AI in India | Chapter 6: How AI is Remodelling Indian Financial Ecosystem
Artificial Intelligence (AI) is no longer an emerging technology segment. There’s an entire suite of leading-edge big data technologies which have brought on the Fourth Industrial Revolution, and inevitably enough, AI has seen proliferation like no other to stake its claim at the forefront. According to a report by MarketsandMarkets, the global artificial intelligence market size is to grow USD 309.6 Billion by 2026 from USD 58.3 Billion in 2021, at a Compound Annual Growth Rate (CAGR) of staggering 39.7 percent.
1. Introduction
Artificial Intelligence (AI) has become so relevant and integral to our ecosystem that there is an AI niche in every possible market, hence marking its presence in every industry possible. It is transforming the strategic as well as the operational landscape of businesses in various industries. AI models combine information (or data points) from various sources, analyze the data, and deliver data-driven insights — all in real-time. From big conglomerates to startups to government agencies, and from Agriculture to Manufacturing to BSFI — data science functions such as Artificial Intelligence and Machine Learning are slowly changing the way business is done.
In 2016, the computer program AlphaGo captured the world’s attention when it defeated 18-time world champion, the legendary, Go player Lee Sedol. The ancient board game of Go is one of the most complex board games ever devised, requiring strategic thinking, creativity, intuition, and imagination — abilities long considered distinctly human. Since then, technologies driven by Big Data such as AL and ML have developed even further, and their revolutionizing influence is increasingly evident across industries.
The Banking, financial services, and insurance (BFSI) industry is one of the early adopters to embrace the potential of Big Data revolutions and the wave of new technology that has come with it. As financial institutions continue to embark upon the next generation of technology, they are looking to streamline and optimize processes ranging from credit decisions to quantitative trading, assist customers with personalized financial services, as well as to have a high level of security. Artificial Intelligence has applications across the spectrum of these functions and enables these financial institutions to deliver optimal experiences for their tech-savvy customers.
While innovation in the finance industry is not new, the focus on technological innovations and their pace have increased significantly. Fintech solutions that use Big Data analytics, Artificial Intelligence, and Blockchain technologies are currently being introduced in the market at an unprecedented rate. These new technologies are changing the way the financial industry functions, therefore creating new opportunities for financial institutions to offer more inclusive access to financial services. In successfully unlocking these capabilities, McKinsey, in a report, estimated that AI technologies could potentially deliver up to USD 1 Trillion of additional value to the world of banking and the financial industry each year.
In this article, we’re setting out to explore:
- The current state of BFSI in the Indian Financial ecosystem,
- Impact of AI on Indian Financial landscape and its potential use cases,
- Major players to leverage AI capabilities in the BFSI industry, and
- Challenges for AI adoption in the Indian Financial System.
2. Overview of the Artificial Intelligence Market in India
Being the world’s fastest-growing business hub with the second largest population, India’s Artificial Intelligence moment is truly here and now. The Indian AI market is entering a new phase where the narrative is shifting from asking whether AI is viable to declaring that AI is now a requirement for most organizations trying to compete globally. The Indian AI market is swiftly adapting to this shifting narrative and driving vital R&D initiatives of AI-integrated technologies across several industries and segments.
With this change in business mindset, the revenue from the AI industry in India has been growing at a healthy annual rate over the last few years. According to The Hitchhikers Guide to Artificial Intelligence 2019–20, the revenue generated in 2019 by India’s Artificial Intelligence industry grew by 80% compared to last year, adding a whopping USD 415 Million in revenues alone.
As the COVID-19 pandemic has accelerated digital uptake and helped bridge the adoption gap, these numbers are expected to grow even further in the coming years. According to a report “AI: An opportunity amidst a crisis” published by PWC, India has recorded the highest increase in AI adoption amidst the pandemic compared to other major economies like the United States, UK, and Japan. Due to the sudden surge in the number of consumers shifting to digital platforms for their day to day needs, more businesses started embracing digital transformation and generating a considerable amount of data through various touchpoints from a customer’s digital journey. This change will assist the AI industry in rapid execution.
