Everyone Bought the Same Machine. Almost Nobody Did the Thinking First.

I have sat on both sides of this table. Client side, watching a budget get carved up by people who couldn't tell you what any of it earned. Agency side, being the friend somebody paid to bring along without ever asking what the friend actually did. From both chairs the pattern is the same, and AI has not changed it. It has just made it faster.

Here's the thing. The whole market is panicking about AI, and the panic is aimed at the wrong target. The fear is that the machine will beat you. It won't. What will beat you is scaling a process you never understood in the first place. Point a machine at a broken funnel and you don't fix the funnel. You break it at volume.

Fifteen years ago the tools were the edge. Having proper data instruments was like having a car on your phone forty years ago. Now everyone has a mobile. The tools are ubiquitous, as ordinary as a phone in your pocket, and near enough every marketer is using them. When everyone holds the same instruments, the instruments stop mattering. The edge moves to the thinking that decides where to aim them. That is the part almost nobody has done.

Notion surveyed more than 6,000 AI decision-makers and users across 10 global markets for its report on AI at work, covered by Search Engine Journal in July. They sorted everyone into four levels of maturity. Level 1 is AI as a thought partner, a better search engine. Level 4 is AI running the work end to end.

The split is brutal. 57% are stuck at Level 1. Another 31% at Level 2, using it as an assistant. Only 10% have it working like a teammate, and 2% have built it into the actual system. Add that up. 88% are using AI as a tool. Only 12% have built it into something that reshapes how the work gets done.

Here's the part that matters. The 12% are not winning on better prompts. They do three unglamorous things at much higher rates. They plug AI into systems they already run (55% versus 37%). They put oversight around it (42% versus 26%). They measure what it actually did with real metrics (37% versus 22%). The advanced lot have even stopped counting "time saved", which is a number you make up to feel good, and started measuring error rates and cycle time instead.

That is not a prompting gap. That is a thinking gap. The 12% did the boring commercial work first. They knew their process well enough to hand a piece of it to a machine and check whether the machine improved it.

The Australian numbers make it worse, and they make it local.

Depending on which survey you read, AU business AI adoption runs anywhere from 37% to 68%. Scale Suite pulled the SME-specific figures together from MYOB, CSIRO and the National AI Centre. MYOB's November survey of over 1,000 small businesses put SME adoption near 29%. The National AI Centre tracker says closer to 37%. So plenty of businesses are switching it on.

Then look at what they do with it. Only 7% of SMEs have put AI into their actual product or service, and that number went backwards. Most AI use is internal admin. Fine. Here is the line that should stop you.

82% of AI-using SMEs say it had a positive impact. But 46% of them do not measure that impact at all, and of those, 74% said measurement was "unnecessary". Run the maths and roughly 38% are claiming a win with no way of knowing whether they won.

That is grading your own homework. It is the oldest trick in marketing, dressed up in new clothes. The attractors always say they're doing a great job. They're incentivised to. Now the machine says it too, and everyone nods, and not one person can point to the number that moved.

Scale Suite's own conclusion, off the back of mid-market firms reporting more than double the revenue growth of smaller ones, is blunt. The gap runs through operational maturity, not the AI. Their line: AI without operational maturity does not deliver. Correct.

I'm not theorising here. We have scored the Australian market across six marketing dimensions, and the same hole shows up over and over. Most businesses cannot tell you their own conversion rate. They do not measure retention. They cannot say which channel actually earns its place versus which one just feels busy.

They are flying blind on the fundamentals. Your website is your shopfront. A thousand people walk in and you have no idea how many walk out with anything. In a physical shop you'd have a camera and a register and you'd work out the conversion rate by Tuesday. Online, if you don't capture it yourself, you lose it. You can't get it back.

Now hand that business an AI tool. It will write more copy, spin up more variations, launch more campaigns, faster than ever. None of it is pointed at anything. That is not sprinkles on a broken cake. That is an industrial sprinkle machine on a cake that was never any good to begin with.

So what would I actually do about it. Nothing clever. The clever move here is refusing to be clever.

Start with two questions. Do you know how much money you make? Do you know how much money you spend? You can work back from there. Everything else derives from those two inputs. I'm not saying you need to be an accountant. All I'm saying is you have to understand what's going on in your own business before you let a machine loose inside it.

Then find the constraint. Every business has one thing that is bleeding out. A mobile experience so bad that fixing the website alone jumps you 20% overnight. A channel you keep funding that would change nothing if you turned it off tomorrow, which is exactly the test. Turn it off. If revenue drops, you learned something. If nothing moves, why were you paying for it.

Only then do you aim the machine. Once you know your numbers and you know your constraint, AI becomes what it should be. A throttle. You scale when you know. You scale back when you don't. The machine doesn't decide the direction. You do. It just moves you there faster once you've decided.

That order is the whole game. Numbers, then constraint, then machine. Get it backwards and you've built a very expensive way to lose with confidence.

Here's my call. The next two years do not belong to whoever has the best AI, because everybody has the same AI. They belong to whoever did the unglamorous work of knowing their own business before they scaled it.

The 12% who built AI into a real system aren't smarter. They just refused to skip the commercial thinking that the other 88% skipped. In Australia the bar is lower still, because near half of the businesses using AI aren't even measuring whether it works.

That is not a threat. Read it again. It is a gift, and it's sitting there for whoever wants it. While the market panics about the machine, the actual edge is the least fashionable thing in marketing. Know your numbers. Find the constraint. Then, and only then, let the machine run. Do that and the panic isn't your problem. It's your opening.