Every survey out this year agrees Australian small businesses are using AI in some form — the range runs from 43% to 66% depending on how loosely "using" is defined. The story worth telling is what happens after the sign-up: most of those businesses stall at the exact same rung, and stalling there is expensive.
Deloitte Access Economics surveyed more than 1,000 Australian SMBs against a purpose-built AI Maturity Index and published the results in November 2025, commissioned by Amazon Australia. Two-thirds of respondents were using AI. But when Deloitte scored how well that use was actually integrated into the business — not just switched on, but embedded in how work gets done — only 5% qualified as "fully enabled." Over 40% sat at the most basic level: a tool here, a login there, no real change to the workflow underneath.
Deloitte Access Economics, The AI Edge for Small Business, commissioned by Amazon Australia, Nov 2025 (n=1,000+ Australian SMBs)
The gap, drawn to scale
Put those two numbers next to each other and the shape of the problem is obvious. It isn't an adoption curve — it's a wall. Most businesses get over the first hurdle (signing up, trying a tool) and then stop, because nobody scoped what "done properly" looks like for their specific workflow.
That distinction has a dollar figure attached. Deloitte modelled what happens to profitability as a business climbs each rung of the index, and the curve is steep enough that it changes the argument for doing this properly the first time.
Deloitte's modelling found that if just one in ten SMBs sitting at the basic or intermediate level moved up a single rung, it would add $44 billion to Australia's GDP — 1.3% of the whole economy, from small, unglamorous businesses doing the same work slightly better organised. That's not a moonshot number. It's what happens when a workflow that already exists gets automated properly instead of half-heartedly.
Businesses across every industry cited the same barrier first: they don't know what AI can actually be used for inside their own business.
— Deloitte Access Economics, The AI Edge for Small Business, Nov 2025
A year on, the barrier has shifted — fewer businesses say they don't know AI exists. More say they don't know how to apply it properly, or trust that they're doing it right.
— reframed from MYOB Business Monitor 2026 and National AI Centre commentary, 2026
Adoption is also softer than the headlines suggest
It's worth being honest about the other side of this too, because the "AI adoption is exploding" framing oversells the trend. The Australian Government's National AI Centre runs a monthly tracking survey — the SME AI Pulse, fielded by Fifth Quadrant across a minimum of 400 SME owners and decision-makers each month — and its most recent quarterly read (December 2025–February 2026) put adoption at 43%, actually down from 45% the quarter before.
NAB's own SME research this year lines up with a similar middle ground: 42% of small and medium businesses currently using AI, 44% not using it, and 14% planning to start. The Australian Industry Group's broader survey across all business sizes put the figure at 52%. None of these move in a straight line upward — adoption is plateauing in places, and the gap between industries is wide.
| Sector | Using AI | Note |
|---|---|---|
| Property services | ~70% | Highest adoption of any sector surveyed |
| Finance & insurance | 64% | Claims and document-heavy workflows lead the way |
| Business services | 61% | Admin and reporting the most common entry point |
| Hospitality | n/a | Lowest current use, but the highest share (~2 in 5) planning to start |
| Transport | 21% | Lowest adoption; 58% report no plans to adopt at all |
Read sector by sector, the pattern is less "AI revolution" and more "the businesses with the most repetitive, document-heavy work adopted first, and everyone else is watching to see if it's worth the disruption." That's a completely rational way to behave — and it's also exactly why the businesses that move now, with a properly scoped rollout instead of a trial-and-error one, end up with a multi-year head start over competitors who wait for certainty that never quite arrives.
A signal worth watching: hiring
CSIRO's research adds a data point that doesn't show up in the adoption surveys: businesses that have already adopted AI are posting 36% more job advertisements than those that haven't. That's not AI replacing headcount — in the SMB data, it correlates with growth. Businesses use the time AI frees up to go after work they couldn't previously service, and then hire to cover it.
What this means if you're actually small
One thing worth being straight about: every stat above is an "SME" number, and "SME" in Australia stretches from a two-person plumbing business to a 150-person engineering firm. Lump them together and adoption looks like a healthy, broad-based trend. Split them apart and the picture for a genuinely small business — a tradie, a two-partner accounting firm, a physio clinic — looks a lot steeper.
| Business size | Using AI |
|---|---|
| Large (200+ employees) | 80%+ |
| Mid-market (50–199 employees) | ~60% |
| Small (5–19 employees) | ~38% |
| Micro (1–4 employees) | ~33% |
That's close to a straight line down as the headcount drops — and it's not really an appetite problem. It's a capacity problem. The "AI adoption" headline is being carried by businesses with someone whose job it partly is to trial new tools: an ops manager, an IT budget, spare hours to experiment. A sole-trader electrician or a three-chair physio practice doesn't have that person. They have a full diary and an inbox that doesn't stop.
That makes the earlier point about scoping matter more for a small business, not less. A business this size genuinely can't afford to sit at Deloitte's basic-use rung, tinkering with a general-purpose tool in spare minutes that don't exist. The small businesses actually seeing a return tend to skip the tinkering stage altogether — they get one narrow, costly problem solved properly and once, by someone else, rather than trying to become their own AI department on top of running the trade or the practice.
What actually separates the 5% from the 40%
Nothing in this data suggests the businesses at the "fully enabled" end are smarter, bigger, or better resourced. Deloitte's report and the sector breakdowns above point at something more specific: the businesses that get real value scoped the automation around one concrete, recurring problem in their own workflow — a missed-call pattern, a quote-follow-up that always slips, an after-hours enquiry nobody's awake to answer — instead of adopting a general-purpose tool and hoping the value showed up on its own.
That's the practical difference between "using AI" and being in the 5%. A generic chatbot subscription is basic-level use no matter how much it costs. A system built to catch one specific, costly gap in how the business actually runs — with someone accountable for how it behaves — is what Deloitte's index is actually measuring when it says "enabled."
Which rung is your business actually on?
We built momentuum to close exactly this gap for Australian SMBs — not by selling a tool, but by scoping the one workflow costing you the most time or leads, and building the automation around it properly. A free 20-minute call, no pressure, no obligation.
Sources
- Deloitte Access Economics, The AI Edge for Small Business: Increased SMB AI Adoption Can Add $44 Billion to Australia's Economy, commissioned by Amazon Australia, November 2025 — deloitte.com/au
- National AI Centre (Australian Government), AI Adoption Insights: December 2025 – February 2026, SME AI Pulse survey fielded by Fifth Quadrant — ai.gov.au
- NAB, SME AI adoption research, 2026, including sector breakdowns for property services, finance & insurance, business services, transport and hospitality.
- Australian Industry Group (Ai Group), business AI adoption survey, 2026.
- CSIRO, research on hiring activity among AI-adopting Australian firms, 2026.
- NAB Economics, AI adoption by business size (large, mid-market, small and micro employers), 2026.
- MYOB Business Monitor, 2026, non-adopter barrier breakdown (cost, time, understanding); National AI Centre commentary on shifting awareness vs. application barriers, 2026.