Consumers are beginning to imagine AI doing more than comparing financial products. Experian has published research, conducted for it by Forrester Consulting, in which 54% of surveyed consumers said they would be comfortable with an AI agent applying for credit on their behalf.
The survey covered 6,247 credit-active, digitally literate consumers across 13 EMEA and Asia-Pacific markets in July 2026. Eighty-two per cent said they trusted AI to compare loans across providers. But the autonomy numbers are more nuanced: 23% would allow an agent to act when pre-agreed rules were met, while only 5% said they would be comfortable giving it full autonomy.
That distinction matters. The findings are not evidence that consumers are ready to hand over autonomous control of their financial lives. They do suggest that a meaningful share of this digitally engaged sample is open to delegating defined parts of the lending journey to software.
For a bank, that changes the trust model. A conventional digital application is roughly person → identity → consent → application. Agentic finance adds another link: person → identity → agent → delegated authority → application. The institution may need evidence not only of who the customer is, but of what the agent was authorised to do, under which constraints and for how long.
Experian itself identifies identity verification, consent management, fraud prevention and explainability as critical requirements as agent-assisted customer journeys develop. It also reports that 75% of respondents would feel more comfortable using AI connected to a financial institution they already trust.
The next competitive question for banks may therefore be less about whether they launch an AI assistant and more about whether they can support secure transactions initiated by customer-side agents. If the agent can compare, apply and eventually negotiate, the bank needs a reliable way to distinguish legitimate delegation from fraud automation.
