AI proving to be a strong force behind disbursals via SBI YONO
SBI YONO disbursed over five thousand five hundred crore of retail credit in just one quarter as compared to three thousand crore in the same quarter last year. Secondly in the year 2019-20, the bank disbursed two billion in the form of pre approved personal loans (PAPL), which is directly a function of efficient use of analytics and AI
The importance attributed to customer experience and data analytics by India’s largest PSU bank, SBI is proved by the statement from Dinesh Kumar Khara, when he said in his first media interaction after taking over as chairman, that data analytics empowered customer experience will be one of his top priorities at the bank. It speaks volumes about the weightage given by the bank to CX and analytics. It’s important to note, data analytics forms the base for building a robust Artificial Intelligence (AI) platform.
SBI YONO disbursed over Rs five thousand five crore of retail credit in just one quarter as compared to three thousand crore in the same quarter last year. Secondly in the year 2019-20, the bank disbursed Rs two billion in the form of pre approved personal loans (PAPL), which is directly a function of efficient use of analytics and AI. The role of analytics is to segment the customers in age groups, decide on the eligibility and then AI consumes that data and converts it into an opportunity on the YONO or other platforms in the form of right offers, based on algorithms. The idea is to make a best fit proposition for the customers.
In PAPL, a Response Surface Methodology (RSM) model is used. It’s a statistical technique (with correlation to AI), which helps in reducing the workload of the branches and also cuts the risk factor. It has resulted in loan buckets which are low on NPA and issued with relatively low branch related activities. More on a phygital model.
SBI runs the country’s largest data analytics engine. It sees the customer not in parts but the sum of the parts, i.e the customer’s behaviour across digital channels in the public domain. The data is built and then AI algorithms are built to pop notifications to the customers.
The bank is looking at using AI in a holistic manner. It’s not just restricted for the retail segment but also the corporate and SME segment. SBI will soon launch video KYC which will also have an AI element. AI is highly compute intensive and the processing is all done on cloud.
Among other applications, SBI has partnered with a startup company for an AI application that runs on the feeds from the CCTV cameras deployed in branches to do sentiment analysis of the customers. This is a kind of an active AI usage, where near to real time analysis is done. In order to take the ethical considerations, the bank has issued a disclaimer to the customers that the usage is for training and quality purposes only.
The #Gharsebanking campaign was recently launched in the wake of the coronavirus crisis to encourage people to download the SBI YONO app and also using facebook messenger which is powered with AI enabled conversational banking.
My biggest learning on using AI is to get the house in order as far as data purification is concerned. It took us some time to clean up the data at our end. SBI is a 211 year old bank with a humongous amount of legacy data.
