The risks of full automation in lending and what firms can learn from BNPL

The risks of full automation in lending and what firms can learn from BNPL

Last week, FinTech Global spoke to several players in the FinTech space about how automation can improve the customer experience within lending. This is the second part, which explores what the risks of full automation are and what traditional lending companies could learn from the buy now pay later (BNPL) providers.

What to learn from BNPL?

There is a love and hate relationship with BNPL. These services have become increasingly popular over recent years and the current financial market has only accelerated this. The ability to pay for products over monthly instalments can be a lifesaver. But for all their praise, they are met with concerns and reports in the media of people landing in major debt. In fact, complaints in the UK to the Financial Ombudsman Service have risen by 36% in the past three years.

Having been missed by a lot of existing regulations, regulators around the world are starting to build new frameworks that aim to boost protections for customers. For example, the UK’s government recently announced it is giving financial watchdogs more power to clamp down on the space and FCA recently affirmed that BNPL heads could face up to two years in jail if they fail to toe the line on financial promotion rules. Elsewhere, the EU approved new rules to the Consumer Credit Directive that included BNPL, the US Consumer Financial Protection Bureau announced plans to increase regulations around BNPL and the Hong Kong Monetary Authority recently issued a circular to banks on BNPL products to implement new consumer protection measures.

Regardless of the concerns, BNPL has proved to be an increasingly popular service. The global BNPL market size is on track to reach $3.98 trillion by 2030, from $90.69 billion in 2020. This represents an impressive CAGR of 45.7% from 2021 to 2030. According to Amit Dua, president of SunTec, this growth is largely being driven by younger generations that are motivated by convenience. They are more likely to use a provider that has a good customer experience and reduces most of the effort in getting a loan.

Dua said, “The magic behind this convenience lies in automation. BNPL processing is playing an increasingly important role. Today’s advanced algorithms and machine learning models enable BNPL providers to quickly assess the creditworthiness of potential borrowers, parsing through their credit history, income stability, and existing debt levels in near-real time.

“This automation is key to the BNPL approval process, not only making it possible to deliver instant decisions to consumers, but also ensuring the financial stability and reliability of the system. By automating the credit risk assessment, BNPL providers mitigate the risk of defaults while fostering responsible borrowing behaviours among consumers. This level of automation is crucial for the sustainability of the BNPL ecosystem as it balances customer convenience with the financial prudence necessary to maintain the integrity of the system.”

From this, traditional banks need to create lending options that can meet the new needs of consumers, particularly millennials and Gen Z, Dua added. They will need to improve their agility, responsiveness and decisiveness to react and even predict customer needs. “Banks will have to switch to a completely customer-driven model in which they develop credit and lending options for a wide range of customers with very specific, unique needs before the customers themselves realise what they need.”

The ability for BNPL companies to provide customers with convenience is something that Mamta Rodrigues, global president of banking, financial services and insurance at Teleperformance, also noted. By being able to insert BNPL into checkout pages for online purchases has helped them gain customers. This flexibility is something traditional lenders need to take note of.

“The traditional lending sector needs to match the speed and agility BNPL schemes have in their loan processes, in order to meet these on-demand customer expectations. With the speed at which technology is evolving, and seeing how quickly BNPL has taken flight, it will be important for traditional lenders to keep up the pace by embracing digital automation and being available and flexible to service customers where, how and when customers prefer to shop.

“This can be achieved by collaborating with trusted partners who specialise in providing digital business services to financial institutions. In fact, 88% of businesses now use third-party providers for at least one component of their digital transformation strategy. These partners will have the expertise in adapting operating models to optimise business results, and the technological capabilities to build a resilient and efficient lending, payment and collection process.”

Echoing a similar sentiment, Chirag Shah, founder and CEO of Nucleus Commerical Finance, also indicated that the growing popularity of BNPL has been driven by the ability for consumers to quickly and easily apply for loans without needing to go through lengthy applications and waiting months for approval. He said, “Traditional lenders can adopt these practices and technology in their own loan application processes, as well as partnering with fintech to provide BNPL as an embedded lending service through other businesses as an additional revenue stream.”

As the pursuit for greater digitalisation continues, the flexibility and agility of BNPL is something traditional lenders could seek to replicate.

Risks of full automation

While the positives of automation have been highlighted across the two articles: there are still risks of relying solely on automation. For example, Williams noted that one of the biggest risks with full automation is frustrating the user. Instead of being a seamless experience, automation can lead to scenarios where a customer is stuck in a loop and unable to connect with an agent.

He said, “We need to understand the analytics coming out of the touch points and be able to amend those quickly, should a flow change. If we have an error or a failure point, full atomisation may not flag this for some time – however, an agent with the correct mindset will.”

