Kathy Gibson reports from AfricaCom in Cape Town – Artificial intelligence (AI) and alternative data sources could transform financial inclusion in Africa.
This is according to Dr Richard van der Wath, chief data officer of MyBucks, who describes alternative data (AD) as just about anything other than the structured data typically found in the IT systems.
AD could be big data or unstructured data, from a variety of sources.
As an example, he asks what data, apart from credit repayment history, can be used to evaluate and predict risk in Africa.
AD sources could include data patterns and social media, mobile money transactions, Web browsing history, behavioural data and psychometric data.
“The key is that many of these data can be obtained from a mobile phone,” Van der Wath adds. “This is important because it needs to be relevant for Africa.”
A huge percentage of Africans are not served by traditional financial services – but in the next couple of years, a further 200-million people on the continent will gain mobile access.
“That means the AD sources will grow; there is a huge market already and it will grow further.”
Gaining actionable business insights from data is where artificial intelligence (AI) comes into play.
Van der Wath describes AI as an advanced form of data analytics using computer models and algorithms that can find patterns in historical data and build a model from those to help predict outcomes when similar patterns are observed in new data.
AI is also capable of machine-learning, where it is able to learn as it goes along.
“It is my view that data and advanced algorithms is the basis of fintech,” Van der Wath says. “It enables smart predictions in near realtime; and helps to understand customers better so the product offering can be better customised.
“And it helps with risk management.”
MyBucks realised that a lot of data was available that was not being used to predict credit risk. It came up with an algorithm that takes any two attributes of loan applicants and uses these to determine the risk. Importantly, the system will learn as it goes along.
The other widely-used function is fraud detection, which is unsupervised. “We are an online company so we don’t see our clients. The trick is to know who is on the other end.”
MyBucks launched its Haraka app in Kenya a couple of months ago. It provides quick instant loans to users typically ignored by traditional banks. It is quick and simple, and can be used anywhere.
Using AD and AI, Haraka is able to perform its risk management.
A customer could apply using a smart phone. The system would verify their ID using social media and build a transaction history from the phone itself. The user is asked why they want the loan as well.
Haraka doesn’t do a credit check, as these are expensive. It uses its own AI algorithms to determine the risk prediction and either grants or declines the loan.
“Fintech is driven by AI and AD – and it is already supporting economic growth in Africa,” Van der Wath concludes.