Autonomous AI 'agentic customers' set to reshape banks

FICO has warned that a rise in "agentic customers"-autonomous AI agents managing day-to-day financial decisions for consumers-will force banks to redesign product pricing, risk assessment and customer relationships.

It cited analysis from Neurons Lab suggesting that, within two to three years, consumers will grant AI agents full financial autonomy. The analysis also links agentic AI to USD $3 trillion in corporate value.

Agentic customers are part of a broader shift towards "machine customers", Gartner's term for non-human economic actors that buy goods and services and make payments. In banking, that could mean software agents comparing accounts, negotiating loan terms and executing transactions with limited or no human involvement.

FICO expects the change to accelerate once early adoption reaches scale. As AI agents become more capable and accessible-and consumers more comfortable delegating decisions-competitive dynamics could shift quickly.

Changing relationships

Retail and commercial banks design many interactions around human behaviour. Product disclosures, customer journeys and service channels assume a person is reading information, asking questions and making final choices. Agent-led decisions could upend that model, with software evaluating offers across providers and acting on pre-set preferences.

FICO argued that banks could face commoditisation if technology companies or specialist fintech platforms build consumer agents that become customers' trusted points of control. Banks would then compete on price and execution while losing control of distribution and the brand relationship.

"Intelligent agentic customers are capable of negotiating deals, comparing offers, executing transactions autonomously, and represent a profound shift in customer behaviour and expectations," said Rachael Hadaway, Vice President of Product Management at FICO.

Hadaway described a future in which "real-time, agent-to-agent negotiation" becomes a standard interaction model. That would compress the time between presenting and accepting offers and put greater emphasis on consistent decisioning across channels.

Operational redesign

Shifting to agent interactions would affect operations across product design, credit policy, fraud controls and servicing. Many traditional processes rely on staged approvals and manual reviews, which can slow decisions and create inconsistent outcomes when customers use multiple channels.

Banks would need to rework technology and processes to deliver approvals and product configurations in real time, FICO said. It also flagged the need to evolve risk and compliance frameworks when autonomous actors initiate and complete actions at machine speed.

Preparing for direct agent-to-agent negotiations would require systems that authenticate counterparties, manage consent and permissions, and apply policy rules consistently. Banks would also need to respond quickly when an agent requests a revised price or different product structure.

FICO also pointed to organisational implications, including building internal expertise to manage and collaborate with AI agents and strengthening defences against attempts to exploit algorithmic weaknesses.

Regulatory friction

Regulatory frameworks will need to adapt as agent-to-agent interactions become more common, FICO said. Many rules assume direct interaction with a human consumer and emphasise disclosures, informed consent and suitability assessments.

It highlighted GDPR, consumer protection laws, know-your-customer requirements and fiduciary responsibilities as areas that may need changes. FICO also warned that regulators typically move more slowly than technology, creating friction as banks try to stay compliant while operating at higher speeds.

For banks, the challenge would include verifying that an agent has valid authority, maintaining audit trails and handling disputes when decisions are made automatically. It would also mean ensuring automated actions stay within agreed risk limits and meet conduct requirements.

Readiness gap

Banks that have already invested in APIs, cloud infrastructure and AI are likely to be better positioned for an agent-driven model, FICO said. Legacy systems and long-running processes, however, could make it difficult for established banks to match the speed and flexibility that agent interactions demand.

FICO framed the issue as a strategic race as much as a technology upgrade. It argued that platforms participating in broader financial ecosystems may have an advantage over closed, proprietary systems, and it emphasised the need for risk management approaches that can adapt quickly to new forms of autonomous activity.

"Banks that invest early in agent-centric capabilities will gain a decisive advantage," said Hadaway. "Those that delay risk being reduced to commoditised service providers. The shift to agentic customers has already begun, and the timeline for change is compressed. "Banks must act now to prepare for a market in which autonomous agents-not humans-become the primary decision makers in everyday financial interactions."

FICO expects continued progress in consumer-facing AI agents, with a growing role in shopping, switching and routine money management. It said banks should plan for a period in which agent-led decisions become a standard pathway for financial products and services.