You have probably already had a go at this. Most tradies have.

You opened ChatGPT, asked it for some posts, and it gave you some posts. They were fine. Perfectly fine.

And that is the problem. They read like every other post in your trade — because they were built the same way, from the same place, by the same thing, as everybody else’s.

Here is the bit nobody explains.

The Prompt Paradox

You have got the engine. That is not the same as being the mechanic.

Everybody has the same tool now. Your competitor has it. The bloke who quotes against you has it. It costs twenty dollars a month and it will write anyone a blog post.

So the tool is not the advantage. The tool was never going to be the advantage.

What makes it work is the thing around it — what it knows about your business, what you have told it, what it remembers, and whether any of that is written down anywhere.

An engine without the engineering around it is just noise. Loud, fast, expensive noise.

Which is why it did not work, and why that is not on you

Nobody sold you the engineering. They sold you the engine and let you assume it came with the rest.

Every article you have read about AI and small business is about the tool. Twenty prompts for tradies. Ten ways to use ChatGPT in your business. Not one of them tells you the tool is the cheap half.

So you did the reasonable thing with the information you had, got something that read fine, and concluded AI was overhyped. That was a correct conclusion from the evidence in front of you. It was just the wrong evidence.

Everything below is the half nobody sent you.

The Knowledge Moat

Everything we learn about your business gets written down, and stays written down.

How you talk. What you charge. Which jobs you actually want and which ones you take because you are quiet. What your customers praise you for in reviews. The objections you get at the kitchen table.

A bloke with a ChatGPT subscription starts from scratch every time he closes the tab. Most agencies start from scratch every month, because the person who knew your business has moved on.

Yours does not. Month two starts where month one finished. Month twelve is not month one again with different words.

That is the moat. After a year, nobody catches up to you without doing a year.

The Context Engine

This is the part that makes it sound like you and not like a robot.

It gets built from your real material — your best jobs, the quotes that got accepted, your reviews, the way you actually explain things to a customer standing in a laundry looking at a burst pipe.

Then everything that goes out is built from that, rather than from whatever the internet thinks a plumber sounds like.

What you get: copy that reads like you on a good day, using your words and your prices and your suburbs. What you do not get is a machine inventing things about your business, which is the single fastest way to lose a customer who checks.

The Velocity Protocol

Testing twenty things in a week instead of one thing in six.

Most agencies run a campaign, wait six weeks, and tell you it went well.

We would rather find out on Tuesday. A headline, a price frame, a photo, an opening line — you can run several versions of each inside a week and keep the one that actually worked.

The point is not speed for its own sake. It is that you stop paying for six weeks of being wrong.

Why a bloke who dragged buckets out from under houses builds your marketing

I started in trades. In Adelaide, underpinning houses.

Five or six hours a day on your hands and knees in a crawl space, dragging buckets of loose dirt the length of the house. The damaged pier is never near the opening. You dig with shortened picks and spades because there is not enough room to swing a full one. Hard, sweaty, backbreaking — and the smell when you first go under is like opening a sealed tomb.

Five months, early nineties. That was enough.

Then I got a job installing signs. Which led to selling them, then to making them — and eventually to more than fifteen years in construction, engineering and manufacturing: project manager on the jobs, business development for the rest of it. Signage and architectural metalwork, two divisions of the same business. Seven Westfield centres. Woolworths. TAB. NSW State Rail. Telstra installations around the country. Once you were in on the signage you usually got the metalwork as well.

And I have run ads for your trade before. Plumbers. Electricians. Builders. Locksmiths. I have been running Google Ads since 2005, which means I was buying clicks for tradies before most agencies pitching you existed.

I have held a White Card, Heights and Safety, and Confined Spaces. They have lapsed, because I am not on the tools any more and I am not going to pretend otherwise. But I sat through the inductions, I signed the paperwork, I wore the PPE — and I have written SWMSs, which is not a sentence most marketers can type.

So when I say I know what the bad end of your week feels like, I am not saying it to warm you up. I am also not claiming your job. I did five months of the worst work on a site and got out. You have done years of it.

And the project management is the half that decides whether you get your money’s worth

On a Westfield job there is no version where it nearly works. The centre opens on the date it was advertised, finished or not.

So you work backwards from that date, and you plan against what you have actually got — the crew, the crane, the lead time, the access window, the trade in front of you who is running late. Wanting more resource does not produce more resource. You sequence around the constraint or you miss.

You check the measurements before you order, not after. You have a plan for the thing that goes wrong, because something always does. You do not find out on the night that the permit was never lodged.

I have had the conversation where the crane is booked and the paperwork is not.

Marketing is mostly run by people who have never had that conversation. Which is why so much of it is a plan with no sequence in it — and why the agency report has green arrows on it and no jobs behind them.

So: two different things, and you need both from me.

The five months under the house is why I know what your week feels like. It is not why you should pay me.

The project management is why the build actually lands — worked backwards from a date, sequenced against what is really available, checked before it is ordered, and finished. Nothing clever. That is the standard this gets built to, and it is the one you can hold me to.

The Deep Research Layer

Trying the offer on before you spend money on it.

Before an ad runs or a price page goes up, we put it in front of simulated customers — built from the real language in your reviews and the real objections in your market — and see where it falls over.

It is not a crystal ball and I will not pretend it is one. It is a cheap way to find the obvious problem before it costs you three hundred dollars in ad spend to find it.

If a price frame confuses a simulated customer, it will confuse a real one.

