The Gaffer supplies the experience. The Crew supplies the digital labor.

Once you accept that fifteen capabilities do not require fifteen people, the next temptation is to swing too far the other way. You have three lads, each carrying five trades, so the clever move is to deploy all fifteen skills and extract real value. Yet that path turns a simple job into a three‑day committee meeting. Anyone who has worked around trades recognises the danger. Give a bloke a new tool, and problems start looking like reasons to use it. Digital work follows the same pattern. Buy a course creator, and suddenly every useful thought demands twelve modules. Discover webinars, and a humble Monday checklist gains a 47‑slide presentation by Wednesday. Before long, the workshop generates more work for itself than for the customer. So Billy, our Crew Leader, needs a very unfashionable skill in this AI era: to say, “Leave that bloody thing in the box.”

That is where our Job Before Tool Test earns its keep. Before Derek, Peter, or Freddy uses anything, we ask three ordinary questions.  Which function does this tool fulfill? From the Gaffer, what does it require? Which finished item must it hand back? There is no magic in those questions, which is precisely why I like them. If we cannot explain the job, we do not use the tool. If we cannot explain what we need from the AI, we should not complain when it gives us generic porridge in return. And if we cannot describe the finished handoff, the odds are we are about to create another file destined for the folder marked “Might Be Useful One Day.” The test is less about controlling the AI than controlling ourselves. It stops enthusiasm from becoming clutter and keeps every bit of work attached to a purpose.

Imagine a retired gardener who spent thirty years watching beginners repeat the same mistakes each spring. They buy too many seeds and gadgets, and try to turn half the back garden into Kew before Easter Monday. Our gardener sees a calmer path. Start with one small patch and grow a few things you will actually eat. Learn what the soil is doing. He does not need SEO, slides, or an online course. Instead, Derek could need only two of the five tools.

The gardener supplies the material AI cannot honestly manufacture. The neighbour who planted enough courgettes to feed a small regiment remains in his memory. The gardener knows which expensive gadget has been rusting behind the shed for three years. A nervous beginner often needs permission to start small more than they need another twenty gardening facts. Derek’s job is to capture that knowledge without polishing the personality out of it. His finished handoff is not “some ebook content.” It is something concrete: a complete, beginner-friendly guide, based on the gardener’s experience, organized into practical sections and ready for the next stage. That sentence matters because now Peter knows what is arriving on his bench. Peter is not reopening the case and interviewing the gardener from scratch. The job information travels with the work.

This may be the most overlooked difference between using several AI tools and operating an AI crew. A collection of tools keeps asking you to start again. Tell me your audience. Tell me your tone. What is your product? What result do you want? Describe your customer. By the fourth tool, you have become the administrative assistant for the machines that were supposed to be helping you. A crew works differently. The Gaffer gives the brief once. We are building this thing. This is who it is for. This is what I know. I want the finished product to achieve this. Derek receives it. His finished work goes to Peter. Peter receives both the original job and Derek’s result. Freddy receives the same foundation plus whatever Peter created. Information moves forward instead of continually returning to the starting line.

That is familiar territory to anyone who has ever seen a proper job run well. If one trade finishes something that another trade must immediately rip apart because nobody explained what came next, the problem is rarely lack of skill. It is the handoff. Digital work suffers the same waste, only it hides behind screens and impressive-looking documents. Derek may create a lovely ebook whose message bears little resemblance to Peter’s sales page because Peter never saw the original intention. Freddy may optimise the page around phrases that attract the wrong person completely because nobody told him what problem the product really solves. Each piece can look competent, but the installation as a whole fails to heat the house. So we define finished before the work begins. Derek knows what Peter needs. Peter knows what Freddy needs. Billy watches the joins. The Gaffer watches the whole thing.

