Artificial Intelligence for Financial Inclusion

Next Tide: We need to be ready

The Indian Rural Banking landscape has just seen a major transition, by way of Financial Inclusion under the guidance of RBI. It started with Small savings accounts, Agency banking channel, Aadhaar number, Pradhan Mantri Jan Dhan Yojana, BBPS, BHIM, AEPS payments, UPI etc.

First time in the history of India’s Banking sector, the RBI gave out differentiated licenses by way of 10 Payments bank1 and 11 Small Finance Bank Licenses2 in 2015.

At the other end, urban banking is witnessing another wave of transitions with a number of Fintech players competing to offer banking services by way of Robotics and AI to automate banking by machines.

HDFC’s Ira, the first ‘humanoid’ branch assistant, City Union Bank’s ‘Lakshmi’, and Canara Bank’s ‘Mitra’ are just a few examples of how banks and financial institutions are integrating AI and robotics in their services.

Keeping in view the above two trends and if both these paths merge, sooner or later it may be wise for the Banks to think of ways to take Automation to Rural banking. Furthermore, regulation might also require compliance that can be significantly simplified with automation. banks and the solution providers need to be ready and need to be able to identify what all can be done to employ Artificial Intelligence to enrich Financial Inclusion.

The Journey has just begun!!!

What sources of data do we have, what processes of rural banking can be enriched, who all could be the immediate stake holders, what enhancements would a technical solution require and finally what is the revenue model? These are some of the questions that banks may not yet have ready answers to, since we are still in the starting phase of our AI journey. As I write this thought paper I do see some ready clues which both banks and solution providers can pick and proceed.

Sources of Data: its already there

As on Aug 2017 Over 9.3 crore3 PAN cards linked with Aadhaar numbers.
Recent initiatives of Govt. to link Pan card with all bank accounts, and then to link Aadhaar to PAN Card and mobile number, sets the right path for the required source of data. This can be a good starting point for our data scientists in the AI Journey. Using the Aadhaar linked Database, the government is able to track the number of SIMS a person may have, the number of gas connections, the amount and volume of transactions done on his/her accounts. It facilitates a single Point of tracking and also an easy KYC.

AI Applications: Give it a thought

From the existing banking landscape and the rural customer experience, some of the applications of Artificial Intelligence in Rural Banking that the solution providers and banks can explore are:

• AI to build credit history: In rural India, there is hardly any credit history. Artificial Intelligence can be applied to collect and study data like Aadhaar linked data, crop turnover, handset details, SMS logs, social network data, GPS data, call logs and contact list etc – to identify the Credit worthiness of customers. The system can recommend a smaller value loan and then to top up further based on the renewed credit worthiness re-estimated by the AI machine.
• AI as a Relationship Manager: Most of the bank staff have urban orientation and do not have inclination and patience to talk to the rural customer. We can have NRLP (Natural Regional Language processing) based AI trained robot- to train and talk to the Rural customers in Regional language: explain them about banking products, can also discuss about the amount of the debt that they have and suggest how much do they need to save. AI trained Robots can become their financial advisors.
• AI assisted Lifestyle based banking: There are a number of Govt. schemes that are rolled out like Gram Sadak Yojna, Swachh Bharat Abhiyan, MNREGA etc. The Incentives for most of the schemes go through the Pradhan Mantri Jan Dhan accounts. Banks can use feeds of all such incentive payments data from the UIDAI database into the AI engine and come up with the best possible products the customer can be offered. The offer can also be combined with a discount. The higher can be the discount with the raising level of financial health of the customer.

Major Stakeholders: What’s in it for me?

