AI fraud threat emerges in customer support, provider warns

By Aarav Garg

Financial services firms are being warned about a growing risk of AI-enabled customer vulnerability fraud, as support services provider MorganAsh says some consumers are using generative AI tools to create sophisticated claims aimed at delaying or avoiding debt repayments.

According to MorganAsh, lenders in the credit sector have reported an increase in highly structured correspondence from customers asserting vulnerability. The messages are said to include polished legal language, references to Financial Conduct Authority (FCA) rules and detailed requests for forbearance or alternative treatment, with characteristics that suggest AI-assisted drafting.

The development highlights an emerging challenge for firms balancing fraud controls with obligations under the UK’s Consumer Duty regime, which requires firms to identify and support vulnerable customers and deliver fair outcomes.

MorganAsh said that where firms lack a prior record of vulnerability assessments or customer engagement, it can be difficult to challenge potentially fraudulent claims without risking regulatory scrutiny or poor customer outcomes. The company argues that the issue is likely to expand beyond consumer credit into broader financial services as awareness of vulnerability protections increases.

The warning comes as many institutions still rely on reactive methods to identify vulnerability, such as complaints, arrears triggers, collections interactions or crisis events. MorganAsh said some firms continue to record only severe cases, despite the FCA’s intentionally broad definition of vulnerability, which can include temporary or less visible circumstances such as bereavement, illness, job loss or financial stress.

That creates two risks: genuine vulnerable customers may be missed until problems escalate, while firms may also struggle to distinguish legitimate claims from fabricated ones.

The provider said passive data sources, including traditional credit information, and tools such as call-centre voice analytics often identify issues only at the point of distress rather than building an ongoing profile of customer needs.

As a result, firms may need to shift toward more proactive and evidence-based models that regularly assess vulnerability throughout the customer lifecycle. This could include structured questionnaires, periodic reviews, staff training and documented support pathways.

Andrew Gething, managing director of MorganAsh, commented, “It is unfortunate to hear of an increase in potentially inaccurate claims of vulnerability, although we can understand the acceleration due to AI. While we mustn’t rule out those using AI to assist with communication or overcome barriers to engagement, we have to be vigilant of vexatious claims. Firms that have adopted a proactive approach to customer vulnerability management are well-positioned to rebuke fraudulent or inaccurate claims. By engaging with customers proactively and recording assessments at the earliest opportunity, firms hold clear evidence of what was disclosed and robust records of characteristics which can be provided as evidence in the event of any future claim.”