To understand the value of AI in credit risk assessment, Analytics India Magazine connected with Sandeep Anandampillai, founder and Chief Product Officer of Crediwatch.
Artificial intelligence (AI), machine learning, and predictive analytics are reshaping the financial services landscape, enabling banks to analyse customers’ data better to provide them with loans efficiently.
As part of this, credit risk analysis and credit risk management are fundamental to banks and financial institutions that provide loans to retail and SMEs. To understand the value of AI in credit risk assessment, Analytics India Magazine connected with Sandeep Anandampillai, founder and Chief Product Officer of Crediwatch.
Unique Selling Point Of Crediwatch
Crediwatch implements actionable credit analytics and dynamic credit assessment insights as a service to financial firms. The company can achieve this without human intervention by applying AI/ML and NLP tools, which produce the most dependable and extensive real-time insights.
Crediwatch is an insights-as-a-service platform that deploys scalable deep learning tools across different digital footprints left by large and small private entities. The startup makes use of over 18 million risk profiles of prominent companies and unregistered small firms.
The platform is intended to deliver sharp insights over the credit lifecycle, from pre-disbursal to post-disbursal assessment, through its tools such as Early Warning System (EWS). The platform plans to give lending companies, corporates (large, MSMEs, SMEs and small business units) the capacity to manage and allocate credit effectively.
According to the founder, India has more than 50 million small and medium enterprises who face the problem of liquidity crunch. Out of this, only 15% get access to formal credit due to the trust deficit that exists and lack of collateral. And for the ones who do get access to formal credit, they have to wait for 4 to 6 weeks to get their loan processed at a staggering rate of 16% – 24%. This scenario creates a debt financing gap of 1 trillion dollars in the market, and hence, these small and medium enterprises are under-banked and underserved.
“There are a few data aggregators in the market who provide platforms to access public data. At the other end of the spectrum, some KPOs use analysts and operations staff to pull information – both of these models do not allow for scale and efficiency. At Crediwatch, we focus on bringing insights to our clients through a ‘zero human touch’ technology platform,” said Sandeep.
Utilising 25,000 Data Points
Crediwatch employs AI/ ML algorithms on alternative data points, such as statutory payment statuses, litigations, media sentiment, GST invoice data, bank statements, as well as traditional data points such as financial ratios, industry outlook etc.
“We realised that concentrating on quality risk and business insights by applying a proprietary AI-based predictive engine is the way forward for the financial services industry and this has helped us differentiate ourselves as a leading analytics player,” Sandeep said.
The platform pulls up data from 25,000 different points, which is the biggest in the industry, from already existing data in the regulatory framework. On the other hand, a typical bank only scans 200 data points. Thus, this scale is what makes Crediwatch unique from additional credit assessment financial institutions, according to Sandeep.
Moreover, the startup has completed the development of the enterprise version of its flagship product, Early Warning System. This product complies with the RBI framework and is based on a proprietary library of more than 190 early warning signals. The software includes a case management module to track alerts and manage post-alert actions from the respective portfolio manager.
The Tech Stack
If you look at the main components of the tech stack, here are some of the technologies used at Crediwatch, according to Sandeep Anandampillai.
• Front end: React, Flutter
• Back end: MongoDB, Redis, Kafka
• Code: Python, Ruby
• Infra: AWS, Azure, DigitalOcean
Crediwatch: The Focus
For now, the focus remains on enhancing offerings to the SME segment as the startup foresees immense potential in this sector. It is building a dynamic “Trust Score” derived from millions of data points that are extracted and analysed across thousands of formal and alternative sources to help lenders assess borrowers and monitor them close to real-time.
“Our vision is to reimagine SME credit by scaling trust through verifiable data, insights and good behaviour. We comprehended the gaps in the market and envisioned a solution that the loan deficit only could be addressed if the lenders have valuable data-points of these small and medium businesses to bank them,” added Sandeep Anandampillai..