Your best tradie has been with you for eighteen years, which is why AI training for apprentices in trades can help preserve his hard-earned knowledge before he retires. He knows which councils are painful, which supplier will courier same-day if you ask the right person, and why you never touch a 1970s switchboard without checking one specific thing first. None of that is written down anywhere. He retires in three years, and the plan is that your apprentices “pick it up off him” before then.
They won’t. Not all of it. They’ll pick up maybe a third, because he’s on the tools, not teaching, and because half of what he knows he doesn’t even know he knows.
This approach is not about replacing that master tradie with a chatbot. It’s about getting the contents of his head into something your apprentices can question at 7am on a job site while he’s forty minutes away. Here’s how I do it, and what it actually costs you in time.
Key Takeaways
- AI training for apprentices in trades works by capturing what your senior tradies already know, then making it answerable on demand rather than filed in a folder.
- Most trades businesses lose 60-70% of a retiring tradie’s knowledge because it was never asked for, only observed.
- Voice interviews beat written SOPs for capture. Your best tradie will talk for 40 minutes and will not write two pages.
- Capture the edge cases, not the textbook. Apprentices get the standard method from their qualification, not from you.
- One structured hour per senior tradie per week for six weeks produces a knowledge base your apprentices can query forever.
- A knowledge base only helps if apprentices can reach it from a phone on site, mid-job, without asking permission.
- The same captured knowledge quotes jobs faster, answers customer questions, and briefs new hires in days instead of months.
Why Trades Knowledge Disappears Instead of Transferring
The apprenticeship model assumes proximity equals transfer. Put the young one next to the old one for four years, and the knowledge moves across. That worked when a tradie ran two jobs a week with a mate. It breaks when your senior guy is running five sites, answering the phone, and quoting on weekends.
There’s a second problem, and it’s bigger. Expert knowledge goes quiet. Once you’ve done something ten thousand times, you stop being able to explain it. Ask a master plumber why he chose that route for the pipe run and you get “it’s just better.” Push him, and you get four reasons he hasn’t consciously thought about since 2009. Cognitive scientists call this the expertise reversal effect: the more skilled someone becomes, the worse they get at teaching the basics, because the steps have collapsed into instinct.
So the knowledge doesn’t transfer by osmosis. An AI apprenticeship works when it helps transfer knowledge at the moment someone asks the right question. And your apprentice, standing there, doesn’t know which question to ask. He doesn’t know what he doesn’t know.
Then the tradie retires. Not slowly, usually. A back goes, or a grandkid arrives in Tauranga, and you get six weeks’ notice on eighteen years of judgement.
What you actually lose
Not the method. The method is in the qualification and on YouTube. What you lose is the local, specific, hard-won stuff:
Which of your suppliers will actually honour a Friday afternoon order. Which building inspector wants photos before the pour and which one wants a phone call. What that customer in Hobsonville is like when the invoice arrives. Why you always allow an extra half day on villas in Grey Lynn. The three jobs you’ve done where the obvious fix was wrong and what the real fix was.
That’s the material. It’s the material nobody documents, because it doesn’t feel like knowledge. It feels like just knowing the place.
How AI Training for Apprentices in Trades Actually Works
Here’s the practical shape of it. Three stages, and none of them require your senior tradie to write anything.
Stage one: capture by talking
Sit your best tradie down with a recorder for 40 minutes a week. Not a training session. An interview. You or your office manager asks questions; he talks. Transcribe the audio, feed it to an AI, and you have written material in minutes rather than never.
The questions matter more than the tech. Bad question: “How do you install a hot water cylinder?” He’ll recite the manual. Good question: “Tell me about a cylinder install that went sideways. What did you spot that a first-year wouldn’t?” Now you’re getting the actual content.
Run six of those sessions, and you have four hours of the most experienced person in your business explaining the parts of the job that aren’t in any textbook.
Stage two: structure it so it can be searched
Raw transcripts are useless. Nobody reads four hours of talking. The AI’s job here is to sort what he said into topics your apprentices will actually look for: job types, materials, councils, suppliers, customer situations, common failures.
This is where most businesses stop and shouldn’t. A folder of tidy documents is still a folder nobody opens. Ask any trades business owner about the SOP binder they made in 2022 and watch the face.
Stage three: make it answerable
The knowledge has to answer questions, not sit in a document. An apprentice on site types or says, “The meter box is on the neighbour’s side of the boundary; what do we do?” and gets your business’s answer, in your business’s voice, referencing the three times you’ve dealt with it before. Not a generic answer off the internet. Yours.
That’s the difference between documentation and an AI brain for your business. Documentation is passive. It waits. A brain gets asked and responds, and every answer is grounded in what your senior people actually said.
What This Does to Your Apprentices, Practically
Three things change, and they compound.
They stop interrupting. An apprentice asks roughly fifteen questions a day. Maybe three genuinely need a human. The other twelve are things the business already knows the answer to, and each one pulls a senior tradie off their own work. When the system answers those twelve, your senior guy gets an hour of his day back and stops feeling like a helpdesk.
They get better answers, faster. A rushed answer over the phone from someone driving between jobs is worse than a considered answer from a system that has your best tradie’s full explanation stored. The apprentice gets the reasoning, not just the instruction, which is the part that makes them independent.
They ramp in months instead of years. Not fully qualified in months; that’s not what I’m claiming. But productive on your specific jobs, with your specific suppliers and your specific quirks, in a fraction of the usual time. The generic trade skills come from their qualification. The business-specific knowledge is what usually takes three years, and that’s exactly the part you’ve just captured.