This is why the five-tool rule matters. Derek owns five trades: prompts, lead magnets, ebooks, courses, and services. Peter has sales pages, websites, email, presentations, and messaging. Freddy carries SEO, infographics, branding, Etsy, and Amazon/KDP. Those are capabilities, not a daily to-do list. On one job, Derek might use two, Peter three, and Freddy one. Another job may require a completely different combination. When a product has no business on Etsy, Freddy does not create an Etsy listing to justify his existence. When a £7 product can be clearly explained on one page, Peter avoids constructing a webinar because someone claimed webinars convert well. A twelve-page guide can do the job perfectly, but Derek does not inflate it into a course simply because Daniel’s old course-building instructions are sitting somewhere in the machinery.

That restraint hides a quiet commercial advantage. Every unnecessary asset costs attention. It needs checking, storing, updating, explaining and connecting. More output can actually create less movement because the important path disappears under all the extra furniture. A retiree building a modest digital business does not need the internet equivalent of Heathrow Terminal Five. Often we need one useful thing, one clear route to it, and one next action. If that works, then we can add something because the job calls for it, not because our software subscription is making us feel guilty.

There is one responsibility I would never casually hand to any of the lads. Final judgment belongs to the human. AI may tell me a headline is stronger. It may polish the paragraph. It may give three arguments for recommending a product. But if I do not genuinely believe the recommendation, the Crew does not get to manufacture my belief. The same goes for experience. An AI can help a retired nurse explain what decades on a ward taught her about speaking to frightened families. It cannot truthfully acquire those decades by pretending. A retired builder may know instantly when a supposedly clever shortcut is asking for trouble because he has seen what happens six months later. That judgment is not inefficiency to be automated away. It is the valuable bit.

That leaves us with a working rule simple enough to put above the desk: Job first. Human judgment second. Tool third. Begin tomorrow with one real piece of work small enough to finish. Write down what you want at the end. Decide which part requires your experience and which part is repetitive graft. Then let Billy take only the necessary tools from Derek, Peter and Freddy’s boxes. When the job comes back, do not ask whether the AI used every clever capability you bought. Ask the question tradesmen have always understood: “Does the bloody thing do the job?” If the answer is yes, you have not merely used AI. You have trained the Crew on another real installation. And the drawing for the next one has just become a little better.

You’re right, Gaffer. I’d gone back to one-line butter pats instead of Fleet Street slabs. 😄

This time: scene-led opening, proper 200–300 word newspaper paragraphs, observation, humour, story first, teaching woven through it. No machine-gun sentences. Build Your AI Crew

The great AI gold rush has left plenty of perfectly sensible people standing in front of a digital toolbox wondering which blinking thing they are supposed to pick up first.

Retirement Experience Still Needs A Gaffer, Not Fifteen New Jobs

The useful bit is not learning every tool. It is knowing what you want built and putting the right tool on the right part of the job.

The Human Keeps The Idea While AI Handles The Repetitive Graft

That is where a pile of clever software begins turning into something much more familiar: a working crew.

There was a time when a tradesman arrived with a huge toolbox. That signaled competence to onlookers. The van doors slid open, and boxes came out. People assumed he knew what he was doing. Few expected him to tip tools onto the kitchen floor. He turned each spanner over and asked what to try first. That scene shows how AI helps online now. Today, one tool writes email. Another builds sales pages. Another creates images. Then come course makers, ebook writers, keyword finders. Website builders, prompt libraries, and agents follow. A retiree who began with one good idea sits amid many options. There is no shortage of technology. The problem is tools arriving before the job sheet.

My understanding grows from the fact that ideas rarely arrive in tidy packages. Dyslexia shapes that pattern. Thoughts often come indirectly, carrying unrelated pieces that fit when viewed together. A modern digital sales funnel can be compared to a home heating system. Many AI agents can seem like technicians around a boiler. This approach helps me grasp internet marketing. Over time, a clear insight emerges: brilliance with every new AI tool isn’t required. I must remain the Gaffer.