Recognizing AI’s potential to transform economies in the foreseeable future and the need for India to strategies approach, NITI Aayog, the policy think tank of India, has adopted a three-pronged approach — undertaking exploratory proof-of-concept AI projects in various areas, crafting a national strategy for building a vibrant AI ecosystem in India and collaborating with various experts and stakeholders. Thus, boosting the state of artificial intelligence technologies in the Indian ecosystem.
As private investments pour in and the government unlocked the AI’s real potential by providing proper governance, more and more organizations started leveraging AI to enhance their existing processes to yield better outputs.
As per an International Data Corporation (IDC)’s latest report on AI, India’s Artificial Intelligence (AI) is expected to grow at a CAGR of 20.2 percent to reach USD 7.8 Billion by 2025.
3. The State of BFSI in India
The traditionally cash-driven Indian economy has responded well to the fintech opportunity, primarily triggered by a surge in e-commerce, smartphone penetration, and affordable data cost. Consequently, India is now among the fastest-growing fintech markets globally, with a market size of USD 2.4 Billion.
As the ongoing COVID-19 pandemic continues to drive changes in business as well as end consumers behaviors; stakeholders across the payment ecosystem such as banks, merchants, and intermediaries are responding quickly, prioritising digital shift to maintain current revenue streams and also looking out for new ones through an entirely digitised customer experience. Therefore, AI-powered technologies are bound to alter how banking, insurance, and other financial services work.
According to an article published in The Week, India processed a total of USD 25.5 Billion in real-time payments compared to China’s USD 15.7 Billion. In addition, with a digital payment rate expected to grow by CAGR 20 percent, India is ranked the highest globally, with China, in the fintech adoption rate.
The BFSI sector is now leveraging AI across various functions such as enriching tailored customer experience & satisfaction, customer relationship and retention, financial risk assessment, push smart notification, and detecting fraud & money laundering transactions. Additionally, the BFSI sector’s AI revenue distribution in the Indian market is among the top three sectors, about 9.6 percent, equivalent to USD 615.3 Million in market value.
The key players in this sector contributing the maximum AI-backed revenues are the public and private sector banks such as HDFC Bank, Axis Bank, State Bank of India, ICICI Bank etc.; non-banking financial services companies such as Aditya Birla Finance Limited, Bajaj Finance, etc.; and many new emerging fintech startups such as Capital Float, Coverfox, Creditmate, Flexiloans, Shubh Loans, among others. These corporations are using AI predominately in:
- Enhancing customer experience and satisfaction by leveraging robotics and Natural Language Processing (NLP)
- Portfolio optimization for institutional and retail investors by adopting algorithms in investments research.
- Financial management services to deliver personalized financial management advice and alerts to customers in real-time.
- Smart transaction analysis programs and tools to detect fraud and money-laundering transactions.
These examples are just a few instances in which AI delivers innovation across the areas of investments, financial advice, and overall BFSI operations. Now, let’s look at how AI is becoming an integral part of the BFSI industry.
4. Why Must the BFSI Industry Become AI-first?
Over several decades, financial institutions such as banks, insurance providers, credit unions etc., have regularly adapted the latest technology innovations to reform how customers interact with them. Banks introduced ATMs in the late 1980s and electronic, card-based payments in the ’90s. The 2000s saw broad adoption of 24/7 online banking, followed by the spread of mobile-based “banking on the go” in the last decade. With the coming of age technologies, these financial institutions can improve their ability to achieve four key outcomes: higher profits, at-scale personalization, distinctive omnichannel experiences, and rapid innovation cycles. Institutions that fail to make AI central to their core strategy and operations — what we call becoming “AI-first” — will risk being overtaken by competition and deserted by their customers.
To meet customers’ rising expectations and beat competitive threats in the AI-powered digital era, the AI-first financial institutions are expanding their use of AI technologies to ensure an omnichannel seamless customer experience and improve their backend process. Thus, driving an evident shift from financial inclusion to financial empowerment.
