Debt collection in Asia typically involves small time agencies who use hostile calls and physical confrontation to get debtors to pay back their loans to banks and other financial institutions. Since long, this field has been bereft of innovation. Asking someone to pay back his or her loan is never easy. On the other hand each unpaid loan represents a near-perfect one-to-one opportunity cost. The bank, financial institution or fintech could have used the same money to lend capital to another consumer who might have repaid it promptly.
Nonperforming loans (NPLs) are a threat to both the financial institution’s bottom line and everyday operations. Introducing radical innovation in the loan recovery process is a Singapore fintech startup that largely eliminates physical action with a Software as a Service solution and specialised services leveraging data and artificial intelligence – AsiaCollect. AsiaCollect’s CEO and Founder Tomasz Borowski tells International Finance in an exclusive interview how AsiaCollect intends to digitally transform and industry given to strong legacy practices.
Tomasz Borowski: The debt collection industry in emerging Asian markets are typically still managed by small, provincial collection agencies, who don’t have the adequate resources to invest in proper infrastructure and systems, relying heavily on manual phone calls and traditional field collection where unqualified agents use hostile and intimidating methods to recover debt. This poses a huge reputational risk to the banks and non-bank lenders that use these agencies.
We saw an opportunity in using advanced technologies like machine learning and AI to transform the way debt is collected and managed, drawing from my experience running operations at a large retail bank in Eastern Europe.
International Finance: One difference of AsiaCollect to other CMS companies is that it buys the NPLs of financial services organisations. How does this model work and what does AsiaCollect do with the purchased NPLs?
Tomasz Borowski: Our integrated credit management suite includes outsourcing and advisory in debt management – this is what is typically offered by other collection agencies as well. As you mentioned, we also purchase unsecured consumer non-performing portfolios. In the first case, we get in touch with originators who are not necessarily banks – they can be finance companies, fintechs, and other lenders. We manage collections on behalf of our clients. It is usually for a period of three months. We are more like a servicer, while the case is still on the balance sheet of the lender.
In the second case where we purchase portfolios, we become the owner of those loans. So the loans are transferred to our balance sheet and we can start performing actions as AsiaCollect becomes the new owner of that debt. Keeping in mind the type of loan or portfolio we purchase, we can work on it for three or four years depending on the model we use.
Before we purchase the portfolio we have to evaluate every single debt from different angles and based on the estimation, prepare the price offer. We spend quite a lot of time on analysing all the portfolio details, it mainly depends on the case and can last between a few days to a month. We analyse the portfolio using our purchasing model. The purchasing model is fed with the information about portfolios that we previously worked with.
Each portfolio we analyse has hundreds of characteristics that we are checking, such as demographic data and behavioural history that we take note of. Following the analysis, we check how the clients with similar characteristics performed previously and whether there is a possibility of debt recovery from them in the future. Grouping such debtors into homogeneous groups and estimating the potential recoveries and costs for each group ends with sending the offer.
Once the portfolio is purchased the process is automised – we upload the portfolio into our platform, apply the strategy that best suits the portfolio and all further actions are executed by our system.
International Finance: What is the process you follow to bid and evaluate the NPL portfolio to achieve the best results for all parties involved?
Tomasz Borowski: With regards to bidding, we need to take into consideration what would be the extent of our investment into this process. This means we need to look into all the actions we will need to perform on debtors, how much it will cost and what could be the potential recovery from the portfolio. This way we will also be able to offer a pricing to all our clients. Then we sign an agreement and the data is transferred. After that we choose and apply the strategy that is the most efficient for portfolio with such characteristics.
One of our differentiating factors is our local experience in working with different portfolios and customising individual strategies. These strategies are constantly challenged by alternative strategies applied simultaneously to find even more efficient ways of higher recoveries: a never ending, champion-challenger approach.
International Finance: Could you describe with an example how AI and ML come into play in AsiaCollect’s CMS solution?
Tomasz Borowski: AsiaCollect CMS solution is fully driven by Data and AI. AI is playing significant roles in our solutions from every aspect: we use predictive calling to improve the efficiency of our agents. We built machine learning models to predict the best timing, channels and strategies to reach the debtors to increase the recovery rate and better experience for the debtors. We use deep learning models to convert voice to text with support of multiple Asian languages to automate our QC process and we are researching and experimenting more exciting AI features such as emotion recognition and psychological profiling.
International Finance: AsiaCollect’s target is to make the debt recovery process in its Asian markets completely automated. Do you think that eliminating the human touch points in debt recovery in a market like Asia is feasible and how will you achieve it?
Tomasz Borowski: I’m not sure the market has reached a stage where the operators could be fully replaced now. But I strongly believe that from three to five years it will be possible. Currently, we already try to minimise the human involvement in the collection process by using the IVMR, Email and SMS campaigns and by sending letters. Our integrated solution for call centre agents is equipped with the dynamic scripts: Using the internally developed voice to text model, we analyse clients` answers so that in the future, we can replace our operators, who currently read the scripts displayed on the monitor by the system, with virtual agents.
International Finance: For banks in Asia that are used to physical collection of debt, moving to an AI and ML based system of recovery with behavioural analysis is a major cultural change. How will AsiaCollect work with banks and financial institutions to bring them to the point of using such third-generation debt recovery solutions?
Tomasz Borowski: I agree that sometimes originators believe that physical contact with the borrower is the only powerful way of collecting the debt. We are trying to convince the financial institutions to give us a chance and to prove that despite the ease of changing the SIM card in the region, we are able, by using our cutting-edge technology to reach their clients without conducting field visits. We have helped almost 50 lenders minimise their losses on outsourced portfolios, amounting to us managing over 265,000 loans every day.
International Finance: AsiaCollect acquired CreditSeva in India. What are your targets for the India market and what is the strategy that you will use to achieve the target?
Tomasz Borowski: Previously, we were focused on expansion in Vietnam and Indonesia. We are still exploring the Indian market, trying to find the best expansion strategy. Large BPOs and hundreds of collection agencies are already on the market but I believe the market in India is large enough to accept a new player – one that is AI-driven and focused on digital collections. Currently, we are focused on offering a Smart Agent SaaS (Software as a Service) solution and convincing originators that debt collection is possible without field collection.
International Finance: The markets you have chosen for your debt collection system, be it Indonesia, Philippines, Vietnam or India are some of the hardest regions to collect debt. What inspired the group of European entrepreneurs behind AsiaCollect to focus on these markets?
Tomasz Borowski: The reason why we reach out to those markets is not because they are the hardest to recover, but because of the very early stage of the collection industry in those markets, thus the opportunity to accelerate the changes. I am really interested in transforming the early stage markets, professionalising the collection activities and helping the borrowers to be more responsible and financially literate. Of course, one of the key factors is market size and growth potential of the markets we have already entered.
International Finance: How do you visualise technology transforming the credit management services market in Asia in the next five to ten years and what role does AsiaCollect plan to play in it?
Tomasz Borowski: I strongly believe that our example would help other professional collection agencies to expand to South Asia and help to accelerate the change from manual, inefficient, and people driven high-risk collection practices to digitalised collection, without reputational risk for lenders and threatening borrowers.
Also, I see ourselves being more active in working with industry players (lenders, collectors) and regulators to make professional, responsible debt collections the norm. We have already started working with industry associations in Indonesia, Vietnam, and in India.
International Finance: In terms of rules and regulations are there any challenges that you face for your CMS in Asia and how do you overcome them?
Tomasz Borowski: I don’t see any obstacles in particular, but in India, for example, we are required to obtain an expensive licence for purchasing debts. From a cost perspective, this is mitigated by a much lower cost of such in Vietnam.