Here's a pair of numbers worth sitting with. According to MYOB's research on Australian SMEs, 82% of businesses using AI say it's having a positive impact. And 46% of them don't measure that impact at all.
Read those together and the story changes. Nearly half the firms reporting that AI "works" are describing a feeling, not a result.
What adoption actually looks like right now
Across December 2025 to February 2026, the National AI Centre's SME tracking found 43% of Australian SMEs reported some level of AI adoption. Not 90%. Not "everyone but you." Forty-three percent — and the number has been drifting sideways, not rocketing up.
So if you run an accounting practice, a law firm, or an advice business and you feel behind on AI, the data says otherwise. You're standing in the middle of the pack, surrounded by firms in exactly the same position: tools purchased, someone drafting emails faster, and no clear line from any of it to revenue, hours, or client experience.
That's what I call the faster typewriter problem. The work got slightly quicker. The business stayed the same.
Why measurement is the whole game
The professional services data makes the gap sharper. Thomson Reuters' 2026 research across legal, accounting and advisory found only 18% of organisations track return on investment for their AI tools. In the legal sector specifically, Clio found around three-quarters of small firms now use AI — but only about 31% report any revenue increase from it.
And here's the finding that should change what you do next: firms with a visible AI strategy are almost four times more likely to see ROI than firms adopting ad hoc. Not better tools — a strategy. Knowing which processes AI should touch, what success looks like in dollars or hours, and what you're deliberately skipping.
The skeptics are asking the right question
The National AI Centre also found that among businesses not adopting AI, the biggest barrier isn't cost or complexity — it's trust. Around 65% cite distrust of AI decision-making or a preference to keep humans in control.
Vendors treat that skepticism as a problem to overcome. I think it's the correct starting position. The firms that get value from AI aren't the ones that trusted it fastest. They're the ones that asked "where does our business actually leak time and money?" before asking "which AI tool should we buy?" — and were willing to hear "AI won't fix that one" as an answer.
A 15-minute test for your own firm
Ask three questions:
1. Name the AI tools your firm pays for. (Easy.)
2. Name the business process each one was bought to improve. (Harder.)
3. Show the before-and-after number for each — hours, dollars, or turnaround. (This is where most firms go quiet.)
If you can't get past question three, you haven't failed at AI. You've just been sold adoption without measurement — like most of the market. The difference between the 82% who feel AI helps and the firms that can prove it is not more software. It's picking two or three places where the numbers say AI will pay, and skipping the rest.
That's the entire premise of how we work.