There’s a fourth effect nobody expects. Your senior tradie enjoys it. Being interviewed about what you know, by someone who’s actually listening, is one of the more respectful things that happens to a person late in their career. I’ve watched a 62-year-old sparky get genuinely animated about a job from 2011 that nobody had ever asked him about. He wasn’t being replaced. He was being recorded, which is close to being honoured.
The Wider Problem This Sits Inside
Knowledge trapped in one head isn’t only an apprentice problem. It’s the same shape as every other bottleneck in a founder-led trades business.
Your quoting knowledge is in your head. Your customer history is in your head. Which jobs are worth chasing and which ones smell like trouble, in your head. When your team can’t act without asking you, the business can’t grow past the number of questions you can answer in a day. Just like modern marketing teams look to a generative AI marketer apprenticeship to systematically capture and scale digital expertise, trade businesses need a reliable way to capture and scale their own operational know-how.
Capturing your master tradie is the same operation as capturing you. Same interviews, same structuring, same result: a business that can answer questions about itself without a specific person being available. That’s what I mean by an AI brain, and the AI workforce that sits on top of it, handling the calls and the follow-ups and the quote chasing, only works because the brain exists underneath.
Start with the retiring tradie because there’s a deadline on it. Then do yourself, because there’s a deadline on that too; you just can’t see it yet.
If you want the wider picture, I’ve written about why the founder becomes the bottleneck and about what AI adoption actually looks like in a smaller business. Both cover the same underlying gap from different angles.
What It Costs You in Time
Six weeks. Forty minutes a week from one senior tradie, plus an hour a week from whoever runs the interviews. That’s roughly nine hours total to capture the core of what someone knows after two decades.
Much like global workforce initiatives where industry leaders collaborate—such as partnerships between NABTU and Microsoft to upskill trade workers—structured knowledge transfer bridges the critical gap between traditional field experience and modern capability. Compare that to the alternative. When a senior tradie leaves without capture, the standard cost is six to twelve months of reduced output while the rest of the team works out what he used to handle quietly. Mistakes go up. Rework goes up. The owner absorbs the difference, usually by working more.
Nine hours against six months isn’t a close call. The reason it doesn’t happen isn’t cost. It’s that nobody has ever sat down and run the first interview, because there’s always a job on.
Where to Start This Week
Pick the person whose departure would hurt most. Book 40 minutes. Ask them one question: “What’s something you know about this trade that you’ve never seen written down anywhere?”
Record it. Don’t organise it, don’t build a system, don’t buy anything. Just get one conversation on tape and see what comes out. Almost every trades owner who does this is surprised by the first ten minutes.
Then do it again next week.
The technology is the easy part now. Turning a recording into structured, searchable, question-answering knowledge for AI training for apprentices in trades is close to solved. The hard part is still the same as it was twenty years ago: booking the time and asking a good question.
Ready to Get the Knowledge Out of People’s Heads?
If you’re looking at a retiring tradie, a growing apprentice crew, and a business that stops when a specific person is unavailable, that’s worth a conversation. Book a 30-minute discovery call, and we’ll talk through what’s actually trapped in your team’s heads and what it would take to get it out. No pitch deck, no prep needed, just a straight conversation about where your business depends on people rather than systems.
Frequently Asked Questions
Can AI actually train an apprentice in a trade?
No, and anyone selling that is overreaching. Hands-on skill comes from doing the work under supervision, and it always will. What AI does is capture and serve the knowledge layer: the reasoning, the local quirks, the war stories, the reasons behind decisions. It answers the questions an apprentice would otherwise have to interrupt someone to ask, which frees your senior people to teach the physical skills properly.
How do you capture knowledge from a tradie who won’t write anything down?
You don’t ask them to write. You record them talking. A 40-minute interview produces more usable material than any tradie will ever type, and it’s a far more natural format for people who think with their hands. Transcription and structuring are handled by the AI afterwards. Your senior tradie’s only job is to talk about work, which most of them will happily do for hours.
What questions should I ask to get the useful stuff out?
Ask about specific jobs, not general processes. “Walk me through the worst install you’ve done and what you’d check first now” beats “how do you install this.” Ask about exceptions, mistakes, and judgement calls. Ask what a first-year would get wrong. General questions produce textbook answers you already have; specific questions produce the knowledge that only exists in that person.
How long before apprentices see a benefit from this?
Sooner than most people expect, because the first interview alone usually produces two or three answers to questions your apprentices ask weekly. A useful base sits at around six sessions, roughly six weeks. Full coverage of one senior tradie’s knowledge takes a few months of steady capture, but it delivers value continuously along the way rather than only at the end.
Is this worth doing for a small trades business with two or three staff?
It’s arguably more important, because a smaller team has less redundancy. In a business of three, one person leaving takes a third of the operating knowledge with them. The capture work is also faster since there are fewer people to interview. The main question is whether anyone in the business is genuinely hard to replace. In most small trades businesses, at least one person is.
What happens to the captured knowledge if we change software later?
It stays yours. The material is text, stored as files you control, not locked inside a vendor’s platform. If a better AI model arrives next year, you point it at the same files. This is worth checking before you commit to any tool: ask where the knowledge lives and whether you can export it. If the answer is vague, walk.
Does this replace formal apprenticeship training?
Not at all, and it shouldn’t. Formal training covers standards, safety, and the qualification your apprentice needs to work legally and competently. What captured knowledge adds is everything specific to your business that no qualification covers: your suppliers, your councils, your customers, your methods. The two sit alongside each other, and the captured layer is the part nobody else can give your apprentice.
About Octavius
Titus Mulquiney is the founder of Octavius AI, where he builds AI brains and AI workforces for founder-led businesses stuck running everything out of their own head. Twenty years in marketing, ex-Sony product manager, ex-GM Zeal NZ. Based in Auckland, working with operators across NZ, Australia, and the US. Connect on LinkedIn.