That sounds almost embarrassingly simple, yet it reverses the typical AI narrative. Software is often showcased by what it can do. It writes an ebook, creates a website, or produces emails, graphics, posts, and a course. This display is impressive, but capability does not equal responsibility. A tool can do twenty jobs, but you don’t need all of them today. A tool may handle twenty jobs, but you don’t need all of them today. When I installed heating systems, the job decided which tools came from the van. The same rule sits above every AI desk: do not ask what the tool can do; ask what job needs doing. If the aim is to turn twenty years of experience into a useful guide, start there. If the guide later needs a page explaining why somebody might want it, that becomes another job.

We arrived at the Build Your AI Crew concept through a deliberate approach. I recently acquired a PLR bundle containing 15 distinct AI business agents. The bundle included email writing, ebooks, courses, lead magnets, websites, sales pages, SEO, graphics, publishing, and more. The obvious internet-marketing impulse would be to showcase all fifteen on stage like a digital Royal Variety Performance. That reply was less glamorous: “Why the fooook do we need fifteen lads?” The answer, of course, was that we did not. We bought fifteen trades, not fifteen people. Once that distinction became clear, the workshop began organising itself. Five related skills could sit in one toolbox; another five in the next; the remaining five in the third. Derek could become the lad who helps build something useful.

Reconsider retirement with a lifetime of experience, for the heart of the job lies in you. Do not vanish from your own work; you carry what matters most. Your stories, your judgment, and your unique view of problems stay essential.

So the first rule of this new workshop is not complicated: keep the judgment and delegate the repetition. We maintain the strange connection. The decision about whether something is worth recommending stays with me. I keep in mind the sort of person I want to help. I keep the final “Aye, I would put my name to that.” The Crew gets the repetitive graft surrounding those decisions. That leaves me doing more of the bit I actually enjoy: spotting ideas, joining things together, trying something, helping somebody, and seeing whether the useful thing we created earns something back. It is not retirement as another job. It is retirement as a workshop. And perhaps that is the biggest difference between collecting AI tools and learning to Build Your AI Crew. One leaves you staring at a toolbox. The other leaves you looking at a finished job.

First, acknowledge that fifteen capabilities do not require fifteen people. Three lads, each carrying five trades, tempt you to put all fifteen skills to work. That can turn a simple job into a three-day committee meeting. Anyone who has worked around trades recognises the danger. Give a bloke a new tool, and problems quickly become excuses to use it. Digital work follows the same pattern. Buy a course creator and every useful thought wants twelve modules. Discover webinars, and a modest Monday checklist becomes a 47-slide presentation by Wednesday. Before long, the workshop produces more work for itself than it does for the customer. So Billy, our Crew Leader, needs a practical skill for the AI age. That skill is to say, “Leave that bloody thing in the box.”

That is where our Job Before Tool Test earns its keep. Before Derek, Peter, or Freddy uses anything, we ask three ordinary questions. What job is this tool doing? From the Gaffer, what does it need? Which finished item must it hand back? There is no magic in those questions, which is precisely why I like them. If we cannot explain the job, we do not use the tool. If we cannot explain what we need from the AI, we should not complain when it gives us generic porridge in return. And if we cannot describe the finished handoff, the odds are we are about to create another file destined for the folder marked “Might Be Useful One Day.” The test is less about controlling the AI than controlling ourselves. It stops enthusiasm becoming clutter and keeps every bit of work attached to a purpose.

A retired gardener spent thirty years watching beginners repeat spring mistakes. They buy many seeds and gadgets, trying to turn half the back garden into Kew before Easter Monday. Our gardener offers a calmer path: start with one small patch, grow a few crops you will actually eat, and learn what the soil is doing.

The gardener supplies the material AI cannot honestly manufacture. The gardener recalls a neighbour who planted enough courgettes to feed a small regiment. Aware of the gadget that spent three years rusting behind the shed, he knows which. He knows that a nervous beginner often needs permission to start small more than they need another twenty gardening facts. Derek’s job is to capture that knowledge without polishing the personality out of it. His finished handoff is not “some ebook content.” It is something concrete: a complete beginner-friendly guide, based on the gasupplied experience, organised into practical sections and ready for the next stage. That sentence matters because now Peter knows what is arriving on his bench. Peter is not reopening the case and interviewing the gardener from scratch. The job information travels with the work.