So what does AI visibility actually mean for a tradie?

Plain answer, because this is where most of the jargon lives. And it works the same whether that is one ute or six.

When somebody needs a tradie now, a lot of them stop opening ten tabs. They ask their phone, and their phone gives them three or four names. Whoever is on that list gets the call.

There is no ad slot in an answer. You cannot buy your way onto that list the way you can buy your way to the top of Google.

What decides it is whether a machine can read you, quote you and trust you — which comes down to unglamorous things. Is your licence number written as text a machine can lift, or is it a photo of a card? Does your site say what you charge, or does it say “competitive rates”? Are your suburbs named, or does it say “servicing Sydney and surrounds”?

That is it. That is the whole of what AI visibility means for a trade business, and anyone making it sound more complicated is selling you the complication.

Has AI actually killed Google for finding a tradie?

No, and anybody telling you it has is describing a different search.

Here is the honest version, and it is against my own interest to tell you.

When someone types “plumber near me”, a 2026 study found an AI Overview appeared only 8% of the time — while the local map pack appeared on 99% of those same searches. The emergency call still goes through the old system.

But everything before the emergency has moved. How much should this cost. Do I need council approval. Is this a repair or a replacement. How do I check he’s licensed. Those questions get answered by an AI now, up to 93% of the time depending on the type.

So the honest position is this. The panic is oversold. The shift is real, and it is further upstream than anyone is telling you. If an agency says AI has killed your local search, they are describing the second thing and charging you for the first.

Both studies are American. Neither has an Australian equivalent, and no Australian study exists — which is part of why we started publishing our own.

How long before any of this works?

Phase one is fast. Phase two is not, and anyone promising otherwise is guessing.

Price bands, photo intake, quote template, follow-up, missed-call text-back — most of that is live inside the first month, and you should feel the difference in your quoting week almost immediately. That is the Kerbside Quote, and it is where the Ute Tax goes — the full working day a fortnight you currently spend driving to prices that could have arrived without you.

Getting named by the assistants takes 60 to 90 days before anything moves at all, and it moves unevenly. Some months nothing happens. We publish our own log of exactly that, including the flat months, so you can see what it actually looks like rather than what a case study says it looks like.

How much of this lands on me?

One sitting, properly, at the start. Your real prices, your licence number, your suburbs, and the three jobs you actually want more of.

That conversation is the build. Without it there is nothing to publish, and I will not invent numbers on your behalf — a made-up price band is worse than no price band, because the first customer who holds you to it is a real one.

After that it is on me. You approve things by text. There are no standing meetings. If a month goes by and I have not needed anything from you, that is the system working rather than me forgetting about you.

Can I just do this myself?

Some of it, genuinely yes.

Publishing your prices costs nothing. Putting your licence number on your site as text costs nothing. Answering the six questions a homeowner asks before she rings anyone costs nothing but an afternoon.

If you do those three things and nothing else, you will be ahead of most of the peninsula. I would rather tell you that than pretend the whole thing is locked behind a retainer.

What is hard to do yourself is the part that needs to be maintained — the monthly measurement, the listing worked until it scores, the content that keeps arriving whether or not you had a bad week. Not because it is clever. Because it is relentless, and you already have a job.

Your Tuesday morning, six months from now

Now	Six months
You quote at night, after dinner	You ballpark from the ute, off a photo she sent
Three quotes go out, one comes back, you never find out why	Every quote gets a follow-up at 48 hours, automatically
Your prices live in your head	Your prices are on your site, and she has half decided before she rings
She asks her phone and it names someone else	You are on the list
“How is that bloke flat out and I’m not?”	You know your number, and you can watch it move

What it costs

The Snapshot	Free. Ninety seconds, self-serve. Most people stop here and that is fine
The Revenue X-Ray	A$670 — credited in full against the build if you go ahead within 30 days
The Speed-to-Quote build	A$2,670 for the founding five, A$3,470 after. Days 1 to 90
Monthly	A$870, six-month minimum, then month to month

Year one at founding pricing is A$13,110.

I do not discount the monthly. If something has to move, the build moves, and never below A$2,610.

The two things I will promise, and nothing else

Every enquiry answered in under three minutes, or the month is free. That covers the automated first response. You can test it by ringing your own number, which is the point.

Founding five only — if your Google listing is not at 18 out of 20 within 30 days, month one is free.

That is the list. No “more calls”. No “page one”. No “ChatGPT will recommend you”. Anyone promising you those is promising something they do not control, and you should ask them how they plan to deliver it.

One per trade, per catchment

Brookvale–Dee Why. Frenchs Forest. Manly. Mona Vale. Terrey Hills. One business per trade in each.

Not scarcity for its own sake. I cannot publish price bands and answer the same homeowner questions for two plumbers in Dee Why without one of them being the decoy.

So the practical version: if a sparkie in Frenchs Forest signs, Frenchs Forest sparkies are closed. Not for a launch window — until he leaves. If your trade is taken in your patch I will say so on the first call, not after you have paid for a diagnostic.

---

When I would tell you not to bother

• If you are genuinely flat out three months ahead and turning work away. Second van, not an agency…
• If you will not publish a price. The entire first phase depends on it. If that is a hard no, the offer does not work, and I will say so on the first call rather than the third month…

Run the free Snapshot → · What the Revenue X-Ray involves →

Or ring me. Rowan Shead, 0420 745 357.

Questions

Wondering whether your trade is already taken in your catchment? That one is answered on the homepage →