Soon an AI will do your customers' shopping and your admin. Here's how not to be left behind

“If you do not get across it and get ahead of it, you’re going to be left behind.” That is how AMP Bank’s Shehan Rajakumar summed up agentic commerce at a recent panel, reported by SmartCompany. It is a blunt line, and the sort that usually deserves a raised eyebrow. This time it is worth taking seriously, as long as you also take the panic out of it.

Agentic commerce means AI agents that do not just answer questions but act on them: search, compare, decide, and complete a purchase on a customer’s behalf, within limits the customer sets. And it has two sides that both reach a Melbourne service business, even one that has never sold a thing online.

Side one: your customers’ agents come looking

“Agentic commerce” sounds like a retail problem, something for shops with a checkout. It is not. Picture someone telling their assistant, “find me a physio in Fitzroy with a Saturday appointment and book it.” The agent does not scroll Google. It shortlists the businesses whose information it can actually read, then acts.

That is the whole game. If your services, prices, availability, location and hours are clear and machine-readable, you make the shortlist. If they are buried in a slow site or a PDF or only live in your head, you do not exist to the agent. We covered the retail version of this when AI shopping arrived in Australia, and the groundwork in getting found in AI search. The encouraging part, as the panel noted, is that this can level the field: a small clinic with clean, readable data can beat a big one with a messy website.

Side two: your own agent goes to work

Here is the side that pays off sooner, and the one most Melbourne businesses are sleeping on. The nearer-term win is not being bought by an agent. It is putting one to work for you.

MYOB’s Sally Davies pointed out that 69% of small businesses worry about cash flow every single day. An agent that chases overdue invoices, drafts the polite follow-up, reconciles payments, or triages the enquiry inbox is time and money handed back to you now, not in some agentic future. This is the same idea we keep coming back to in putting AI to work inside your systems, not just a chat window.

The best advice from the panel was refreshingly unglamorous. Do not automate everything at once. In Rajakumar’s words: “Pick one thing in your business that you want to try and get an agent to look after. Train your agent over time to do that job really well.” One task, done properly, beats a grand plan that never ships.

The honest caveat: the rules are still being written

None of this is fully baked. The panel was clear that the trust layer, who is liable, how approvals work, what an agent is and is not allowed to do, is still being built by banks, regulators and payment providers. So start where the stakes are low and a human still signs off. Let an agent draft your invoice reminders, not empty your bank account. Widen the leash as the trust, and the standards, catch up.

Where a Melbourne business should actually start

1. Get machine-readable. Make sure an agent can read your services, prices, availability, hours and location. Clean structured data and consistent information across your site and listings, the same groundwork that gets you found in AI search.
2. Pick one internal job. The most annoying repetitive task on your week, quote follow-ups, invoice chasing, enquiry triage, is your first candidate. Just one.
3. Keep a human in the loop. Let the agent prepare and draft while you approve. Trust is earned one task at a time, for software as much as for staff.
4. Measure it. Hours saved, invoices paid faster, enquiries answered sooner. If you cannot see the gain, change the task.

The func.digital take

“You’re going to be left behind” is a scary phrase, and a fair warning. But the response is not to panic-buy AI. It is to do two calm things: make your business readable to the agents on the outside, and put one agent to work on the inside. The winners over the next couple of years will not be the businesses that bought the most technology. They will be the ones that picked one real problem and solved it properly, then picked the next.

If you would like help choosing that first job, or checking whether an AI agent can even find your business yet, that is exactly what a free digital systems audit is for. We will find the one task worth automating first and tell you where your data is letting you down. Get in touch, and get ahead of it rather than left behind by it.