Forrester: The banks that win will learn to tell authorized AI agents from malicious ones, not block them outright
Forrester calls blocking a prudent short-term fix, but says it will not work in the long run.
What's happening
According to Forrester's blog post, 50% of surveyed consumers in the US and Canada still do not trust AI with sensitive financial information. In addition, 40% of surveyed consumers are concerned about data security and confidentiality. The approach Forrester recommends: authenticate both the customer and the agent, verify the agent's provenance, assign a unique identity to each agent, and limit access to only the permissions the task requires.
Behind the number
Forrester asked consumers in the US and Canada how much they trust AI with financial information and tasks.
Whose desk it lands on
The executive in charge of digital channels at a bank. How will the bank recognize an AI agent the customer has authorized?
Our view
The CX Brief's interpretation, not a claim made by the source.
The post's recommendation is to move away from the block-or-allow binary and grant trust in stages. Since half of consumers still do not trust AI, there is no need to rush, but banks that reject agents from the outset may find themselves excluded from the experiences customers prefer.
Source
Source:Forrester·Sample: US and Canadian consumers
Reported by: Forrester Blogs·October 27, 2026
Read it on Forrester Blogs →
