Traditional credit data alone may bias AI credit risk determination algorithms—but adding alternative data can boost financial inclusion
The use of AI and machine learning in credit decisioning is a boon to micro, small- and medium-sized enterprises to secure loans for survival and growth.
Today, a growing segment of executives in financial services see AI-enabled risk decisioning as the cornerstone to organizational improvements—including fraud prevention, automating decisions across credit lifecycles, improving savings and operational efficiency, and more competitive pricing.
As such, many organizations have come to terms with the stark advantage that alternative data, AI and machine learning offer to resolve not only the challenges of the enterprise, but also providing better and more desirable products and services to the customer.
Bharath Vellore, General Manager, Asia Pacific, Provenir, tells the team at DigiconAsia.net more…
DigiconAsia: What is “alternative data” exactly, and how can this improve credit risk decisioning?
Bharath Vellore (BV): While traditional data covers traditional fraud, credit, and open banking identity, alternative data usually consists of new types of data such as behavioral, ecommerce transactions, social media data, and more. This variety of data is essential for the discovery and action on new insights that generate business value and drive superior customer experiences.
- Firstly, through access to a data marketplace, financial organizations can quickly access data, including alternative data from global data providers—to expand their offerings horizontally, across various countries, or vertically into segments previously obscure to the enterprise.
By selecting specific data sources, businesses can create rich and customized datasets to meet any business need. Such data sources can be acquired almost instantly, and firms can quickly test data across multiple decisioning processes easily.
- Secondly, alternative data provides the opportunity for financial inclusion, a challenge that is still prevalent in South-east Asia as some 60% of micro, small and medium enterprises (MSMEs) have been unable to obtain a loan due to unavailable credit history. Alternative data provides more opportunities and allows these MSMEs to now obtain the financial assistance they require through improved credit decisioning. At the same time, financial inclusion also means that other segments such as younger consumers, minorities and immigrants, especially those without a credit history, will also be able to receive the financial support they need for career growth and hire-purchase loans.
According to our own studies, 64% of lenders and credit providers using alternative data are already seeing tangible benefits within a year, more financial institutions will jump on the bandwagon to supercharge their credit decisioning.
DigiconAsia: Fraudsters seem to always be one step ahead of financial institutions. As AI with alternative data boosts financial inclusion, what about fraud detection and prevention?
BV: A key benefit of using AI and ML in financial services involves enhanced fraud detection, and the accuracy and speed gets better with each transaction processed. So, even when fraudsters evolve their methods, AI models can use real-time data to identify new patterns, learn, and adapt decisioning to maximize the right fraud alerts and minimize false positives.
In some studies, most financial institutions that had already adopted AI were being alerted to fraud before it happened, had fewer false positives and indicated that the reduction of payment fraud was a key outcomes of their AI systems.
DigiconAsia: Which countries in the region are leading in the adoption of AI financial and credit decisioning?
BV: When it comes to AI, innovation is not limited to front-end and customer-facing solutions.
In South-east Asia, especially in the Philippines, Indonesia, Malaysia and Singapore, some of the major advancements in AI technology have been for backend operations in the financial sector. It is in this backend that financial institutions can experiment and test new AI-powered solutions and develop a more AI-friendly culture. This provides the testing ground before more potentially brand-advancing AI technologies are deployed for customer-facing functions.
Meanwhile, in other Asia Pacific regions: China already constitutes the biggest share in AI spending and will continue its dominance through 2024, closely followed by other countries such as Australia, India, Korea, Singapore and Hong Kong.
DigiconAsia thanks Bharath Vellore for his insights.