Here are the potential use cases of AI transforming the BFSI industry:
4.1. Personalized Marketing and Support
4.1.1 Relevant Customer Engagement through Chatbots & Voicebots
The first generation of chatbots for the BFSI industry proved the hypothesis of customer convenience with reduced costs with 24*7 service without any delays. With the advent of conversational AI and NLP, AI-based chat and voice bots are enhancing the customer experience with hyper-personalization, mimicking the branch experience with the ability to converse in customer’s preferred language, providing familiarity and comfort with complicated banking services.
Financial institutions have a potential cross-sell and up-sell opportunity with a large segment of the banking population. Voice-enabled solutions eliminate the barriers faced by end consumers posed by device knowledge, texting capabilities, and, more importantly, the language barrier. Today, voice bots converse and transact in regional dialects and language, empowering millions of people left between the urban-rural divide. Powered by conversational AI platforms, banks can deploy bots that support over 100 languages and dialects, enabled with sentiment analysis and smart analytics. Voice-enabled banking virtual assistants can handle payments, transfers, credit card activation, password resets, and pay alerts and reminders for customers from anywhere, freeing-up customer-service teams to focus on more complex customer enquiries and enhance productivity.
As per Gartner’s AI and ML Development Strategies Study, by 2023, 80% of consumer apps will be developed with a “voice-first” philosophy. Financial institutions are revamping their digital strategy and embracing conversation-led banking to deepen customer engagement and differentiate services.
4.1.2. Smart Wallets
Another area of interest for the banks is Smart Wallets, where they will provide AI-enabled digital wallets to their customers. These wallets will analyse customers’ spending habits and learn from their behavior to provide competent advice and recommendations for future spending. It will encourage savings and responsible spending on their credit & debit cards in the form of predictive alerts and recommendations. Additionally, AI can also detect if a customer is likely to switch their products or services based on their activities. This early signal will help banks offer them more suitable products, which may help retain the customer.
4.1.3. Advisor for Financial Products
Online platforms backed with an AI tech stack require zero to minimal intervention in offering financial advice, re-investing dividends, automatic portfolio creation/re-balancing, etc., for its customers. Consequently, financial institutions are constantly pressured to reduce their commission rate on individual investments as machines/algorithms can do the same work for a single down payment.
4.1.4. Personalized Financial Services
AI-powered platforms monitor customers’ goals, give them personalized suggestions, and advise which stocks or bonds to buy/sell, irrespective of their risk appetite. These intelligent systems track income, essential recurring expenses, and spending habits and come up with an optimized plan and financial tips. In the current marketplace, multiple mobile banking apps help its customers plan their expenses, push customised reminder alerts to pay bills on time, streamline the way they interact with their bank, from getting information to completing transactions.
4.1.5. Portfolio Optimization
Artificial Intelligence is fundamentally transforming asset allocation, trading processes, risk management and other areas of portfolio management. In fact, many Robo advisors already use AI-enabled technologies to deliver portfolios with better out-of-sample performance for investors by conducting a textual analysis of annual reports, economic reports and other meaningful information. AI is now used to identify hidden correlations between asset classes and then pinpoint stocks that could outperform or underperform based on those correlations.
4.2. Automated Business Process Management
4.2.1. AI-enabled Trading
AI-powered application has a wide variety of use cases in high-frequency trading, where inputs are taken from multiple financial markets to make an investment decision in real-time. According to a report by MarketandMarket, the global algorithmic trading market size is projected to reach USD 18.1 Billion by 2024, at a CAGR of 11.1 percent during the forecast period of 2019 to 2024.
Financial institutions have been relying on data scientists and algorithms to predict future patterns and outcomes in the market. As a domain, trading and investments are heavily dependent on the accuracy of these predictions. AI-enabled technologies can do this perfectly as they can crunch a vast amount of data in a short while. These AI-driven models can analyse historical data to find financial patterns and predict how & when such patterns might repeat in the future. While outliers such as the great economic depression of 2008 do exist in the historical data, AI-based cognitive systems can be trained to study the data to find triggers for such outliers and plan for them in future predictions as well. Depending on individual risk appetite, AI can suggest personalized portfolio solutions to meet each customer’s targets.
