DAVOS, Switzerland — Financial services executives who gathered recently at the World Economic Forum were increasingly focused on a new phase of artificial intelligence that could reshape banking, payments and insurance operations: agentic AI, systems designed not only to generate insights but to act autonomously across complex financial workflows.
At panel discussions and private meetings in Davos, bank CEOs, fintech leaders and regulators described agentic AI as a potential inflection point for the industry, capable of automating credit decisions, fraud monitoring, compliance reviews and customer interactions at a scale far beyond today’s generative tools.
Unlike earlier AI models that support human decision-making, agentic systems are designed to initiate actions — such as moving funds, adjusting credit exposure or escalating compliance issues — with limited human intervention. Executives said the technology could deliver major efficiency gains in an industry facing rising regulatory costs, margin pressure and growing customer expectations for real-time service.
Opportunity, Risk are ‘Enormous’
“Financial services are built on workflows, controls and decision chains, which is exactly where agentic AI is moving,” said one banking executive attending the forum. “The opportunity is enormous, but so is the risk if it’s not governed correctly.”
Risk management and trust emerged as central themes in Davos discussions. Regulators and bank leaders cautioned that autonomous systems operating in financial markets raise new questions around accountability, explainability and consumer protection — particularly in areas such as lending decisions, anti-money-laundering monitoring and payments authorization.
Amplification of Errors?
Several executives noted that while agentic AI could strengthen fraud detection and operational resilience, it could also amplify errors or bias if models act on flawed data or unclear rules. As a result, many firms are exploring “human-in-the-loop” frameworks that keep people involved in high-impact decisions while allowing AI agents to handle routine actions.
The discussions also highlighted a growing divide within the financial sector. Large global banks and payments companies with strong data infrastructure are moving quickly to test agentic systems, while smaller institutions risk falling behind due to legacy cores, fragmented data and regulatory uncertainty.
More Investment Needed
Speakers at the forum said that unlocking value from agentic AI will require more than technology investment. Banks must redesign processes, modernize governance frameworks and work closely with supervisors to ensure autonomous systems align with existing regulatory obligations.
As global financial leaders continue meetings through the end of the week, agentic AI is emerging as both a competitive opportunity and a regulatory challenge — one that could redefine how financial institutions operate, manage risk and serve customers in the years ahead.
