Safety & Security

Adactin maps six fraud defences for Australian businesses facing AI scams

Adactin has listed six fraud defence priorities for Australian firms, warning that AI deepfakes, real-time payment scams and expanded AML rules are outpacing manual controls.

Adactin urges Australian firms to step up fraud defences
Adactin urges Australian firms to step up fraud defencesPexels / pixabay
By Sophie Lindqvist6 min read

Why it matters

  • Adactin published six fraud defence investment priorities for Australian businesses on 16 November 2026.
  • Head of AI & Data Engineering Vishal Bhanage said AI should support better decisions, not replace accountability.
  • Australia's expanded AML rules now cover services from lawyers, accountants, real estate professionals and conveyancers.
  • Adactin flagged AI-generated voices, deepfakes, account takeovers and real-time payment scams as leading fraud vectors.

Adactin has named six investment areas it says Australian businesses must fund to cut fraud, as AI-enabled scams multiply and compliance obligations tighten across newly regulated sectors. The technology and consulting firm published its recommendations on 16 November 2026, targeting a shift in criminal tactics away from stolen card details and toward impersonation, social engineering, and weaknesses in everyday business systems.

The six priorities are concrete: combining business data, strengthening identity checks, monitoring payments more closely, tightening internal processes, setting clear rules for AI, and working with industry and government. Adactin warned that AI-generated voices and deepfakes, real-time payment scams, fake and stolen identities, account takeovers, and attacks that target AI tools themselves are likely to shape fraudulent transactions in the period ahead.

Vishal Bhanage, Head of AI & Data Engineering at Adactin, said businesses would need to move beyond traditional fraud prevention methods as criminal tactics evolve. He pointed to faster payments, more convincing impersonation attempts, and the growing use of automation in business processes as the forces driving that change.

The stakes are straightforward. Fraud has become cheaper to run and harder to detect, and the businesses least prepared are often the ones still approving payments by hand.

Why are fraud attempts getting cheaper and harder to catch?

Adactin's core argument is about scale. AI can now help criminals produce convincing messages at scale, which lowers the cost of each attempt and raises the volume a business must filter. That makes companies more attractive targets, especially where payment approvals and supplier checks still rely heavily on manual processes.

The firm flagged five fraud types to watch:

  • Business email and invoice fraud
  • Impersonation scams
  • Identity fraud
  • Refund and payment fraud
  • Investment and cryptocurrency scams

Mid-sized businesses and organisations with complex supply chains face particular exposure, Adactin said. The danger concentrates where staff handle supplier payment changes or urgent transfer requests — precisely the moments impersonation scams are built to exploit.

Australia's real-time payments infrastructure adds another layer of pressure. Money can move before suspicious activity is identified, so controls that react after the fact are largely useless. In that environment, Adactin concluded, businesses may need stronger verification steps before authorising transfers or updating supplier bank details.

The takeaway: the window for catching a fraudulent payment has effectively collapsed, so the checks have to happen before the money moves.

Is the real problem technology or coordination?

For many organisations, Adactin argued, the challenge is not simply a lack of technology. It is weak coordination across systems, information, and staff. Data on customers, suppliers, payments, and suspicious activity often sits in separate platforms, which makes it harder to spot unusual patterns or connect related incidents.

This is why Adactin's first investment priority is combining business data rather than buying new tools wholesale. The firm's recommendations place particular weight on basic controls over system replacement: improve how information is collected, shared, and used across payment systems, customer records, and supplier databases, so suspicious behaviour surfaces sooner.

The specific controls Adactin recommended include:

  • Multi-factor authentication
  • Stronger customer and supplier verification
  • Extra checks when bank details change
  • Call-back verification for payment amendments
  • Dual approval for high-value transactions
  • Staff training to recognise impersonation and deepfake scams

None of these require exotic technology. Most require discipline and someone accountable for making the pieces talk to each other.

What do privacy and AML rules demand?

Governance pressure runs alongside the fraud problem. Adactin pointed to privacy, governance, and anti-money laundering obligations that constrain how businesses can respond. Organisations need to understand what information they can share, with whom, and under what circumstances, while still meeting privacy requirements.

Automated decision-making carries its own disclosure duty. Some organisations will need to state in their privacy policies how automated systems make, or help make, decisions that could significantly affect individuals' rights or interests. That requirement lands directly on companies deploying AI in fraud detection or customer decisions.

Expanded anti-money laundering and counter-terrorism financing rules add a further pressure point. Adactin identified sectors newly brought into Australia's regulatory framework, including some services provided by lawyers, accountants, real estate professionals, and conveyancers. Those businesses now need customer checks, risk assessments, and reporting processes that match their obligations — a compliance build-out for firms that previously operated outside the AML regime.

The takeaway: sharing fraud intelligence is now a legal tightrope, and a growing set of professional services has to build compliance machinery from scratch.

Can AI fight the fraud it enables?

Adactin's answer is yes, with conditions. The same technology helping criminals craft scams can help businesses detect suspicious activity earlier and speed up investigations, the firm said. Potential uses it listed include:

  • Analysing payment patterns
  • Spotting suspect identity documents
  • Reviewing emails and invoices for signs of impersonation
  • Helping staff connect related fraud cases

AI may also cut false alerts by separating normal customer behaviour from genuinely suspicious activity. That matters operationally: fewer dead-end alerts mean fraud teams can spend more time on higher-risk cases instead of clearing noise.

But Adactin drew a hard line on accountability. Systems can produce misleading results, make mistakes, or be manipulated if deployed without oversight, testing, and clear rules on data access and decision-making. For organisations already using AI tools, the firm said they should know which systems are in use, what information those tools can access, and what decisions they are allowed to make. Regular testing and monitoring should form part of that process.

"The aim should be to use AI to support better decisions, not replace accountability," Bhanage said.

That quote is the hinge of the whole advisory. AI can watch payments and flag anomalies, but a human still owns the outcome.

What should businesses actually do first?

Adactin's sequencing favours foundations over flash. Combining data across payment systems, customer records, and supplier databases comes first, because every later control — AI detection included — depends on seeing the full picture. Identity verification and payment monitoring follow, backed by the manual safeguards of call-backs, dual approval, and trained staff.

The context makes this urgent rather than optional. Australia's real-time payment rails mean fraud losses are near-instant and hard to reverse, its expanded AML regime pulls thousands of newly regulated professionals into scope, and AI-generated impersonation is lowering the cost of attack for criminals while raising the detection burden for everyone else.

Bhanage's framing suggests the practical test for the year ahead: businesses that pair AI-assisted detection with human accountability and basic verification discipline will absorb the pressure, and those that leave manual approval chains unguarded will keep paying for it.

Source: itbrief.com.au

Share this article:

More from Sophie Lindqvist

Sophie Lindqvist

Show full bio

Staff writer covering regulation, policy and their impact on artificial intelligence and machine learning at AI In Context.

628 articles

Related articles

  1. Australia Warns AI Will Industrialise Scams as Losses Hit $2.18 Billion
  2. INTERPOL: AI Supercharges Existing Cybercrime, Not a New Threat
  3. Apate's AI agents have wasted 17 years of scammers' time
  4. Two-thirds of Australian firms hit by AI-linked cyber threats: ESET
  5. JPMorgan, Amazon and Kohl's Are Blocking AI Agents. Experts Say They Can't Win That Fight

« Previous article