4.2.2. Robotic Process Automation
Financial institutions, especially banks, are using AI-powered machines to take over high volume, back offices processes and time-consuming repetitive tasks. Artificial intelligence-enabled software can verify data and generate reports according to the given parameters, reviews documents, and extract information from forms (applications, agreements, etc.) at a scale. Thus, saving time by increasing accuracy, reducing operational cost, and driving efficiencies.
4.2.3. Data-backed Lending Decisions
Historically, credit and loan applications were processed by humans, introducing the risk of intentional and unintentional bias. As lending becomes increasingly digital, financial institutions are looking to AI-powered technologies to help make credit decisions based on historical data. Beyond just analyzing data, artificial intelligence and machine-learning algorithms are deployed to make faster, more efficient credit decisions. By incorporating alternative data, eliminating human bias, and training the models in real-time, AI-enabled algorithms are able to predict a consumers’ creditworthiness more accurately, regardless of factors like race, gender, religion, or sexual orientation.
4.2.4. Talent Acquisition
AI is being deployed to save hiring manager’s time in various talent acquisition processes. For example, an AI-based tech stack can analyse previous pipeline data for a specific role and predict the right channels, keywords, and personas for that role. These systems can also be used for shortlisting resumes from social media sites, pre-screen candidates over chat, determine candidate dropout chances, etc.
4.2.5. Investment Research
Synthesizing information from multiple data sources and building proprietary quantitative models takes enormous human effort and time. With the expansion of AI applications and ever-increasing data sources, investment research and analysis methodologies are evolving rapidly. AI tools enable large-scale data processing at a rapid rate and integrate traditional data sources with new ones such as web search trends, online traffic, and social media engagement. Portfolio managers and analysts save time and uncover hidden signals by quickly processing an enormous amount of financial statements, call transcripts, press releases, investor presentations, blogs, news articles, and sell-side reports for investment research — contributing to improvements in forecasting, investment decision-making, and idea generation.
4.3. Security & Compliance
4.3.1. Fraud Detection Through Machine Learning and Pattern Recognition at Scale
Traditional businesses relied on human resources to detect and block fraudulent payments. But these operations often generated false positives, had limited outcomes, were inefficient and very hard to scale. With a massive uptake on eCommerce activity and online transactions recently, the utility of AI-based systems in battling financial fraud is critical. These fraud detection systems analyze customers’ behavior, buying habits, location and trigger a security mechanism when something contradicts the established spending pattern — all in real-time.
4.3.2. Biometric Identification Through Speech and Image Recognition
Traditionally, banks have used knowledge-based authentication methods like one-time passwords (OTPs), PINs, and passwords. These authentication methods are have been proven effective when done right but depend hugely on customers’ vigilance. It requires customers to put in extra effort to choose strong login credentials, preferably unique for each account, and update them frequently. It creates a sense of frustration among customers and leaves them with a poor online banking experience, especially when they have to often come up with passwords that meet the security criteria. Consequently, the demand for contactless and seamless technologies has permeated even to banking security, especially after the COVID-19 pandemic’s impact. Voice and Image biometrics in banking is a promising online banking solution, above and beyond the retina and fingerprint scanners that require people’s physical presence. Hence, reducing the chances of fraud and improving customer experience.
4.3.3. Compliance Monitoring
Although compliance is a business function relatively untouched by digital technology, AI’s computing power to process vast amounts of data with speed and accuracy can transform regulatory compliance. AI and ML can be used to capture, extract and analyze several key data elements to flag potential issues and help in reducing costs in today’s data-driven compliance environment.
5. Roadblocks in AI Adoption in Indian Financial Ecosystem
5.1. Model Interpretability
AI-based neural networks and other machine learning models are quite complex, making them more challenging to understand and explain than traditional ones. This intrinsically leads to some level of risk and requires an increased level of governance. Financial institutions must be able to explain their models and the rationale behind them — in-depth — to ensure a smooth transition, remain compliant, and minimize the risk of making bad business decisions due to these advanced technologies.
