Fraudsters used a WhatsApp message and an AI-generated voice clone to persuade a senior executive at Fideuram – Intesa Sanpaolo Private Banking to authorize more than $100 million in overseas transfers. In a September 29, 2026 report citing Reuters, TechRadar said authorities recovered about $60.14 million, while roughly $40.8 million remained missing after apparently being converted into cryptocurrencies.
An Urgent Request From a Familiar Name
The attack reportedly began in February 2026 when Fideuram chairman Paolo Molesini received a WhatsApp message that appeared to come from Intesa Sanpaolo CEO Carlo Messina. The sender asked for help with an urgent overseas transaction. A later call from someone claiming to be a senior partner at a prominent law firm appeared to confirm the instruction.
According to the report, the caller used AI tools to imitate the lawyer’s voice. Believing the request was authentic, Molesini contacted the bank’s finance department and arranged multiple transfers, mainly to accounts in mainland China and Hong Kong.
Recovery Efforts Left Millions Missing
Fideuram identified the fraud soon after the transfers and attempted to block them. Cooperation among law enforcement agencies in China, Portugal and Italy helped recover approximately $60.14 million, but the report said around $40.8 million had not been recovered.
Molesini stepped down as president of Fideuram and Intesa Sanpaolo Private Banking in March 2026, citing personal reasons. Reuters reported that neither he nor other Fideuram executives were under investigation. Local authorities were reportedly investigating a foreign national living outside Europe for possible fraud.
Business Compromise Moves Beyond Email
The method extends the familiar pattern of business email compromise, in which attackers impersonate senior leaders and pressure finance employees to make confidential transfers. In this case, the deception crossed messaging and voice channels, allowing the criminals to present one apparent identity through multiple forms of communication.
The incident shows a material weakness in approval processes that rely on recognizing a sender’s account or voice. AI-generated impersonation can make those signals easier to counterfeit, particularly when an urgent request is reinforced by a second supposed authority figure.
AI-generated editorial illustration; not a documentary photograph.
Source: Italy's top bank hit by an AI messaging scam which cost it nearly €100 million
SIGNAL TITANS
Leave a Reply