Dua echoed a similar point. While customers often prefer to have an automated interaction where they can do everything quickly and by themselves, that is not always the case. Sometimes they want to speak to a human, whether it is for guidance or just reassurance. This is even more the case when finance is involved. Speaking to a human is essential and can boost the trust they have with a lender.

“Automated systems will need the right blend of human interaction to gain a contextual understanding and provide the same level of personalised service. Automated systems are also often inflexible, following predefined rules and algorithms,” Dua added. “This rigidity may limit their ability to adapt to unique customer circumstances or handle exceptional cases. As a result, the system’s recommendations or decisions may not align with the customer’s best interests. Customers with specific needs or complex situations may not receive the tailored assistance they require, potentially leading to frustration or dissatisfaction.”

Another potential risk from full automation is inaccuracy. Rodrigues stated that lenders need to constantly review parameters for automated decision making in automated onboarding processes as they need to ensure valid applications are not erroneously rejected. She stated that the initial parameters used to assess applications might not be perfect and lead to the rejections. To counter this, advanced technologies like machine learning and AI helps firms automate onboarding as well as learn from customer behaviour to reduce errors.

“This iterative approach allows institutions to learn from patterns and trends in customer data, ensuring a fair and accurate assessment of applications, removing probable algorithm biases. Ultimately, the goal is to strike a balance between efficiency and accuracy, resulting in a smoother onboarding experience for customers.”

Some of the other risks pointed out were approving loans that do not meet the lenders’ underwriting standards, missing personalisation opportunities, incomplete loan documentation, risk of non-compliance, accountability concerns, bias in AI and many more.

The need for lending firms to balance their automation with human oversight was emphasised by Dua. Rather than letting automation completely takeover, a hybrid approach that combines automation and human support is the best implementation of the technology that can maximise efficiency and reduce risk.

The power of AI

It is unsurprising that AI will play an important role in the digitalisation for lenders, but it will also have an important role in their automation efforts, namely through chatbots.

Chatbots have become a common sight and many companies leverage the technology to offer customers a quick solution to answer some basic questions. These are often basic tools that can respond to simple queries, point customers to relevant FAQs or connect to specific support staff. However, the rapid development of AI is helping to transform chatbots into powerful tools that can provide valuable services to customers.

The world has been captivated by the rapid development of generative AI tools. Chat GPT has gained a lot of attention, but there are countless tools in the market or in development. Interest in the market is creating a huge potential for growth, with Bloomberg’s research suggesting generative AI could have a market size of $1.3trn by 2032. This represents a colossal CAGR of 42% as it rises from a market size of $42bn in 2022.

It is a safe bet that most industries will seek to use generative AI in some form. Within lending, transforming the chatbot to offer improved support to customers could be a potential, but it is still in the early stages and hard to know where the technology will expand to. One thing that is certain, firms will need to experiment to see how it could transform their operations.

Rodrigues said, “Generative AI is disrupting the lending space with its ability to provide hyper-personalization at scale, through data-driven insights and machine learning. Having the expertise and resources to do this will act as a key differentiator moving forward, and the lenders that can provide these solutions will become the favourable lenders for the next generation.”

On top of this, the technology can help automate data extraction processes. Rather than teams needing to sift through emails, chats and social media, the technology can collate this information and generate insights for staff or provide personalised offerings directly to the customer, she added. “By understanding the customer’s previous interactions, bots can provide tailored solutions and recommendations, replicating a human-like understanding of the customer’s needs.”

The AI could also analyse data and behaviour from a customer or business and split them into groups based on their propensity-to-pay and connect percentage, making a prediction on the probability of them making a debt payment. With this information, it could build a personalised repayment plan and recommend channels for engaging with them. This can improve overall collection rates, while lowering cost of collections.

Williams acknowledged the potential gains from AI within chatbots but also noted there are risks, such as a customer asking a question that confuses the technology, leading to inaccurate answers or errors.

“In the short/medium term, mapping out your chat flow correctly addressing quick fixes will, over time, significantly reduce customers needing to engage with an agent, whilst having someone there swiftly should they need them. There are other potential uses of tech within the lending space that I am hopeful about – however, AI needs to be regulated before that happens. The possibilities are exciting.”

Aside from AI, there are several other types of technology that are helping to transform the customer experience within lending. Shah pointed to open banking and open accounting that allows lenders to connect directly with an applicant’s bank account or accounting software to automatically get the financial information they need to make instant decisions on loan eligibility, amount and terms.

Keep up with all the latest FinTech news here.

Copyright © 2023 FinTech Global

Enjoying the stories?

Subscribe to our daily FinTech newsletter and get the latest industry news & research


The following investor(s) were tagged in this article.