AI Personalised Pricing: The Australian Consumer Law Line

Two customers, same job, different price — decided not by you but by a tool reading their postcode, their phone, their history with you. On 20 July 2026 the federal government put that practice on notice, naming "retail surveillance pricing" among five AI consumer-safety priorities it will examine under Australian Consumer Law. The technology that does this is cheap now, and it is already sitting inside the quoting and pricing software vendors are selling to trades and professional-services firms.

The stakes aren't abstract. Get personalised pricing wrong — charge differently in a way that misleads a customer about the price or why it changed — and you're in misleading-conduct territory under the ACL, where the maximum penalty for a company is the greater of $100 million, three times the benefit, or 30% of annual turnover. That last limb is the one built for your size.

The 20 July announcement, from Assistant Minister Andrew Leigh's office, set out five AI consumer-safety priorities: a duty of care, privacy reform, AI safety in the workplace, consumer protections, and a framework for automated decision-making. The pricing line sits under consumer protections: the government will "examine options in Australian consumer law to address consumer risks, such as retail surveillance pricing and agentic commerce."

It doesn't land in a vacuum. The ACCC's 2026–27 enforcement priorities already target "manipulative and false practices" in digital markets and pricing that limits genuine consumer choice, and the Unfair Trading Practices Bill 2026, introduced to Parliament on 1 April, would prohibit manipulative practices outright. Two sides of the same shift are now named: surveillance pricing is the selling side — the price you show — and agentic commerce is the buying side, where a customer's AI shops on their behalf. This note is about the first.

Want to know what this means for your business? We'll map it in one call.

Start with what's fine, because most of it is. The ACCC's position is that dynamic or personalised pricing is not illegal in itself — businesses "must be clear about the price consumers will pay and must not make false or misleading pricing claims." Charging more for an after-hours callout, a complex job, or a rush turnaround is normal, defensible pricing. The reason sits in the job.

The line is crossed when the reason sits in the customer. An AI tool that quietly shows a higher price because someone's postcode, device, or spend history marks them as willing to pay more — and misrepresents or hides why — is exactly the conduct Section 18 of the ACL prohibits. Intent is irrelevant; you don't have to mean to mislead, you only have to mislead. There's a second trap the overseas evidence keeps surfacing: proxy discrimination. A model that never sees a protected attribute can still use postcode or device as a stand-in for one, producing outcomes you'd never sign off on if they were written down. A US Consumer Reports investigation, as reported by Louder, found 74% of grocery items carried more than one price, with some shoppers shown prices up to 23% higher — a scale of variation few customers would guess was happening.

This is a compliance exposure that maps straight onto the rest of the rules that already create AI liability in Australia: the law is technology-neutral, so the AI vendor's tool doesn't inherit the risk — you do. If you're weighing an AI pricing or quoting tool and can't yet explain how it decides a number, that's the conversation to have first.

You don't need to avoid AI pricing tools. You need to be able to explain, in one sentence, why two customers see different numbers — and the reason has to be about the job, not about who the customer is. Write your pricing rules down. If a rule is one you'd be uncomfortable stating to a customer's face — "you're in a wealthy suburb," "your phone says you'll pay it" — that's the part carrying the risk, and it doesn't matter that a model, not a person, applied it. Keep a human able to justify any quote the system produces. The businesses that treat AI pricing as a transparency question, not just a margin question, are the ones that won't be rewriting their approach after a complaint. Book a call and we'll pressure-test how your pricing tool decides.

Key takeaways

Common questions

Related notes

Weighing an AI pricing or quoting tool?

AI should sharpen your pricing, not expose you to a misleading-conduct claim. A short conversation can pinpoint whether your tool's pricing logic is explainable enough to stand up. Book a call.