5.2. Lack of Credible and Quality Data
AI and ML models collect data points from various sources, analyze the data, and then deliver data-driven insights — all in real-time. These models automatically become more efficient through this process and, therefore, result in greater accuracy and predictability over time. But the lack of credible data sources and good quality data can negatively impact these models’ ability to enhance decision-making and provide a better service for all its players.
5.3. Legacy IT Infrastructure and Core Technology
The IT infrastructure of financial institutions such as government banks, credit and loan associations, etc., needs a rehaul as financial organisations’ data centre technologies do not perform well on fast data / big data-type technological tools designed for large-scale standard systems. Their IT teams are generally struggling with inflexible tech infrastructure. On top of it, the complex process involved in AI such as data ingestion, analysis, transformation and validation, model development, validation and monitoring, and logging and training, among others — lead to delays in AI adoption at scale, in turn increasing the time between implementation and reliable ROI.
5.4. Unskilled Existing Workforce
Interestingly, the investment cost and convincing top management to deploy AI seem to be of lesser importance as it can add immense business value. But being a new technology, there’s still a shortage of available skillsets in the market, with most respondents facing an issue in finding the quality talent pool with experience with cognitive systems.
5.5. Narrow Focus & Use Cases
By virtue of their design, cognitive systems work best at solving a specific problem and cannot deviate from what they were built for. An algorithm designed to detect fraudulent payment would not necessarily help detect any other suspicious activity related to trading. In most use cases, separate algorithms are required to solve different tasks. This, in turn, creates a challenge for financial institutions to create a more general AI which can tackle more than a specific genre of use cases.
5.6. Multiple Stakeholders
Since cognitive systems are trained to quickly make multiple decisions at scale in real-time, there is a lack of accountability for decisions made by these AI models. When it comes to larger financial organizations embarking on the journey of AI implementation, there is no lack of leaders ready to involve themselves in the project. In fact, there are too many leaders with too many opinions, which often turns into a bigger problem than a lack of leadership. These leaders have overlapping responsibilities, different long-term AI visions, and work areas, which means that there is often no single owner of the AI projects. This can lead to uneven adoption of technologies, which can drive up the costs over time and may conflict with an overall AI adoption strategy.
5.7. Regulatory Hurdles
In order to gain acceptance from the customers, regulations and compliance are required to ensure financial services are transparent and what kind of user sensitive data is being fed to AI-based cognitive systems. However, an increase in regulations may threaten the Indian financial ecosystem by stifling innovation and raising operating costs. Therefore, the regulators have a critical responsibility of creating an innovation-friendly environment while adhering to customer protection, data security, and privacy concerns.
6. Uses Cases of AI in Action
Prominent players of Indian origin are leveraging AI-powered platforms in conceptualizing, designing, and executing end-to-end digitalized financial services. With the advancement in technology and its adoption, these Ai-backed cognitive systems can now manage functions like financial risk assessment, compliance monitoring, push smart notification to customers, portfolio enhancement and detecting fraud & money laundering transactions.
The organizations leveraging AI-empowered solutions across the BFSI sector include:
6.1. Razorpay Software Private Limited
6.1.1. Fraud Prevention
Razorpay uses ThirdWatch, to reduce fraudulent transactions owing to the AI-driven fraud analytics solution. It collects and analyses hundreds of parameters like the customer’s email address, shipping address, pin code, and how much time the customer took while selecting the product, among others. This AI solution compares these pieces of information to normal shopper behaviour data points. If something looks like an outlier, it immediately flags customer and seller for a possibly fraudulent transaction.
6.1.2. Determine creditworthiness
Moreover, Razorpay also utilises AI-based models to determine creditworthiness and help businesses avail loans. The standard credit bureau data is inadequate to provide a holistic picture of SMEs’ (SMEs) creditworthiness. As a result, most SMEs fail to get a loan from traditional financial institutions. Razorpay uses machine learning models to collect and analyze transactional data generated on its platform by such SMEs and predicts the business cash flow for coming years. Based on these estimates, Razorpay offers credit or loan facilities to SMEs.
