Home Insights The hardest part of AI is choosing where to point it first, and most businesses have nobody whose job that is
AI Strategy

The hardest part of AI is choosing where to point it first, and most businesses have nobody whose job that is

By QuantalAI Solutions Team · 11/08/2026

Most SMBs can't afford a full-time AI lead, so the first step never gets chosen. A technology partner brings that judgement at a fixed monthly cost.

Most owners we speak to have already decided AI can help them. What they can’t decide is where to start. There’s no shortage of tools, no shortage of people telling them the tools are urgent, and especially no shortage of people telling them that their tool is the best. What’s missing is someone whose actual job is to look at how the business runs and say, let’s start here.

That gap shows up in the data. Goldman Sachs surveyed 1,256 US small business owners in early 2026 and found 76% already using AI, while only 14% had it built into core operations. Close to half named the same two blockers, a lack of technical expertise (49%) and difficulty choosing the right tools (48%). Australia looks similar. A January 2026 AI Lab Australia report, which models 500 Australian enterprises using Deloitte, Tech Council, ABS, and National AI Centre data, puts regular AI use at 64% of SMBs while only about 5% of adopters are what it calls fully enabled.

So the tools have arrived but the judgement about where to aim them hasn’t. That judgement is the part that decides whether any of it produces a return on investment.

Buying everyone a subscription isn’t a strategy

Lets use a Perth equipment hire business with 40 staff as an example. The owner did what most owners do first and bought the whole team AI subscriptions for use in their day to day work. The team used them to tidy up emails and draft ads, which saved maybe 2 - 3 hours a week each. That matches what the AI Lab report describes for basic users, an incremental gain of one to two hours a week and not much else. Nothing about the business got faster.

The reason was that the hours that were saved were never the constraint. Quotes still took three days to go out, because one person had to pull availability, past hire history, and pricing out of three systems that don’t talk to each other. Faster emails just meant enquiries reached that queue sooner. This is the most common way AI work goes wrong. AI speeds up whatever you point it at, so pointing it at the wrong step delivers work to the bottleneck faster and makes the bottleneck worse. The business didn’t have a tool problem, they had a targeting problem, and nobody in the building was experienced enough to solve it.

The person who could tell you where to start is a serious hire

Working out the right first step takes someone who has done it before inside a business, not someone who has played around with it on a weekend project. It needs a proper hire, and in the current market it costs somewhere around $250,000 AUD a year. Most SMBs can’t carry that, and most don’t have enough AI work to keep that person busy full time even if they could. Your own tech team isn’t the answer either, through no fault of their own. They’re good at running what you already have. They aren’t putting new AI systems in every week, so they aren’t current on what’s changed, because they’ve been busy doing the job you hired them for. Learning on your money is slower and dearer than it looks.

The AI Lab report puts numbers around this. More than half of the SMB workforce has only basic or novice AI literacy, and just 10% have advanced skills. Small businesses can’t outbid banks and tech firms for the people who do. What the report says those businesses need is an AI Translator, someone who understands the actual work and can see where AI genuinely fits. Very few SMBs have one on staff, and that’s the whole problem in a sentence.

What a technology partner does instead

This is why we built our Technology Partner service. You get an allocated block of expert hours each month, and every AI decision and every piece of AI build work sits with that one role. The ongoing running costs of what we put in are rolled into the same fixed monthly price, so you know the number before the month starts and there’s no meter running in the background.

The work starts with your process rather than with a tool. For the Perth business, the first piece of work wasn’t a chat interface. It was joining those three systems so a quote could be built in one pass, with a person still checking the price before it goes out. Quote turnaround moved from three days to same day, and the 12 hours a week that used to go into chasing figures went back into taking bookings. On numbers like those, the monthly fee is covered inside the first few months, which is why we treat this as an efficiency line rather than a technology cost. The role isn’t there to replace anyone. It’s there to make the team you already have get more done.

Two honest limits. A fractional partner isn’t in your building or business every day, so this works when you’re clear about what you want and it drifts when you’re not. And accountability stays with you. The National AI Centre’s Guidance for AI Adoption, released in October 2025, is plain about it, which is that someone is responsible for the AI’s output, customers know when they’re dealing with AI, and a person can review the decisions that matter. We build that review in rather than leaving it for you to bolt on later.

If you’re weighing up agencies, more of them are moving to this fixed price, outcome-driven shape, which is a good thing. The one check worth making is whether they build the technology or only talk about it. Ask what they’ve put into production recently and what changed for the client.

What this means for you

You probably don’t need to hire an AI champion. You need the judgement one brings, for the few days a month you’d actually use it, and you need it aimed at the right problem the first time. Find where work banks up in your business before you buy anything, because a subscription for everyone is a cost, not a plan. That’s the piece a technology partner is there to own.

If you’re not sure where your bottleneck sits, 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 step and what it’s worth doing.

Frequently asked questions

What is a technology partner in AI?
A technology partner is an outside expert who owns your AI decisions and AI build work for an agreed block of hours each month, at a fixed price. Instead of hiring a full-time AI lead, you pay a fraction of that cost and get the same judgement about what to do first. Ongoing running costs for the AI you put in can sit inside the same monthly figure, so the number is known before the month starts.
Can't our existing IT team handle AI?
They can run what you already have, and that's a different skill. Installing and tuning AI systems is something you get good at by doing it every week across many businesses, and your team has been busy doing the job you hired them for. Most SMB tech teams end up learning on your money, which is slower and more expensive than it looks.
How do we choose the first AI project?
Start with where work banks up, not with a tool. Map the steps in a process, find the point everything queues behind, and check that speeding it up would actually change the output. AI Lab Australia's 2026 report gives the same advice, which is to start with your most expensive repetitive problem rather than asking how to use AI.
Who is accountable for what AI does in our business?
You are. The National AI Centre's Guidance for AI Adoption, released in October 2025, sets three plain expectations, which are that someone is accountable for the AI's output, customers know when they're dealing with AI, and a person can review decisions that matter. A good partner builds that review into the work rather than leaving it to you afterwards.
How do we know an AI agency is any good?
Ask what they've built and who is using it right now. More agencies are moving to fixed price, outcome-driven arrangements, which is a good sign, but only if the people involved have shipped working systems rather than talked about them. Ask for the last three things they put into production and what changed for the client as a result.