ID Finance, the digital finance, credit scoring and emerging markets company has developed and introduced a self-learning chatbot for MoneyMan— online lending platform. Since launch at the beginning of July, over a third of customer requests are already being processed automatically.
The chatbot interacts with new customers at the loan application stage and with registered users when they log in to their personal account. The chatbot helps to locate the information required to determine loan eligibility, and provides recommendations of relevant products tailored to the individual’s requirements and financial prudence. General advice on personal budget planning and financial literacy is also offered.
The chatbot works within the NLP (Natural Language Processing) and NLU (Natural Language Understanding) AI frameworks. Information is processed based on statistical matches covering a wide range of frequently asked questions. And the NLU platform enables analysis of messaging flow so the meaning of the information can be sought out in context.
Additional capabilities include finding non-trivial links in dialogue with users and providing relevant answers to questions unrelated to credit and finance. Thanks to the machine learning technology, the number of questions the chatbot is able to answer increases by 20 per cent daily. The average response time is around ten seconds and if a question cannot be answered the message is automatically forwarded to an available client support operator.
“We are constantly working on improving our customer service. Machine learning and AI technologies are allowing us to automate processes and reduce costs, but to also significantly increase customer loyalty,” commented Boris Batine, CEO and Co-Founder, ID Finance.
“Initial results of the chatbot launch have been positive, both in terms of the reception by our customers and the proportion of requests that can be handled automatically. The service is showing enormous potential with over a third of MoneyMan customer requests already being processed automatically. We expect this figure to increase to 70 per by the end of 2017.”