Banks are ready for AI agents, but their customers aren't: only 14% would let an AI open an account for them

By Cybersecurity Insiders · November 1, 2026.

79% of European financial services providers say they are ready to handle product applications initiated by AI agents. Only 14% of consumers are comfortable letting an AI open financial products on their behalf. The figures come from “The Battle to Onboard 2026: The AI Reality Check,” a report by Signicat, a European digital identity and fraud prevention provider.

The study, conducted by Sapio Research in August 2026, surveyed 4,050 consumers and 450 onboarding, compliance and digital identity decision-makers. It covers nine countries: Denmark, Finland, France, Germany, the Netherlands, Norway, Spain, Sweden and the United Kingdom. It is worth remembering that Signicat sells precisely the solutions (identity verification, anti-fraud) that its own report presents as necessary. The data are the company's and we have not independently verified them.

Consumer preference is clear: assistance yes, autonomy no. According to the report, 65% support AI helping with financial applications, for example by gathering information or preparing the submission for their approval. Only 11% would let it complete the entire application on its own.

Trust also depends on risk. 42% would delegate creating recipes and 38% writing emails. For purchases, 19% would trust AI with small amounts and only 8% with high-value purchases. Only 15% would allow an AI to use their digital identity to verify themselves with their bank, and 14% would let it sign contracts on their behalf.

Another finding: 60% of European consumers abandoned an online financial application in the past year. In 2022 the figure was 68%. After years of investment in digitalization, the improvement is modest. The reasons cited in the report are lengthy processes and requests for too much information.

There is also a counterintuitive finding. Norway, with Europe's strongest preference for digital identity, also shows the greatest rejection of an AI opening products for the customer: 79% say they are uncomfortable. It is followed by Finland (78%), Sweden (73%) and the United Kingdom (69%). Being used to digital does not mean being willing to delegate.

Our take: the gap is not because consumers do not understand the technology. It is because delegation in finance is an act of responsibility. Signing, identifying oneself and committing money are actions with legal and financial consequences, and users want control at that point. This fits a thesis we hold: agentic reliability is built with external layers of oversight, not by trusting the model. Here the key layer is human: the model prepares and the customer approves.

The winners, according to the data, are incumbent institutions. Consumers trust a financial provider (44%) almost twice as much as an AI provider (23%) to safeguard their personal data. If the agent becomes the interface, the bank can retain the relationship as a trusted custodian. But there is a risk: that a large platform's agent captures the customer and relegates the bank to regulated plumbing.

In general, the history of digital finance suggests these gaps close. It happened with online banking and contactless payments, as Signicat's own executives point out. But they close when the system proves its security, not through industry insistence. That only 11% accept full autonomy today is not a permanent ceiling. It is a snapshot from August 2026, before there are benchmark incidents, agent identity standards and clear liability guarantees.

In the short term, the risk is twofold. A bank that enables agents before solving identity and fraud exposes itself to automated deception at scale. And one that ignores the demand for control will lose customers who are already abandoning applications. In the long term, frictionless onboarding, where a trusted assistant handles tedious paperwork with safeguards, is part of the abundance we pursue: less time on paperwork and more on what matters. The path lies in earning trust step by step, starting with assisted, low-risk tasks.