How South Africa’s leading Online Creditor utilizes Machine Learning to stay ahead in the cut-throat Fintech Space

Enhanced by the substantial volumes of data available in today's digital world, machine learning offers the potential to completely revolutionize the service offerings of financial institutions. This shouldn't be news to anyone reading this post.

However we're quick to forget that until very recently, this technology was accessible only to tech giants and corporations. In the last decade, access to this power has become increasingly available to financial institutions of all sizes, including community banks and credit unions. Machine learning and artificial intelligence applications have already made permanent impacts on the banking industry.

Machine learning leverages extensive data analytics to create dynamic predictive models. In essence, when the right data is fed into computers equipped with the appropriate algorithms, they possess the ability to autonomously learn and iteratively improve over time. This "learning" aspect of machine learning enables computers to uncover hidden insights without explicit programming directives. Of course, the secret success lies in what data is processed and how the algorithms work.

The use of exceptionally large data pools enables computers to identify anomalies and patterns within the information, leading to quicker and more precise data-driven decision-making. Significant advancements in machine learning have made the process of data analysis more efficient, reliable, and cost-effective. This technology is fostering innovation across various sectors, with the banking industry at the forefront, as it's really the perfect environment for machine learning to flourish.

Today, we have the pleasure of speaking with Michael Lazarus, the Chief Technology Officer of personal loan provider Wonga South Africa. As a senior leader in the business, he is responsible for the development and evolution of Wonga's core platforms and architectures, emphasizing the pivotal role of technology in the highly competitive credit industry.

He explains to us that, traditionally, predictive modelling techniques have been mainly used in credit businesses for risk assessment, however Today's reality is different—having access to so much data allows Wonga to employ predictive models to optimize various aspects of the business, that is to say Machine learning's potential extends much further.

Here are four ways Wonga has utilised machine learning applications to enhance their customer's experience and brand satisfaction.

One: Improved Marketing and Consumer Engagement:

Machine learning provides marketers with unparalleled insights into both existing and potential customers. This enables granular audience segmentation and affords opportunities for more effective application of marketing budget across different platforms. E.g. we can predict with more accuracy than ever before whether a targeted email campaign will perform better for a specific audience cohort than, say, reaching them through Facebook or Paid Search campaigns.

Two: Predicting Customer Challenges

One can often predict the reason for a customer's phone call and pre-emptively offer solutions. We can ensure a customer receives adequate support through our online knowledgebase before they need to escalate the level of support required. The processed data we have access to gives our support staff strong predicative indicators about a customer early on in the customer service funnel which can reduce call routing errors, wait times, and improves customer satisfaction.

Improving Risk Analysis and the Collections Process:

It's such an integral part of digital lending that it must be mentioned again. The ability to accurately predict which borrowers can borrow comfortably within their means is fundamental to the success of any personal loan provider.

Of course even the best predicative analysis cannot account for the chaos of real life, which is why helping customers struggling with the collections process is an equally essential branch to providing an optimised consumer experience. This is where Wonga's stellar customer support service launches into action.

Fraud detection and customer protection:

Online loan providers are a prime target for all manner of scamming activity. Modern fraud detection systems can harness machine learning to detect anomalies in online activities and behaviour, often providing real time protection and risk minimisation. Wonga uses a number of additional safeguards to keep customer's data secure.