How to use the RBA’s predictions to get ahead of the curve with AI

Lee Robson, Co-Founder of The Agitators, explores the Reserve Bank of Australia’s warning on slow economy-wide AI productivity gains and shows how small businesses can pull ahead by redesigning…

Lee Robson
Expert · 16 Nov 2026 · 2 min read
Above How to use the RBA’s predictions to get ahead of the curve with AI. Dynamic Business

The Reserve Bank says it expects artificial intelligence to take much longer to lift productivity than many following the pace of progress might expect. In a speech on 9 November1, Assistant Governor (Economic) Sarah Hunter said there's limited evidence so far that AI has raised productivity across the economy, comparing it to the arrival of electricity, whose biggest gains took decades to arrive.

 That’s one – though not the only – way of reading the economy as a whole. But for an individual business, perhaps the most valuable thing to take away from this comparison is the idea that the most impactful gains are likely to go to firms that redesign how their work is done around this new technology, and that SMEs are capable of redesigning their workflow faster than almost anyone.

 What the electricity story shows

When electric motors arrived in American factories, most owners pulled out the steam engine, put a large electric motor in its place and changed nothing else. The overhead shafts and leather belts that carried power to every machine stayed where they were. At the turn of the twentieth century, electric motors supplied less than 5% of US factory mechanical drive, and it took decades more before half was electrified.

The real payoff came when engineers gave each machine its own small motor. The shafts came down and factories were rebuilt around the flow of work.

The RBA expects AI to follow the same pattern, something that economists call a J-curve: a flat stretch while firms reorganise, then gains once the technology is built into how the work gets done.

For many businesses, AI mostly looks like that motor bolted in where the steam engine used to be. Someone drafts emails with ChatGPT. Someone else summarises meeting notes. The work arrives, moves around and gets approved the same way it did in 2022, with one or two steps running faster. The minutes saved tend to disappear into the next email.

The wait can be shorter

There are signs it already is in the US. In February, Stanford economist Erik Brynjolfsson argued in the Financial Times that US productivity grew around 2.7% in 2025 , close to double its average over the previous decade, and that the economy is entering a "harvest phase".

 Small businesses have an advantage here. A factory owner in 1910 had to rebuild a plant to switch to individual motors. A 15-person business can redraw a workflow on a whiteboard on a Friday afternoon, with no steering committee and no change program.

Pick one piece of work that comes round every week or month: a client report, a proposal, an invoice run. Map every step it goes through and who touches it.

Then ask the factory engineers a question: if you designed this work today, knowing what the tools can do, which steps would still exist, and who would do them? Some steps will vanish. Others, such as checking the output and deciding what it means, will matter more.

Make one or two changes then watch where the freed-up time goes. If it goes straight back into email, you've sped up a step. If it goes into work you couldn't get to before, you've started rebuilding the factory.

What to expect on the way

Rebuilding costs something. Time spent redesigning a process is time not spent billing, and the first version usually works worse than the old one. That's the dip the RBA describes, and nobody skips it entirely.

Hiring is where the change may show first. Most firms in the RBA's business liaison program say AI hasn't changed their headcount, and they don't expect it to over the next year. Hunter does point to early, modest signs of weaker hiring for younger workers in highly AI-exposed occupations, while stressing it's too early to say AI is the cause. If you're hiring fewer juniors because the entry-level work has shrunk, you're already reorganising, and you'll need to decide where your next senior people will learn the job.

Jobs and Skills Australia estimates that around 4% of Australian workers are highly exposed to automation by AI, and 31% to AI augmenting their work. Those figures measure what AI could do to a job. Which way a particular job goes depends a good deal on how its employer redesigns it.

 Hunter wraps up by noting that the largest gains from transformative technologies often take longer to arrive than expected. Across the whole economy, she's probably right. For your business, the clock starts when you pick the first workflow to rebuild. In the 1910s, the factories that got to the real gains first were the ones brave enough to rebuild the entire machine, not just replace a few parts

Lee Robson
From the floor
Closer to this story than we are?
If you're building in this space — or watching it reshape your market — pitch us. We edit it; you get the byline.
More from the desk

Keep reading.

News ◆

Five business groups say this tax bill could hit 350,000 small businesses

Five business groups have written to the Senate warning that a proposed tax change could hurt hundreds of thousands of small family businesses and their families. The change would introduce a 30 pe…

Yajush Gupta · 2 min
News ◆

ACS renews call for sovereign IT review after Microsoft outage

Microsoft says the issue is fixed, but ACS says Australian small businesses deserve more than a status page during an outage like this.

Yajush Gupta · 2 min
News ◆

A retailer’s checklist for a scam-safe, compliant Black Friday sale

Australians lost $2.18 billion to scams in 2025. Here’s the official Stop, Check, Protect framework before you shop this Black Friday.

Yajush Gupta · 2 min
0 people like this

Comments

Loading comments…