No essays. No 10,000-word assignments. How the AI apprenticeship is actually assessed.

On a client call this week, a senior leader who recently completed a data apprenticeship described the experience: three interim projects of a few thousand words each, a final project, a presentation, then presenting the lot again for assessment. His verdict was blunt: you lose a lot of learning time describing what you have just done. His teams had discussed it internally, and the writing burden was actively putting people off enrolling on anything similar.

He is not wrong, and he is not alone. We hear the same story from HR directors, operations managers and IT leads every week. So here is the straight answer to the question underneath it: what does assessment actually look like on the AI & Automation Practitioner Level 4?

Many established standards evidence learning through written assignments because, for a long time, writing was the only auditable trace of thinking. The result, on some programmes, is the experience our caller described: thousands of words of interim projects, a written final project, then a presentation about the writing. The learning is real, but a meaningful slice of programme time goes to documentation instead of capability. For busy operational staff, that trade-off is exactly backwards.

The AI & Automation Practitioner Level 4 is a new standard, and it was designed task-first. Every module ends in a doing-task set in your business, not an essay about the topic. Month 2’s task is mapping the processes in your own team and ranking what would survive automation. Month 3’s is shipping a live, no-code automation that removes a repetitive job, permanently. Month 4’s is building an AI agent on your own data that does work rather than answering questions about it. The write-up for each is short and practical: what was built, what changed, what you would do differently, with screenshots. Writing supports the work; it never replaces it.

There are no assignments at all on this one, just tasks they actually do. By month four they’ll have built an AI agent to help them with their role. The project is a problem they’ve spotted in your business, we sign it off around halfway through, and the only bit at the end is a showcase, where they show off what they built. Rod Doyle, Director, TESS Group

Let’s be precise, because this is a regulated apprenticeship and we will not pretend assessment vanishes. It is a 12-month programme plus a 3-month End Point Assessment window, delivered with BCS, The Chartered Institute for IT. What changes is the shape. The workplace project is not an academic exercise: it is a problem the apprentice has spotted in your business, the kind that eats six hours every week, and the solution is the thing they have learned to build. That project is typically agreed and signed off between halfway and three quarters of the way through the programme, not crammed into a terminal exam season.

What remains at the end is a showcase: the apprentice presents the portfolio of what they actually built across the programme, the automation, the agent, the assured rollout, the project. For the employer in the room, it looks less like a viva and more like a demo day, because that is what it is.

Three things. First, the evidence trail is made of working artefacts in your business, so the programme pays for itself as it runs rather than promising value at the end. Second, manager support time goes into coaching around real work instead of proof-reading assignments, which is most of what the manager’s guide covers. Third, the people most likely to refuse an essay-heavy programme, your experienced operational staff, are exactly the people this standard was designed for. The full L4 guide walks through every module; the 8-month accelerated route exists for teams that need it faster.

If your teams have sworn off apprenticeships because of the writing burden, this is the one to re-open the conversation with. Show them the module list on the Claude Apprenticeship page, point at month 3 and month 4, and ask one question: would you rather write about automation, or ship one? For leaders who want to feel the teaching quality before committing anyone, the free AI Leadership taster days in July and August are the no-obligation sample.