Most small businesses already use AI, but few have built it into their workflows. Why Australian SMEs move faster than enterprise, and where to start.
Over the weekend the Guardian ran a piece by Gene Marks on the split between how big companies and small ones are using AI. Enterprise is actively cutting roles with it. Small businesses are using the same technology to hold on to the people they already have and help them get more done in a day. We work with Australian small and medium businesses, so that’s the seat this is written from, and the short answer is that from what we see day to day, Marks has it right. Your advantage in AI isn’t budget, and it isn’t talent. It’s that you can make decisions. Right now that’s worth more than any tool on the market.
Marks describes a window and door retailer who spent about US$10,000 on an application that listens to showroom conversations and drafts the quote for the salesperson to check and send. “It allows my salespeople to talk to more customers and spend less time doing paperwork,” the owner told him. “And it cuts down on errors.” Another business he knows connected Claude to a folder of specs, manuals, and technical sheets for the equipment it sells, so the support team gets answers in seconds instead of hunting through a printed guide.
Why small businesses get further, faster
Neither of those is world beating technology, but they are ordinary problems solved by someone who was able to say yes. We work closely with a Gold Coast building products supplier with 22 staff. Quotes were taking too long because each one needs a look through pricing sheets, and two people in support spend their days finding answers in supplier PDFs. Nobody’s drowning exactly, but there’s no room for error either, and the errors that were showing up started to cost money. When Lachlan (owner) decided to fix quoting, we ensured that the people who knew the quoting process were in the room, and often it’s not just the owner. These people know what the process is meant to achieve, where it breaks, and what fixing it is worth, so the conversation is short because it is a small, concentrated group, and the work can begin in earnest.
Put the same idea inside a large enterprise and it collects layers instead. Procurement, legal, IT, risk, and a few business units all hold a view. Everyone is partly a decision maker and nobody is fully one, so the meetings multiply, the risk register grows, and a year later there’s a pilot. The article makes a related point about why the job cuts land at head office and not on Main Street. Big companies carry slack in departments like PR, marketing, IT, and customer service, so automation can quickly reduce headcount. In a 20-person business there’s rarely any slack, everyone is already pulling their weight, and that’s exactly why the owner wants AI to help staff do more rather than to replace them. The survey numbers tell the same story. In the Goldman Sachs 10,000 Small Businesses survey, 76% of small businesses already use AI tools in daily operations, and 93% report a measurable gain in efficiency, productivity, or customer service. 87% of owners see AI as a way to help the staff they have, not a way to shed them.
The gap between using AI and running on it
Then there’s the figure that matters most. Only 14% have fully built AI into their core, end to end operations. Read those numbers together and the picture is clear. Nearly everyone has tried AI, nearly everyone found something useful, and almost nobody has changed how the work actually runs. Our Gold Coast supplier fit that description exactly before we started working with them. For over a year, AI there meant a few staff rewriting emails and drafting policies with a chat interface. Useful, and completely beside the business rather than inside it. Teh Guardian sees the shift starting now, from asking questions and reviewing contracts last year to real applications that show a return.
Access and cost aren’t the hurdle any more. Free tiers, low code integration, and cheap cloud infrastructure took care of that, which is why small businesses have closed the technology gap with big corporates in months, where cloud computing and e-commerce took them years. What’s left is training, data hygiene, and implementation guidance.
Two shifts change what “built in” looks like. The first is Agentic AI, meaning software that carries out a multi-step workflow across your systems rather than answering one question, so it pulls the client’s data, drafts the proposal, updates the CRM record, and sends the invoice. The second is a move away from general purpose chat interfaces toward tools built for a single industry, because broad models struggle with industry compliance and the domain rules a legal, trade, logistics, or accounting business can’t be loose about.
However there are major caveats that apply to both shifts. An agent doesn’t fix bad data, it acts on it faster and in more places, so records that are half wrong become wrong across four systems instead of one. Trust is the other, and it’s earned slowly for good reason. Marks’s clients worry about breaches, about models trained on private information, and about whether their pricing and costs are exposed by sitting on someone else’s platform. In Australia that worry is also a legal duty, because the Privacy Act 1988 and the Australian Privacy Principles keep you accountable for personal information you hold, whichever model touched it. So a person approves what goes out the door, and some work stays with people: an unhappy customer, or a call that needs someone’s experience on the job.
Where to start
Lachlan, our building supplier started with one job. Something that was contained and the data was already sitting in a folder. Four and a half weeks later that workflow was mapped, the agents could pre-fill all the data, and then the quoting work followed, with AI drafting each quote from the conversation and a person checking the price before it went out. What made it work wasn’t the tool. It was knowing which job to point it at first, having the pricing data clean enough to trust, and being honest about the two ideas that got parked because they were never going to work in a business that size.
That last part is where most owners are stuck, and it’s why 76% usage sits next to 14% integration. Not knowing where to start is a reasonable place to be. It’s also the whole value of a technology partner who works your business out with you, says which jobs are worth doing and which aren’t, and has already watched the pitfalls play out somewhere else. You get the benefit of that experience without buying it one failed pilot at a time.
What this means for you
The return on AI for small and medium businesses isn’t really in question any more. What’s open is whether you act while acting is quick and cheap, because your competitors are working through the same list right now with the same short chain of command. If you know AI should be doing more in your business but aren’t sure what to point it at, take our AI Roadmap Interview. You’ll talk through your team’s goals and where the time goes, then get a plan for the first job worth doing. You can also read how we work as your technology partner.