6.2. HDFC Bank — Electronic Virtual Assistant [Eva]
HDFC Bank’s EVA has become India’s largest banking chatbot. EVA is a chatbot created by senseforth.ai using conversational AI and assist users in accessing the correct information by decluttering the access information usually clouds banking websites & navigating them to the right page. Deployed in March 2017, EVA successfully addressed over 2.7 million customer queries within six months of its deployment. It is available on all digital platforms of the bank, including the website, mobile site and a dedicated customer portal.
6.3. Mswipe Technologies Pvt Ltd
To make merchant’s onboarding simple and real quick, Mswipe uses an AI-driven F2A2 solution (developed by Signzy). It records merchant profile information and KYC papers and instantly authenticates them using over 40 government databases, making it one of the unique solutions in this space.
6.4. ICICI Lombard
ICICI Lombard leverages AI-based robotic process automation to automate backend mundane, time-consuming tasks — making it human error-free and more efficient. As transaction volumes increased, these automation processes have helped reduce time to respond to user management requests, improved customer satisfaction index, and increased cross-sale.
6.5. Cognext Analytics Private Limited
Cognext is one of the fastest-growing fintech startups in India, based out of Mumbai, Maharastra. It has developed an AI and ML-based technology platform, Platform X, to provide configurable, scalable, and cost-effective solutions for regulatory compliance. It assists financial institutions like banks, non-banking finance firms, and neo banks to remain in control of risk, finance and regulatory compliance, establish integrity & transparency of results, and adapt & respond quickly to regulatory changes through a fully declarative configurable and scalable technology framework.
6.6. HDFC Life Insurance
India’s leading private sector life insurance company, HDFC Life, launched an insurance Email Bot in 2017 named SPOK. SPOK uses AI and NLP methodologies to mimic human cognitive abilities in reading, comprehending, interpreting, and conversing. Besides answering queries quickly, it also helps generate more profound insights into customer needs by identifying patterns in their email interactions.
6.7. PolicyBazaar
PolicyBazaar, India’s one of the most prominent online life insurance and general insurance aggregator, has launched an AI-powered Chatbot named PBee, based on Google’s DialogFlow, on its platform to ensure smooth customer engagement and sell insurance online. Due to its launch on PolicyBazaar’s platform, the per-call talk time has been reduced by 50 percent, leading to increased productivity levels of callers to 200 percent. At the same time, the conversion rate of their callers has gone up by 25 percent due to a smoother chat customer experience.
6.8. Capital Float
Capital Float is a leading fintech startup based out of Bengaluru, Karnataka, that provides working capital finance to SMEs. It extensively deploys AI and ML models to determine credit risk across its product, such as loan management system (LMS), loan origination system(LOS), and decision engine (DE). Its small-ticket size loans are entirely automated and are built using AI models. Moreover, Capital Float’s personal finance management function is algorithmically designed to generate financial insights on its users, which helps their marketing team to deliver personalized credit proposals to customers proactively.
7. What’s next?
New-age technologies like Artificial Intelligence, Machine Learning, Deep Learning, Natural language processing etc., have already started disrupting the BFSI industry. Early adopters are gaining a competitive advantage and widening the gap with those still doubting AI capabilities. The need for technological evolution in the industry has acted as a catalyst for Indian financial institutions to seek and deploy next-generation technologies. These technologies have empowered the BFSI sector to build a delightful digital user journey across the customer lifecycle by ensuring personalized customer engagement, improving security & compliance, and automating back-end processes at scale.
For many Indian financial institutions, ensuring the adoption of AI technologies isn’t a choice anymore but a strategic imperative. However, it would be shortsighted to consider AI adoption as a mere technology upgrade exercise over the existing IT infrastructure. Financial organizations have to leverage AI to reimagine the business value chain through the right human-machine interplay. And the journey to becoming AI-first for these organizations starts with evaluating how their short term and long term strategic goals (e.g., growth, customer engagement, customer retention, profitability, innovation) can be materially enabled by the range of AI capabilities.
