ANZ adds governance layer as use of AI agents grows
The bank says its enterprise AI control plane will impose operating limits and halt activity that falls outside them.
By Quang Dinh
ANZ Banking Group has introduced a central governance system for its growing use of AI agents, with group CEO Nuno Matos outlining the initiative at the bank’s annual results presentation.
Matos said the enterprise AI control plane would establish boundaries for agent activity and halt activity outside those boundaries. He said human oversight remained essential as ANZ expanded its use of agents.
The system would make AI decisions traceable and provide oversight and technical controls for models, agents, users, tools, platforms and costs, Matos said. It adds to safeguards already operating at ANZ, including a governance council and several AI gateways.
The underlying software and any involvement by external providers in developing or configuring the system remain unclear. CIO Donald Patra did not provide further architecture details in a blog post.
Matos said ANZ was targeting projects with clear business outcomes rather than adopting AI across every activity. The bank’s aims include more personalised customer experiences, improved tools for bankers, higher productivity and better risk management. ANZ identified customer relationship management, engineering and contact centre applications as areas where it was already using AI.
