How is an AI-powered debt collection solution helping banks to reduce delinquency?
Nitin Purswani is the CEO & Co-founder of Medius AI. Having spent a significant amount of time developing e-commerce & fintech products, he says, “I love building products from 0-1 & then scaling them from 1-10.”
With roots running deep in the world of technology, Nitin began his entrepreneurial journey nearly a decade ago with Zepo, a SaaS e-commerce platform used by millions of consumers and thousands of businesses.
Lately, Nitin got fascinated by financial technology, particularly the use of artificial intelligence and machine learning in conjunction with traditional banking systems, which inspired him to found Medius, an AI-powered debt collection solution for banks.
In a rendezvous, Nitin answered some of our questions about his startup. Here are excerpts of the chat:
Describe the industry challenges you’re attempting to address with Medius. What observations have you made that compelled you to enter this space?
The banking sector is booming as the Indian economy expands rapidly. Credit is expanding rapidly too, but so is delinquency.
Covid 19 also had a significant impact on the lending industry. Banks and other financial institutions have been issuing warnings of rising NPA. A substantial portion of their losses is due to this bad debt, which costs each major bank billions of dollars. Collection and recovery are also some of the most expensive aspects of loan servicing.
The banks are challenged; everyone knows that and can read about it daily. Their NPA ratios are expected to rise four to six times. They are aware of the serious strain on their balance sheet and are unsure how to address it.
It is the challenge we have taken on – resolving India’s NPA crisis.
How will you solve this issue? What insights did you use to build your solution?
Data-driven debt management can turn this crisis into an opportunity. Smart debt collection and recovery is critical now more than ever, and technology can help lenders retain revenue and customers.
For decades, collection management has been rigid and out-of-date. COVID-19 brought that to light. Historically, the banking industry has been slow to innovate. Most banks were unprepared for the spike in delinquencies. New laws, mobility restrictions, smaller teams, and dispersed operations have made managing it even more difficult.
Lenders typically rely on third parties to collect. However, if we enable banks with new-age technology that helps them reach out to delinquent customers at the right time, the repayment rates would improve with the right strategy. Strategies for collection management that rely on a single channel, or even multiple channels asynchronously, are prone to failure. It’s past time for banks to go omnichannel.
Could you please elaborate on Medius’s products and solutions?
We use the P-A-R model (Predict-Act-Reduce) to help banks reduce NPAs.
Firstly, we predict by combining historical bank data with newly generated borrower information. It gives us the Medius Score, which determines whether a delinquent account will self-cure with little to no activity, roll deeper into delinquency (no matter what we do), or make a payment if properly treated.
Second, we use machine learning to train our conversational AI to collect debt quickly. Our collection and recovery models use deep analytics to realign strategy for better results at every collection stage – early, late, and recovery.
Third, we use legally integrated workflows to resolve past due accounts. Pre-litigation and mediation are used first to try to resolve disputes outside of court. Then we help banks generate and manage large volumes of cases with a single click.
Since Covid, the situation has gotten even more difficult for many banks. How did you handle collection & recovery during the pandemic?
Post-Pandemic, banks were forced to adopt new collection methods due to mobility restrictions on field teams. After my partner, a well-known lawyer, became inundated with requests for collection assistance, we determined that we needed to leverage technology to resolve this crisis.
Our expertise is in data analysis, forecasting consumer behaviour, and optimising customer interactions. Leveraging our strength, we reduced the bank’s reliance on collection agencies, allowing for faster account resolution.
We designed the workflow to handle the massive scale of default effectively. Customer engagement was also a major issue. We live in a 24/7 environment; thus, this process must be digital, omnichannel, and interactive. Using our conversational AI, we were able to get the desired response from clients in 9 out of 10 cases.
Our solutions successfully made an enormous impact on how our banking clients manage, monitor, and mitigate delinquent activity.
How much growth has Medius experienced over the last six months?
Typically, non-performing assets are the third-largest liability on a balance sheet. As a result, when we improve our banking clients’ collection by a significant percentage, they see a significant impact on their balance sheet. And because our growth is directly related to the results we deliver, we have been able to scale our business quickly.
Over the last 2-3 quarters, we’ve seen tremendous growth. We have been able to onboard some of India’s largest banks and NBFCs as clients. The allocated loan portfolio has nearly doubled in size as our client base has grown. As we continue to make technological advancements, our loan recovery rate has been increasing quarter over quarter.
We launched two new verticals along with collections and began assisting our banking clients with Pre-litigation and Litigation activities. It resulted in increased collections for our banking clients and a twofold increase in our revenue.
What is the USP of your solution? How are you different from the competition?
On a technical level, AI and machine learning are critical components of our solutions. No other player currently offers a fully AI-driven solution, to my knowledge. To be honest, our goal is to add value with our unique and comprehensive solution rather than compete.
With our omnichannel conversational AI, we reach out to borrowers via seven to ten channels, are interactive, operate primarily in digital modes, are available 24/7, and support channel hopping – all of the things that customers now expect.
As we’ve seen in the market, other solution providers aren’t driving the right insights to contact the right customers with the necessary resolutions to assist them in overcoming their financial difficulties. That’s why their recovery rates remain low.
18% to 20% of the delinquent portfolios we receive are typically recovered in less than 30 days, significantly better than the industry average.
Do you have any immediate plans to raise funds?
Every business requires the appropriate resources to grow. And so would we, to bolster our teams and product offerings even further at the right time.
We are currently self-sustaining and profitable, and our primary focus is on expanding our product offering and client base. We plan to begin fundraising by the end of this year.