• Business Impact for Banks: Banks can think of newer revenue channels and find ways to encash AI enabled features. The AI enabled credit rating feature can be offered as a monetized API to other enterprise in similar businesses like MFI, NBFC, Credit Rating agencies etc. In Chile4, for instance, supermarket chains have started writing credit histories for their unbanked clients.
• Enhanced Rural customer financial engagement: AI trained Robots can better engage with rural customers and there by elevate them on the financial and economic ladder. Quoting an existing application in agricultural field: Microsoft & ICRISAT’ s pilot5 cloud based app can predict the best sowing week depending on weather conditions, soil and other indicators. This has benefited the farmers with better crop yields. Similarly, Banks can provide AI enabled guidance on possible product harvest returns that the farmer may get based on image analysis of the Crop. The same can be then linked to the customers earning capacity which will in turn be linked to a number of other banking related decisions.
• Government Scheme Effectiveness: With the suggested approach of AI Engine to work in conjunction with the UIDAI database, there are a number of ways solution providers can estimate the effectiveness of various Government schemes linked to Aadhaar. For instance: the latest image analysis technology can be adapted to study the effectiveness of Swachh Bharat Abhiyan and accordingly map with the incentive announced by the Government through the Jan Dhan Accounts.
• Solution Provider Dimension: Technology solution providers are already there in the AI journey by way of AI assisted Risk assessment, fraud detection, better customer services and sooner or later the divide between rural and urban banking solution will be minimized. To be specific, in the field of smart agriculture, there are drones and other IOT devices6 which collect on field information about the crops or the MSMEs. The technology solution providers will have to be ready with the built in interfaces with these sensors, drones etc. to bring data into the Big data space which can further act as an input to machine learning algorithms.
• Regulatory Bodies View: Take one step at a time. RBI has also expressed keen interest in the latest technology developments including block chain, although with caution. Firms like IDRBT7 has even gone to the extent of forming a core team to draft a white paper on Digital Currency, Identity management & KYC and Trade Finance. All the ecosystem players like Regulators, Banks, financial solution providers, Fintech players are still in the learning curve – and the much awaited legal clarity will take due time. Hence it is time to experiment cautiously on a smaller scale.

Word of Caution: The below view holds good for both rural and urban banking.

• With the initial excitement, there is also an inherent fear of handing over decision making to AI. Accepting the fact that even a human can’t be 100% right, AI can’t be a pursuit of perfection, it is more to bring in greater quality at scale and speed.
• The intelligent machine learning algorithms comes with a risk to the Line of Business aims. The Three laws of Robotics8 (also known as Asimov’s laws) will definitely hold good, where necessary caution has to be taken to not break the existing law of the land.
• How much is too much? – with IOT devices and Artificial intelligence algorithms working on everything and anything which humans touch or see, banks and the solution providers will have to draw the line on the extent that technology can intrude the privacy of their customers.

Is it worth it?

Considering the small ticket size or value of transactions, banks and technology solution providers might have a tough time testing the economic viability of AI enabled robots in rural areas and villages. So is it really worth it? Yes, and No. Although the transaction values are low the volumes are definitely significant. The initial investment may be high and one may need to wait for the results to come in. However, considering the enthusiastic Indian crowd and the way we welcome new things – the robot will be the center of attraction and can be the overnight hero of the village. As the future unfolds itself, the early risk-takers will have an added advantage as Automation and AI are here to stay for a way long time.

References:

• https://www.rbi.org.in/Scripts/BS_PressReleasedisplay.aspx?prid=34754
• https://www.rbi.org.in/scripts/BS_PressReleaseDisplay.aspx?prid=35010
• http://economictimes.indiatimes.com/news/politics-and-nation/over-9-3-crore-pans-linked-with-aadhaar/articleshow/60045489.cms
• https://rbi.org.in/scripts/PublicationReportDetails.aspx?ID=836
• http://www.icrisat.org/microsoft-and-icrisats-intelligent-cloud-pilot-for-agriculture-in-andhra-pradesh-increase-crop-yield-for-farmers/
• https://www.postscapes.com/smart-agriculture/
• https://www2.deloitte.com/in/en/pages/financial-services/articles/banking-on-the-future.html
• https://en.wikipedia.org/wiki/Three_Laws_of_Robotics