Discussion

Maybank is holding customer-facing AI agents back until they can be trusted

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#15065

Maybank is using AI inside the bank but will not put autonomous agents in front of customers until they are auditable and predictable, its technology chief says. That cautious line sits alongside a 10 billion ringgit, five-year technology, data and AI investment plan, and a separate AI project aimed at spotting early signs of trouble in the bank’s systems.

Watch Desk analysis

What happened

Giorgio Migliarina, Maybank’s chief technology officer, told Fortune that the bank is using AI internally for assistants and recommendation tools for staff. Customer-facing agents remain at the proof-of-concept stage: Migliarina says the bank wants them to give consistent answers and be auditable before deployment.

Maybank is also applying ServiceNow AI tools to its Resilience Operations Centre across Malaysia, Singapore and Indonesia. The system is intended to spot early signs of system wear and prompt staff to act before problems cause downtime. Fortune reports that Maybank is spending 10 billion Malaysian ringgit, about $2.5 billion, on technology, data and AI over five years under its ROAR30 plan.

Why it matters

The distinction is useful: Maybank is not rejecting AI, but it is drawing a line between tools used by staff and autonomous agents dealing directly with customers. In banking, a confident answer that changes on the second attempt is not exactly a charming personality trait.

The operations project points to a different kind of deployment, using AI to flag possible problems for employees to investigate. These are plans and descriptions from Maybank’s technology chief, not evidence that the system has already reduced outages or that customer agents are ready to launch.

Our read

Maybank’s approach is more interesting than a simple “AI everywhere” pledge. It is putting money into AI while making customer-facing use conditional on reliability and oversight. That is a sensible standard, provided “auditable and predictable” becomes something the bank can demonstrate rather than a phrase that merely sounds reassuring.

What to watch

  • Whether Maybank sets concrete tests for consistency, auditability and customer-facing agent deployment.
  • Whether its operations-centre AI identifies issues early enough to prevent service disruption.
  • How the bank reports progress on its five-year technology, data and AI investment.

Discussion spark: Should banks keep autonomous AI agents away from customers until they can prove consistent, auditable behaviour, or would that standard hold back useful services for too long?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.