According to Okwechime, A.I. can be used for credit-risk assessment, automating a significant portion of the credit-lending under-writing process, enabling lenders to offer loans more rapidly, and even extending services to more customer segments without needing to increase the workforce. He explained that by leveraging data, a Credit Decision Engine is built using sophisticated statistical methods to generate risk ratings for customers based on various data attributes. “Such methods can approve loans without visiting a bank branch or any other access points. This is whilst also being able to manage risks to acceptable levels and provide appropriate credit lending limits tailored to the customer’s circumstances,” he said
Over the years, as digital channels for operating businesses have increased considerably, A.I. has emerged as an essential tool for preventing fraud and financial crime. Okwechime revealed that A.I. has the innate ability to analyse massive amounts of data and uncover fraud trends, which can subsequently be used to detect fraud in real-time. He explained that this is especially useful in identifying fraudulent bank transactions and detecting transactions that do not fit the normal behaviour of specific customers. “These systems can also self-learn, detecting transactions that have similar trends of previously identified fraudulent activities,” Okwechime said. He added that A.I. also plays a vital role in identity fraud prevention, enabling digital channels to verify identity via face matching to government-issued IDs, biometrics, and even voice matching.
Okwechime noted that A.I. is helping marketers with new levels of understanding and lasting commercial advantage, especially in an era where communicating to customers is easier, with several channels for dialogue and customer acquisition. He said, “A.I. uses customer data and profiles to learn how best to communicate with customers, then serve them tailored messages at the right time without intervention from marketing team members, ensuring maximum efficiency.” He added that businesses could also use A.I. to prioritise leads, particularly relevant when companies have hundreds of leads a day but limited resources. “Using data such as drop-off information gathered during customer onboard, A.I. can prioritise leads from the most likely to convert, ensuring efficient use of limited marketing resources, whilst maximising conversion rates.”
Chatbot (Conversational Agents or Dialog Systems)
A new wave of change in how consumers interact with services is on the horizon, driven by the growing demand for rapid responses from customers. According to Okwechime, AI-powered Chatbots leverage Natural Language Processing to understand conversations and their contexts to such an advanced level that appropriate responses can be generated. He explained that recent advances have gone beyond text-based chats, with companies like Apple (Siri), Google, and Amazon (Alexa), introducing voice-controlled virtual assistants, able to answer an infinite number of questions proficiently in natural language dialogue. “In a market like Africa where literacy rate falls heavily behind the world’s average, voice-controlled products and services will revolutionise how a consumer interacts with technology platforms, reducing entry barriers and improving technology adoption,” he said.
Lastly, Okwechime explained that A.I. for operations combines sophisticated methods from deep learning, data-stream processing, and domain knowledge to analyse data gathered from operational pipelines to enhance efficiency, reduce costs, and maintain an optimal delivery and quality of services. Depending on the type of business, this covers several verticals, including but not limited to:
- Using images or videos to:
- Do quality control on products or raw materials.
- Count high volumes of stock.
- Tagg essential information about a product such as visible license numbers, colour, make, and manufacturer.
- Using historical data to estimate prices when purchasing goods or services.
- Using macro-economic factors to forecast value and any pending appreciation or depreciation.
- Predict periods or high demand locations with sufficient notice to prepare accordingly to maximise returns.
- Liquidity management, forecasting future events of liquidity challenges based on predicted or projected growth.