Key Takeaways
- Intesa Sanpaolo is developing AI-based tools to improve suspicious transaction detection.
- The bank's pilot involves the Anti-Financial Crime Digital Hub and Italian public institutions.
- The initiative has identified more than 30 potential tax-fraud cases through an AI-based algorithm.
Intesa Sanpaolo Uses AI to Detect Suspicious Transactions
The initiative focuses on developing new analytical approaches for identifying financial-crime risks.
According to Intesa Sanpaolo, criminal activities can rely on complex schemes involving multiple financial operators, making system-wide cooperation important for financial institutions.
The pilot project aimed to identify money-laundering and terrorist-financing risk patterns and to test innovative AI-based models for detecting suspicious transactions, with particular attention to tax-fraud schemes involving false invoicing.
Public-Private Cooperation Behind the Project
The initiative was developed through the Anti-Financial Crime Digital Hub.
The consortium was established by Intesa Sanpaolo together with Intesa Sanpaolo Innovation Center, Politecnico di Torino and the University of Turin.
The project also involved cooperation with Italian public institutions, including the Guardia di Finanza, the Anti-Mafia Investigation Directorate, the Bank of Italy and the Financial Intelligence Unit.
The collaboration combined banking and research expertise with public-sector input on the parameters used to develop the algorithm.
AI Algorithm Identifies Potential Tax Fraud Cases
Intesa Sanpaolo said an algorithm developed using artificial intelligence techniques was tested by the bank as part of the initiative.
The bank reported that the algorithm was tested on more than 40,000 corporate customers in Piedmont and Valle d’Aosta and identified more than 30 cases of potential tax fraud across Italy.
These are potential cases for further investigation, rather than findings that establish criminal activity.
Why Artificial Intelligence Is Being Used
The pilot focused on testing innovative models for detecting suspicious transactions and on monitoring transactions through dedicated algorithms.
The collaboration between Intesa Sanpaolo, the AFC Digital Hub, the Intesa Sanpaolo Innovation Center, Politecnico di Torino and the University of Turin enabled the sharing of technologies and expertise.
The resulting AI solutions were aimed in particular at monitoring transactions and identifying potential tax-fraud schemes.
Tackling More Complex Financial Crime
Intesa Sanpaolo said criminal activities can rely on complex schemes involving multiple financial operators.
This complexity can make suspicious activity more difficult to identify using conventional approaches alone.
The bank's AI initiative therefore focused on analysing transaction activity through dedicated algorithms and improving the detection of potentially suspicious transactions.
Cooperation With Italian Institutions
The project brought together financial-sector and public-sector organisations.
The participating institutions included the Guardia di Finanza, the Anti-Mafia Investigation Directorate, the Bank of Italy and the Financial Intelligence Unit.
Piero Boccassino, chairman of the Anti-Financial Crime Digital Hub, said new European anti-money-laundering legislation creates opportunities for system-wide cooperation, including data sharing on higher-risk customers under specified criteria.
Pilot Results Reported
Intesa Sanpaolo said the pilot partnership aimed to identify money-laundering and terrorist-financing risk patterns and test AI-based models for detecting suspicious transactions.
It also reported that the tested algorithm identified more than 30 potential tax-fraud cases.
The bank described the pilot as Italy's first significant public-private partnership initiative aimed at combating financial crime. Following the pilot, the Bank of Italy, the Financial Intelligence Unit and the Italian Banking Association launched a nationwide partnership involving major Italian banks.
TwikUp's Perspective
The pilot illustrates how a bank can combine artificial-intelligence tools with institutional expertise to identify activity that may require further investigation.
However, identifying a potential case through an algorithm is not the same as proving fraud. Human investigation, regulatory procedures and appropriate evidence remain necessary before conclusions can be reached.
What Happens Next?
The nationwide partnership launched by the Bank of Italy, the Financial Intelligence Unit and the Italian Banking Association extends the public-private cooperation model beyond the initial pilot.
Intesa Sanpaolo's initiative provides an example of how artificial intelligence is being incorporated into banking risk and compliance processes, particularly for monitoring transactions and detecting patterns linked to potential financial crime.