By Sudhir Padaki, Director-Data Analytics, Asia Pacific, Altair
Over the past few years, the use of data has grown exponentially as businesses adopt a digital-first approach.
In digital and data-driven organisations, analytics is critical for success. Analytics lets businesses understand and segment their target customers better, upsell and cross-sell, predict customer acquisition and churn, innovate product offerings to improve competitiveness, and prevent fraud.
At the moment, a mere 12% of all enterprise data is used to make business decisions. Only 29% of businesses are successful at connecting analytics to action. And one in five enterprises believe they have lost customers due to using incomplete or inaccurate data. Moreover, while machine learning (ML) is widely acknowledged as key to overcoming these failings, a staggering 96% of enterprises admit to struggling with aspects of the technology.
Manging Fraud and Risk in Financial Institutions
Today’s financial institutions increasingly rely on analytics and data-driven capabilities to support risk and regulatory compliance priorities.
The modern advances in predictive analytics and machine learning empowers financial institutions to make better decisions mapped to defined business objectives. By accurately predicting future consumer behaviour, they let credit risk analysts, financial marketing analysts and fraud detection teams better deploy strategies to take advantage of opportunities while preventing disruption to their business models.
Financial fraud takes countless forms and involves multiple aspects of business, including insurance and government benefit claims, retail returns, credit card purchases, under and misreporting of tax information, and mortgage and consumer loan applications.
Accessing the data necessary to detect fraud is often very difficult. Necessary data are frequently found in semi-structured content in PDF files, text-based reports, or 3rd party systems. Failing to access and transform this data accurately for analysis reduces the detection of fraudulent and anomalous transactions and events.
In the past, banks used rules-based systems to tackle fraud, but these are generally ineffective today since behaviours and tactics change so rapidly. AI-based fraud detection and prevention techniques are much more adaptable and can substantially reduce the harm caused by anomalous and fraudulent behaviour.
Modern AI techniques can find associations and patterns across large numbers of complex datasets. The technology is powerful and can decipher subtle trends, patterns, and exceptions within enormous volumes of data. AI allows financial institutions to build robust strategies to counter fraud. Self-service capabilities are critical to this effort. Business users, compliance officers, and analysts within the banks who understand the myriad types of fraud must be able to update their strategies continuously without having to wait for custom code or help from an IT department.
Smart deployment of the right AI-based fraud detection systems help financial institutions cope with the increasing speed of data and can potentially stop fraud in near real-time. For example, they can interrupt the use of stolen credit cards immediately and detect fraudulent information being entered in an online loan application before it can even be submitted. Daily reports summarizing findings and trends and real-time alerts are critical for such applications.
Digitising Banking Landscape Drives Technology Adoption
The Indian banking landscape is swiftly becoming more competitive.
Thanks to the opportunities it offers to both mature and developing markets, digital banking is growing rapidly across the region. The growth is driven by evolving customer expectations and enhanced digital penetration. Since 2015, for instance, the number of digital banks in Southeast Asia has grown by 190% and is supported by significant investment and supportive regulations. The growth of online transactions that will come with the entrenchment of digital banking will only increase the requirements for rapid yet accurate fraud detection techniques to screen customers.
Unlike traditional brick-and-mortar branch clients, millennial digital banking customers will not wait for days to get an account opened. To acquire and retain customers, digital banks must be able to perform due diligence in a matter of seconds − AI makes this feasible.
Banks have many options for making AI integral to their account management and fraud detection systems, including specialized products and open source libraries like R and Python. Simply having the necessary capabilities is not enough, however.
AI is complex, and the best technology selection will support self-service model development and maintenance, intuitive workflows, high levels of flexibility, a short learning curve, and user interfaces that make it easy for everyone involved in the process to understand exactly how the system makes its decisions. In addition, a complete system must be able to access, join, clean, and manage data from a wide range of disparate sources efficiently and with minimal manual intervention, while still supporting complete traceability for all data operations.
Fortunately with the advanced state of technology today, banks can implement robust, mature tools that will support all of their current requirements, and give them the flexibility to continuously update their workflows and systems to meet the fraudsters’ ongoing onslaught.
The blog has been authored by Sudhir Padaki, Director-Data Analytics, Asia Pacific, Altair.
Disclaimer: The blog has been authored by Sudhir Padaki, Director-Data Analytics, Asia Pacific, Altair. BFSINxt does not necessarily subscribe to the views expressed by the author. BFSINxt will not be responsible for any damage caused to any individual/organisation.
