Article

Could big data analytics and deep learning have detected India’s largest banking fraud?

In short yes. Souma Das breaks it down.

Souma Das
Souma Das
4 avril 2018 5 min de lecture

That’s a Rs. 11,000 crore (and counting) question and the answer to that in a single word is – ‘yes’! What emerges from media reports so far with regards to the case is that there is a strong likelihood of human connivance in the misappropriation of public money. Therefore, what also emerges is that this misappropriation did not involve the use of sophisticated technology. It was, in all probability, (unless investigations prove otherwise) an inside job. But detecting fraud when the deed was done is already too late. What banks need is a business outcome focused, technology enabled analytic solution that identifies the likelihood of fraud before it is committed. So once again, it is good to ask – could mandatory application of technology actually prevent loss? The answer to this is also, subject to relevant conditions.

Understanding banking fraud

It is no secret that mitigating fraud is a top priority for banks. According to the Association of Certified Fraud Examiners, businesses lose more than US$3.5 trillion globally each year to fraud. The problem is pervasive across the financial industry including the insurance industry and is becoming more prevalent and sophisticated. As customers conduct more banking online across a greater variety of channels, devices and geographies, there are more opportunities and more “surface area” for fraud to occur. Adding to the problem, fraudsters are becoming more creative and technologically savvy. They’re also using advanced technologies like machine learning and new schemes to defraud banks. Conventional approaches to fraud detection and remediation are necessary but they remain effective to a point, as conventional tools simply cannot effectively and economically process what is known as big data. “Big data analytics” can enable companies to deploy and integrate rich and new data types to produce new and more sophisticated analyses against the fraudsters and continuously improve the loop of legacy approaches to the war on fraud. In a few test cases, these analytics are extremely effective at exposing not just the fraudsters themselves, but their networks and the people, places and processes they touch or will touch.

It is no secret that mitigating fraud is a top priority for banks.

The Reserve Bank of India (RBI) has strongly recommended the use of technology to curb fraud. Key recommendations include setting up a transaction monitoring group within the fraud risk management group, alert generation and redressal mechanisms, dedicated e-mail ID and phone number for reporting suspected frauds. Old methods for identifying fraud, such as using human-written rules engines, catch only a small percentage of fraud cases and produce a significantly high number of false positives. False positives, as the term suggests, is linked to cases that show up as fraud but do not consequently require a substantial investment of time, people and money to investigate what turns out to be dead-ends. While it is not possible to ascertain what systems and processes were used in the recent fraud case as details are hard to come by, it stands to reason that red flags either did not crop up or if they did, were missed or were willfully ignored. Banks, like many other companies, also face the challenge of having a very small team tasked with investigating an overwhelming number of fraud cases. Whatever the case may be, would human intervention alone ever be capable of countering financial fraud?

Fraud detection & mitigation

Timing is critical. Bank officials need to identify fraud before making a payment or providing a loan because it’s difficult to recover money once its paid. The solution is therefore to make a strategic decision to apply innovative analytic techniques like deep learning including neural networks, machine learning and Artificial Intelligence (AI), to better identify instances of fraud while reducing false positives. Today, AI-driven fraud platforms can analyse incoming transactions in less than 300 milliseconds. Given that banking is now completely electronic, fraudsters whether individual or corporate, despite their best efforts to stay undetected and appear legitimate, will leave tiny data traces which can be correlated to millions of other transactions and with the help of data visualization, a connection and correlation can be made to identify the fraud. Bankers can identify suspicious behavior and withhold payments or loans that appear fraudulent. Fraud investigators armed with technologies such as machine learning and advanced analytics could then review those cases for further action. AI and deep learning models will also allow the bank’s engineers, data scientists, lines of business, and investigative officers from Interpol, local police, and other agencies to collaborate to uncover fraud, including sophisticated fraud rings.

Fraud detection & mitigation

Timing is critical. Bank officials need to identify fraud before making a payment or providing a loan because it’s difficult to recover money once its paid. The solution is therefore to make a strategic decision to apply innovative analytic techniques like deep learning including neural networks, machine learning and Artificial Intelligence (AI), to better identify instances of fraud while reducing false positives. Today, AI-driven fraud platforms can analyse incoming transactions in less than 300 milliseconds. Given that banking is now completely electronic, fraudsters whether individual or corporate, despite their best efforts to stay undetected and appear legitimate, will leave tiny data traces which can be correlated to millions of other transactions and with the help of data visualization, a connection and correlation can be made to identify the fraud. Bankers can identify suspicious behavior and withhold payments or loans that appear fraudulent. Fraud investigators armed with technologies such as machine learning and advanced analytics could then review those cases for further action. AI and deep learning models will also allow the bank’s engineers, data scientists, lines of business, and investigative officers from Interpol, local police, and other agencies to collaborate to uncover fraud, including sophisticated fraud rings.

In June 2017, RBI’s Financial Stability Report called frauds in banks and financial institutions one of the emerging risks to the financial sector. RBI data also shows state-run banks have reported 8,670 “loan fraud” cases totaling Rs 61,260 crore (US$9.58 billion) over the last five financial years up to March 31, 2017. Interestingly the financial industry is also one of the most data-driven of industries. A Deutsche Bank study revealed that at the end of 2012, it was estimated that financial and securities organisations were managing 3.8 petabytes of data per firm. Since then, data sets have grown immensely in terms of size, type and complexity and therefore fraud detection and mitigation has become an equally complex and mammoth challenge. The good news is that AI-enabled big data and analytics can address the problem and stay a step ahead of those on the other side of the law.
Tags

À propos de Souma Das

Souma Das is Managing Director at Teradata India. He brings with him more than 29 years of technology industry leadership experience that comprises of enterprise software knowledge, general management, sales and business development, strategic consulting and professional services and executive management experience.

At Teradata, Souma is responsible for providing leadership & overall strategic direction to the company’s India business overseeing field operations that include sales, customer management, marketing, professional services and customer support.

Souma is a results-oriented executive who enjoys building, coaching and nurturing teams to create high performing talent driving new growth revenue lines for businesses.

Before joining Teradata, Souma was the Regional Vice President and Managing Director for Qlik for its India operations and was responsible for leading the team to drive growth, revenue and customer satisfaction for organisations leveraging Qlik’s analytics platform.

Prior to joining Qlik, Souma was the Regional Vice President and Managing Director of Infor in India. Souma also, built and headed the Indian operations for Citrix Systems for close to a decade as their Vice President. He  started his career with Wipro Technologies and moved on to work for IBM across various roles and functions.

Souma holds a Post graduated in Executive Management  from  Duke University – Fuqua School business and has a MS  in Computer Science and Applications from Jadavpur University.

Voir tous les articles par Souma Das

Restez au courant

Abonnez-vous au blog de Teradata pour recevoir des informations hebdomadaires



J'accepte que Teradata Corporation, hébergeur de ce site, m'envoie occasionnellement des communications marketing Teradata par e-mail sur lesquelles figurent des informations relatives à ses produits, des analyses de données et des invitations à des événements et webinaires. J'ai pris connaissance du fait que je peux me désabonner à tout moment en suivant le lien de désabonnement présent au bas des e-mails que je reçois.

Votre confidentialité est importante. Vos informations personnelles seront collectées, stockées et traitées conformément à la politique de confidentialité globale de Teradata.