This may be the most overlooked difference between using several AI tools and operating an AI crew. A collection of tools keeps asking you to start again. Tell me your audience. Tell me your tone. What is your product? What result do you want? Describe your customer. By the fourth tool, you have become the administrative assistant for the machines that were supposed to be helping you. A crew works differently. The Gaffer gives the brief once. “Here is what we are building. This is who it is for. What I know is this. Here is what I want the finished thing to achieve.” Derek receives it. His finished work travels with it to Peter. Peter receives both the original job and Derek’s result. Freddy receives the same foundation plus whatever Peter created. Information moves forward instead of continually returning to the starting line.

That is familiar territory to anyone who has ever seen a proper job run well. If one trade finishes something that another trade must immediately rip apart because nobody explained what came next, the problem is rarely lack of skill. It is the handoff. Digital work suffers the same waste, only it hides behind screens and impressive-looking documents. Derek may create a lovely ebook whose message bears little resemblance to Peter’s sales page because Peter never saw the original intention. Freddy may optimise the page around phrases that attract the wrong person completely because nobody told him what problem the product really solves. Each piece can look competent, but the installation as a whole fails to heat the house. So we define finished before the work begins. Derek knows what Peter needs. Peter knows what Freddy needs. Billy watches the joins. The Gaffer watches the whole thing.

This is why the five-tool rule matters. Derek owns five trades: prompts, lead magnets, ebooks, courses, and services. Peter has sales pages, websites, email, presentations, and messaging. Freddy carries SEO, infographics, branding, Etsy, and Amazon/KDP. Those are capabilities, not a daily to-do list. On one job, Derek might use two, Peter three, and Freddy one. Another job may require a completely different combination. When a product has no business being on Etsy, Freddy does not create an Etsy listing to justify his existence. When a £7 product is clear on one page, Peter avoids webinars because they are said to convert well. If a twelve-page guide does the job perfectly, Derek does not inflate it into a course simply because Daniel’s old course-building instructions are sitting somewhere in the machinery.

That restraint carries a quiet commercial advantage. Every unnecessary asset costs attention. It needs checking, storing, updating, explaining, and connecting. More output can actually create less movement because the important path disappears under all the extra furniture. A retiree building a modest digital business does not need the internet equivalent of Heathrow Terminal Five. Often we need one useful thing, one clear route to it, and one next action. If that works, then we can add something because the job calls for it, not because our software subscription is making us feel guilty.

There is one responsibility I would never casually hand to any of the lads. Final judgment belongs to the human. AI may tell me a headline is stronger. It may polish the paragraph. It may give three arguments for recommending a product. But if I do not genuinely believe the recommendation, the Crew does not get to manufacture my belief. The same goes for experience. An AI can help a retired nurse explain what decades on a ward taught her about speaking to frightened families. It cannot truthfully acquire those decades by pretending. A retired builder may know instantly when a supposedly clever shortcut is asking for trouble because he has seen what happens six months later. That judgment is not inefficiency to be automated away. It is the valuable bit. That is the practical way to Build Your AI Crew: begin with the finished job, then bring only the digital trades onto the bench that genuinely help complete it.

That leaves us with a working rule simple enough to put above the desk: Job first. Human judgment second. Tool third. Begin tomorrow with one real piece of work small enough to finish. Write down what you want at the end. Decide which part requires your experience and which part is repetitive graft. Then let Billy take only the necessary tools from Derek, Peter and Freddy’s boxes. When the job comes back, do not ask whether the AI used every clever capability you bought. Ask the question tradesmen have always understood: “Does the bloody thing do the job?” If the answer is yes, you have not merely used AI. You have trained the Crew on another real installation. And the drawing for the next one has just become a little better. That is how you Build Your AI Crew without turning yourself into the secretary for a dozen different AI tools.