimum.

MonthWhat is covered

Phase 01OrientationMonths 1 to 2

The Practitioner Role, and How AI and Automation Actually WorkThe occupation and their own role mapped to the 15 duties. Off-the-job training explained properly, so the log starts in week one.

Responsible AI: Law, Ethics and GovernanceUK GDPR in automated processes. Automated decision-making rights, and where the law says a human must stay involved.

Phase 02DiscoveryMonths 3 to 5

Process Discovery, Mapping and Opportunity AnalysisCheckpoint build 1As-is before to-be. Interviewing the people who do the work and cannot describe how they do it.

Data Foundations: Quality, Bias and ErrorProfiling a real dataset live. Preparation, conversion, and spotting the errors that break an automation later.

Designing for People: Requirements and Human-Centred DesignCheckpoint build 2Gathering requirements from people who do not know what they want. Writing user stories that survive contact with reality.

Phase 03BuildMonths 6 to 9

Working with AI Tools: Prompting and EvaluationSpecifying output precisely. Grounding answers in real documents, and scoring quality rather than judging by impression.

Building with Low-Code and No-Code AutomationCheckpoint build 3Choosing a platform for the organisational context. Building live, including letting things fail on purpose.

Connecting Systems: APIs, Integrations and AI ServicesEvent-driven patterns. Choosing an AI service, and connecting systems that were never designed to talk.

Testing, Assurance and DocumentationDesigning test cases that find real defects. Assessing viability, and making an honest go-live recommendation.

Phase 04EmbedMonths 10 to 11

Adoption, Enablement and ChangeResistance treated as information rather than obstruction. Engagement planning, and training your real users.

Governance, Security and Life After Go-LiveGovernance and accountability. Identifying vulnerabilities, and monitoring what happens once you stop watching.

Phase 05ValueMonths 12 to 13

Proving the Value: Productivity and Cost-BenefitBaseline, measure, attribute. Counting time properly and honestly, including what did not work.

Fairness at Scale and Portfolio ConsolidationAlgorithmic impact assessment end to end. Workforce equality monitoring, and consolidating the evidence.

Phase 06CompletionMonths 14 to 15

Consolidation and EPA practiceNo new teaching. A group consolidation session, portfolio finalised, and practice for the end-point assessment.

GatewayThe formal gate into end-point assessment, which is then carried out separately by the assessment organisation.

Full course facts, including entry requirements, assessment and costs: read the fact sheet or download the PDF.

The obvious questions

What business owners actually ask.

What is the catch, if the government pays?

There is not one. This is exactly what apprenticeship funding exists for and has done for years. This particular standard is just new. You are not the exception, you are precisely who it is for.

We cannot spare anyone for fifteen months.

It is not full time. It is around seven hours a week, and most of that is the project work someone would be doing anyway if this problem were being solved properly. You are not losing a person, you are pointing one at a problem you already have.

We already use ChatGPT or Copilot.

So does most of the market, and most of it reports no measurable gain. Using a tool and knowing how to build something durable and governed with it are different skills. This is the difference between someone who prompts and someone who ships.

What if the process turns out not to be automatable?

Then you will know that for certain instead of guessing, which is worth having on its own. Recommending against automation where it will not help is explicitly part of what they are assessed on, and there is a second process ready to go.

What if they finish and then leave?

Around two thirds of apprentices stay with the employer who trained them. And even where they do not, the automation stays: it is built inside your systems, documented and handed over before anyone goes anywhere.

We are too small for this.

Smaller businesses are exactly who it is built for, and the under 25 rule means many pay nothing. Size is not the barrier. Having a real problem worth solving is the requirement, and most businesses have several.

A UK government approved qualification in AI and automation.Up to 100% funded.