Discussion

FIS chief says banks want AI, but trust and oversight set the pace

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#14680

FIS chief executive Stephanie Ferris says banks are weighing how quickly to adopt AI, with priorities ranging from making online services accessible to agents to automating back-office work. Her example of an AI agent assembling records for suspicious-activity reports shows how that could work: the system recommends whether to file and gives its reasons, while a person reviews the recommendation.

Watch Desk analysis

What happened

Ferris made the comments in an interview with PYMNTS at The Clearing House’s annual conference on 18 November. She said banks are also considering tokenised deposits to reduce their own transaction costs, while customer-facing networks may come later. How quickly a bank invests depends on its business, she said, and secure infrastructure is a requirement either way.

Ferris said FIS’s anti-money-laundering agent has reduced manual data retrieval by 90%. She described it as gathering records and recommending whether to file a suspicious-activity report. In testing, the agent and a human reviewer sometimes gave different reasons for filing, reflecting the bank’s own risk policy. Ferris also said FIS uses AI to help engineers address software vulnerabilities.

Why it matters

This is a practical picture of AI entering banking: not simply a chatbot bolted onto an app, but tools that could gather evidence, support regulated decisions and act on a customer’s behalf. That raises a basic question before an agent can move money or apply a bank’s rules: what exactly has the customer authorised?

Ferris’s account also makes the boundary between automation and oversight more specific. She expects individual steps to run without someone checking each one, but says a person should still oversee how decisions are made. The 90% figure is her account of FIS’s results, not an independently established measure of performance across banks.

Our read

The most useful detail is the AML example. An agent that gathers records and explains a recommendation could free staff from paperwork without handing the final judgement to a machine. That is a more grounded prospect than promising that AI will run the bank. The harder test is whether the reasoning, permissions and review process hold up when a decision matters.

What to watch

  • Whether FIS publishes more detail on how it measured the reported 90% reduction.
  • How banks define customer permission before agents can act on accounts.
  • Whether human review remains meaningful when agents handle more steps in regulated work.

Discussion spark: Should banks let AI agents carry out individual steps without human review if a person still oversees the process, or should every consequential decision require a human check?

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.