The whole market has bought the same AI tools, so the tools stopped being the edge. The edge moved to the thinking that decides what to point them at, and almost nobody has done it. Here is why knowing your own numbers beats bolting AI onto a process you never understood.
The Take: Every serious competitor now runs the same AI stack, so the software stopped being anyone's edge months ago. The contest moved back to something older: whether an operator understood their own commercial reality before letting an automated system loose on it. Most Australian operators skipped that step.
Fifteen years watching budgets get spent, from the client seat and the agency seat both, has taught me one thing that repeats with almost boring consistency. The instrument changes. The mistake doesn't. Someone gets a faster piece of kit and uses it to do more of whatever they were already doing, without first checking whether the underlying activity was any good.
AI hasn't broken that habit. It's just given it a turbocharger.
Why doesn't licensing AI count as an advantage anymore?
Because the software is now near identical across every serious rival in a category. Once everyone in a market runs the same generation of instrument, the instrument stops mattering and judgement becomes the whole contest again, exactly like it always was before the tool arrived.
Clean analytics used to separate serious operators from the rest of the pack. Now it's assumed, boring, table stakes. Automated systems are riding the identical curve, only faster. Boardroom anxiety keeps aiming at "will a machine replace us," when the sharper risk is quieter: scaling an unexamined process at a pace that used to require a whole department to reach.
Point automation at a leaky funnel and the funnel doesn't get repaired. It gets broken harder, faster, with a straighter face and a bigger invoice.
What proportion of firms have actually embedded AI into how they operate?
Just 12% of organisations have wired automated systems into a genuine operating rhythm. The remaining 88% treat it as an add-on, sitting outside the workflow rather than threaded through it.
Notion's research surveyed over 3,000 knowledge workers worldwide and ran 85-plus interviews with leaders piloting these systems day to day. The gap between appetite and reality is wide. 93% call automation essential to modern productivity software. Just 20% say it's genuinely woven into their company's existing stack. Wanting a capability and building it into daily practice turn out to be separate undertakings entirely.
88% of firms treat automated systems as a bolt-on. Only 12% have woven them into a workflow that actually changes outcomes
The frontrunners in that research don't win through cleverer prompting. Notion's findings show they connect automation to systems already in place at far higher rates. They also wrap review steps around outputs and track results against real metrics rather than gut feel. Separately, 40% of the teams surveyed admit they lack basic approval steps for using these tools responsibly, a shortfall that predates any of this by years.
What does the local adoption picture reveal about the accountability gap?
Close to four in ten Australian small businesses using automated tools claim a positive result while running no method to confirm it. That isn't proof of success. It's an absence of anyone checking.
Local adoption estimates swing between 29% and 68%, depending on the survey and how loosely the word "adoption" gets stretched, according to Scale Suite's review of MYOB, CSIRO and National AI Centre figures. The tightest, most relevant read comes from MYOB's November 2025 survey of 1,087 small Australian firms: genuine usage sits at 29%. A mere 7% have built these systems into a customer facing product or service. The rest is largely internal admin.
MYOB pressed further, asking what happens once a tool goes live. 82% of users report a favourable outcome. Yet 46% track nothing at all, and 74% of that group called tracking "unnecessary." Multiply that through and close to 38% of local operators are reporting a win they have no way to substantiate.
Roughly 38% of Australian small businesses using automated tools report a positive result with no way to verify the claim
That reflex predates every piece of software in this story. Whoever runs a channel tends to insist it's paying off, because they have every incentive to. Now the software backs them up too, and there's no yardstick left in the conversation to push back.
What does New Rebellion's benchmark data show about this exact pattern?
A meaningful slice of scored businesses are pouring effort into acquisition through channels they cannot properly measure, the identical pattern showing up in a completely separate dataset.
Across New Rebellion's benchmark scoring of Australian and New Zealand businesses, the typical business scores noticeably stronger on Acquisition Performance, the discipline of attracting customers, than on Data & Tracking, the discipline of measuring what happens once they land. In roughly one out of twenty businesses scored, that split runs 15 points or wider.
An operator chasing acquisition hard while measuring weakly is precisely the profile most exposed once automation gets switched on. More campaign variants, more copy, more spend, all aimed at a funnel too blurry to correct.
Methodology note: figures drawn from New Rebellion's benchmark dataset of Australian and New Zealand businesses scored across six marketing dimensions (Digital Maturity, Acquisition Performance, Conversion Efficiency, Retention & Loyalty, Brand & Positioning, Data & Tracking), scored as at May 2026. The 4.5% figure and dimension averages come from 729 businesses with completed scoring status. Treat any single industry split as directional rather than definitive given varying sample sizes by industry. Full methodology at How We Score.
So what should an operator actually change
Nothing clever. Refusing to be clever is the actual move here.
Start with two questions that have nothing to do with software at all. Does the business know what it earns? Does it know what it spends, and where? Everything downstream traces back to those two answers, and an operator who can't state them plainly has no yardstick to judge an automated change against, regardless of how capable the tool is.
Then locate the single constraint doing the most damage. Sometimes it's a mobile experience so rough that fixing it alone lifts conversion 20% overnight. Sometimes it's a channel that would change nothing if it disappeared tomorrow, which is exactly the test worth running. Switch it off. If revenue drops, that's real signal. If nothing shifts, the payment was buying nothing at all.
Only once both are answered does automation get pointed at anything. At that stage it becomes a throttle rather than a wager, scaling up when the numbers justify it and pulling back when they don't. Direction stays a human call. The system just executes it faster once the call has been made.
Numbers first, the constraint second, the machine last. Reverse that order and the result is an expensive way to lose while sounding certain about it.
Where this heads over the next couple of years
The next two years won't belong to whoever licenses the flashiest system, since most rivals will be running something close to identical. They'll belong to whoever put in the unglamorous work of understanding their own operation before scaling anything through it.
The operators wiring automation into a real rhythm aren't sharper than the field. They simply refused to skip the commercial groundwork most competitors are still skipping. In Australia that bar sits lower still, given close to half the local businesses using these tools aren't confirming whether the tools are earning their keep.
That's not a caution. It's an opening, sitting there for whoever picks it up. While most of the market frets over the software itself, the real edge is the least fashionable habit in marketing: know the numbers, locate the constraint, then let automation run.
Want to see where your own numbers actually sit before pointing anything automated at them. That's exactly what New Rebellion's Hub is built to show. Six dimensions, benchmarked against real Australian businesses, no guesswork required.
Frequently asked questions
Is AI actually the same across every marketing tool now?
The underlying models and capabilities have converged fast enough that most automated marketing tools produce comparable output quality. The real differentiator has shifted from which platform a business licenses to what it's pointed at and how well the result can be checked.
Why do so many Australian businesses skip tracking AI's impact?
MYOB's research found 74% of businesses that don't track AI's impact call tracking "unnecessary." In practice this usually reflects a missing baseline to compare against, not that tracking itself has no value.
What should a business fix before spending more on AI tools?
Establish two numbers first: what the business earns and what it spends, broken down by channel. Then identify the single constraint holding back growth. Scaling automated tools only makes sense once both of those are clear.
How does New Rebellion measure Data & Tracking maturity?
Data & Tracking is one of six weighted dimensions in New Rebellion's benchmark scoring model, alongside Digital Maturity, Acquisition Performance, Conversion Efficiency, Retention & Loyalty and Brand & Positioning. It draws on signals including analytics configuration, tag management, consent setup and attribution sophistication. Full methodology at How We Score.