With ChatGPT leading the conversation as its poster boy – seems like ‘Artificial Intelligence’ has become everyone’s buzzword-of-the-day everyday.
However, amidst the AI frenzy, a paradoxical truth emerges: the majority fail to grasp its limitations. Instead, AI has become the in-vogue label, slapped onto everything from toasters to toothbrushes. In this wild, AI-infused ride, we need to unveil the enigmatic allure of AI and the need to unmask its true potential, emphasising that while AI is supercharged, it’s not the magical fix-all potion many believe it to be – (Atleast not at the moment, that is!)
Don’t get me wrong, I have nothing against the technology in itself, in fact I myself enjoy ‘prompt-engineering’ on OpenAI‘s ChatGPT and/or DALL-E Open Ai. I seldom hold back when it comes to creating AI-generated images of dogs wearing ties!
This being said, AI has emerged as the reigning bandwagon, on which everyone wants to jump on – capturing the attention of businesses and individuals alike. And, behind the façade of AI obsession lies a disconcerting truth. Many organisations have fallen into the trap of using AI as a blanket statement to mask their inefficiencies and incompetency in delivering meaningful technological advancements.
In the realm of debt recovery, especially with the rise of Digital Lending, the obsession with AI has become an even more pervasive trend among organisations. In an attempt to explore and understand how the allure of AI can sometimes serve as a convenient shield for shortcomings in the collections industry we need to emphasise the crucial need for genuine innovation and technology that surpasses superficial AI adoption, stressing that true success in debt recovery relies on effective practices, historically tested methods, and customer centric approach and not just being shoehorned into a fake promise of artificial intelligence.
But then, how do you use technology to make your collections effective and ethical?
One basic concept behind understanding the process of Debt collection is that it comprises 2 important elements, namely – Segmentation and Journeys.
Segmentation is defined as the process of segmenting a borrower into mutually exclusive categories based on borrower behaviour, borrower response, the Loan details and the Customer information.
Whereas, Journey is defined as the steps and actions taken for each category of borrower in order to maximise payment outcome and minimise cost and optimise for resource allocation.
The current technological challenges faced by almost all Lenders – Banks, NBFCs, Fintechs alike when it comes to Debt recovery are multi-faceted in nature: Inefficient campaign strategy automation leading to generic journeys, limited integration of traditional/digital communication channels, and an inability to cater to unique borrower circumstances at scale. These challenges collectively impede optimal resource allocation and hinder payment outcomes. Adding to this, if the activity is done at scale, it only leads to more operational inefficiencies, and in turn more losses.
So, how do you harness the power of tech? Well, the only way to use technology in this case is the same as any- Create a logic for what you know! What this means essentially is – if you know what to do, you can program a machine to do the same. If you know what to learn from, you can program a machine to learn from the same.
But now comes the bigger question – Do you know what to do and what to learn from?
Many believe, creating a Digital Infrastructure that brings together various communication channels that campaign to customers in the form of urges and reminders are enough. However, this approach is not only common across the industry but has become pretty old-school now. Even when it works, it is revealed as ineffective way of doing things.
Despite Lenders having a great omni-channel presence with large-scale Tele-calling and Field Agent setups, most fail to turn it into a high-performance venture. On the contrary, they end up in huge losses due to poor resource allocation and/or not recovering the debt altogether.
This only goes to show that no matter how robust the communication engine is, if not executed properly it will fail.
In view of this, currently there is a great way of figuring out the answer to the ultimate question – “How do you know what to do and what to learn from?”
What you do is – segment borrowers in a manner that is mutually exclusive and universally inclusive, upon doing this, you shall be posed with a situation to run different journeys for these segments/categories.
What you learn from is – the Borrower Behaviour and response. Behaviour towards the journey you took them through and response to their idiosyncratic needs.
With the current functionality within collection systems with Lenders – Journeys are very similar to each other. In other words, even if there are a variety of segments, the corresponding journeys are still the same, leading to no efficiency improvements.
The idea here is to categorise borrowers into mutually exclusive categories based on
- Behaviour
- Input or response
- Loan type and details
- Customer-level details
and apply specific strategies unique to each borrower segments’ pain points.
The effective use of technology in debt collections and recovery goes beyond the superficial adoption of AI. It starts with a deeply understanding borrowers, segmentation and the crafting of tailored journeys to address the unique circumstances. By focusing on empathy, communication, and personalised solutions – Lenders can figure out what they want to do and how they want to learn.
Today, AI doesn’t know what to do and how to do it. It’s nowhere near what we’re dreaming of. It can start with being a promising aid to improving Operational Efficiency, enhance monitoring, quality monitoring, asses collections risk, and a few more things. For now, it has to be trained, monitored, improved and trained again with human generated insights.
Tomorrow, the possibilities are as far as you can dream.