Highlights

  • The RBA has identified agentic commerce as an emerging issue within its payment regulation review.
  • Transaction authorisation and accountability are key concerns surrounding autonomous AI systems.
  • International technology and payment companies are developing frameworks for safer AI-driven transactions.
  • The RBA is encouraging industry-led solutions while considering potential competition and efficiency implications.

Artificial intelligence is introducing new questions about how financial transactions are initiated, authorised and monitored. Speaking at the AFIA Conference on 20 October 2026, Reserve Bank of Australia payments policy head Ellis Connolly highlighted agentic commerce as an emerging area of concern, particularly as autonomous AI systems become capable of conducting transactions on behalf of consumers.

AI Agents Introduce New Payment Authorisation Questions

Agentic commerce refers to systems in which artificial intelligence can perform commercial activities on behalf of users, potentially including searching for products, selecting services and initiating payments.

Unlike conventional online transactions, these systems may operate with varying degrees of autonomy.

This creates questions about how payment providers establish whether a customer has genuinely authorised a transaction.

Connolly highlighted concerns about transaction safety and accountability when autonomous systems behave unexpectedly.

These issues may become increasingly relevant as businesses integrate AI-driven purchasing tools into their services.

Consumer Trust Remains a Central Challenge

Payment systems depend on reliable authorisation processes and clear responsibilities between customers, merchants and financial institutions.

Autonomous transactions introduce additional complexity because an AI agent may carry out instructions without the customer directly approving every individual step.

Potential disputes could involve transaction authorisation, incorrect purchases or unexpected activity.

The RBA is therefore considering how existing payment safeguards may apply to AI-driven commerce.

Connolly indicated that the scale of stakeholder concern about agentic commerce had been greater than anticipated.

Industry Developing New Safety Frameworks

Global payment networks, technology companies and AI developers are already exploring frameworks intended to improve the safety of autonomous commerce.

These initiatives may help establish standards for authorisation, transaction monitoring and accountability.

The RBA has expressed support for industry-led approaches, particularly because artificial intelligence technologies are evolving rapidly.

Connolly explained that traditional regulatory processes can take between 18 months and two years, creating challenges when technology changes more quickly than formal rules can be developed.

Competition Could Become Another Important Issue

Beyond consumer protection, the RBA is also examining the potential effects of agentic commerce on payment competition.

Autonomous systems could influence which payment methods consumers use and which providers merchants accept.

The central bank wants to preserve payment choices as new technologies become integrated into commercial transactions.

Future policy considerations may therefore extend beyond security to include competition, efficiency and market accessibility.