How to improve data-sharing mechanisms in the financial sectors by using AI, Data Analytics? | The Financial Express

How to improve data-sharing mechanisms in the financial sectors by using AI, Data Analytics?

The next phase of digitisation of India’s banking and financial services sector will depend on how the country establishes standardised data management across the spectrum. The proposals in the recently presented Finance Bill lays the roadmap for that journey as well.

artificial intelligence in banking
FM Nirmala Sitharaman urged entrepreneurs to “Make Artificial Intelligence in India and Make Artificial Intelligence work for India.”

By Rahul Singhal

India will strengthen its economy by fully digitising its banking and financial services industry. An efficient use of data and information on customers (both individuals and MSMEs) will prove to be the game changer. In a recent budget speech, the Finance Minister of India, Nirmala Sitharaman, urged entrepreneurs to “Make Artificial Intelligence in India and Make Artificial Intelligence work for India.” She put data-enabled digital banking and financial services infrastructure at the centre of her growth plan. However, to set up such a robust digital framework, the country needs to establish an easy and safe financial data-sharing mechanism.

The financial sector is becoming one of the early adopters of Artificial Intelligence. Financial institutions and Banks are leveraging the benefits of AI for holistic transformation spanning several layers, including operations, customer support, marketing, risk management, and compliance. Based on a detailed understanding of the customer’s past behaviour, AI is enabling a consistent experience, offering relevant services and products beyond banking.

Industry leaders are bringing in technologies such as interactive voice response, pattern recognition, AI-based self-service applications, smart chatbots, biometric fraud detectors, and data mining tools to improve backend processes and personalize customer support. These algorithmic solutions boost revenue generation and promote cost reduction while reaching beyond human accuracy and productivity levels.

Financial institutions and banks can benefit if they have accurate data. It will enable them to have a correct credit risk evaluation and risk-based pricing, better product delivery with a wide range of customer services, and sound fraud protection. In her speech, the minister said that a one-stop solution for reconciliation and updating of identity and address of individuals maintained by various government agencies, regulators and regulated entities will be established using DigiLocker service and Aadhaar as foundational identity. Moreover, the PAN will be used as the common identifier for all digital systems of specified government agencies.

What are the challenges we have today?

The significant challenge with implementing AI in banking and financial institutions is maintaining data quality. AI is a data-driven technology, as data quality affects the prediction power of an algorithm. Lack of adequate and credible data (called Rogue Data) encumbers its appropriate processing. It calls for insightful data management using advanced analytics and end-to-end AI modelling.

Data analytics firm Experian has reported that human mistakes are to blame for more than 60% of incorrect data, and that miscommunication between departments is a factor in around 35% of incorrect records. If different teams are entering related information into separate data silos, it would be hard to keep downstream data warehouses clear of errors even if there is a good data strategy in place. Records can be replicated with unrecognized data like alternative spellings of names and addresses. Data silos that lack strong restrictions can cause dates, account numbers, and other private information to appear in various formats, making it tricky or even impossible to reconcile them automatically.

The same report elaborates that all over the world, incorrect data can cause a company to lose between 15% and 25% of their revenue. Given that the global banking industry’s income is more than US$2.2 trillion, it is estimated that rouge data costs them over US$400 billion. Additionally, erroneous data can lead to particular hazards that are exclusive to the banking industry. Inconsistent information between data storage areas in an organization can result in incorrect or even fraudulent transactions.

The most challenging problem in cleaning up the data is the cleaning of invalid entries and duplicate data. Careful error correction is needed to ensure that no data is lost while improving the consistency of existing valid data. Moreover, all of the metadata corresponding to data correction should be maintained alongside the integrated data itself.

Amidst the rising creditworthiness of millions of people, banks face a daunting challenge in their efforts to successfully implement the use of AI and a standard data source in financial services. To achieve success, they must overcome the obstacles of erroneous data entry, which can lead to unreliable results and costly mistakes.

What needs to be done?

Traditional banks have now recognized the need to implement AI advancements in their vision, planning and how they engage with their customers. They are striving to deliver intelligent services and personalized customer journeys at breakneck speed. Realizing and fulfilling this vision requires a robust and progressive technology stack.

At all times during this undertaking, banks must stay attuned to customer perspectives and how the bank can create personalized value for each customer. In this light, the proposed Data Empowerment and Protection Architecture (DEPA) under India Stack, the country’s open Application Programming Interface (API), will help financial institutions like FinTech start-ups, personal finance management platforms, lending start-ups access high-quality, authentic data without spending a massive amount of money on bank and financial institutions while maintaining data security. The entire mechanism will be based on accessing open networks. The latter will operate on a customer consent method. An example of this would be OCEN – India’s Open Credit Enablement Network, where the flow of credit between borrowers, lenders, and credit distributors is codified under a common set of standards.

Further, automating data cleansing processes through rigorous data governance policies and the adoption of advanced technologies such as machine learning and natural language processing will help institutions remove erroneous data. Eventually, it will enhance the customer experience.

Thankfully, the country is on the right track to implement the digitisation strategy. Through proposed changes, financial institutions, including lenders, are better poised to capture consumer data. But much is needed to scale up with efficiency. A recent McKinsey study, ‘Financial Data unbound: The value of open data of individuals and institutions,’ says the country is likely to add significantly to its growth if it successfully implements the digital strategy.

Increasing access to finance to millions will only mean customers will be able to buy or use a wide range of financial products from banks and financial institutions. Here again, the traditional documentation process of a bank may disqualify many first-time borrowers or consumers. Hence, Aadhar and PAN will be the base of a unified financial data platform.

The period between 2014 and 2022 witnessed the establishment of a robust FinTech ecosystem and UPI-led payment infrastructure. From here on, the underserved and unbanked population of the country will only witness enhanced access to available capital and FinTech services.

(Rahul Singhal, Partner, SN Dhawan & Co. Views are author’s own.)

Get live Share Market updates and latest India News and business news on Financial Express. Download Financial Express App for latest business news.

First published on: 05-02-2023 at 09:24 IST