Swedish philosopher Nick Bostrom, in the book Superintelligence said, “Machine learning is the last invention that humanity will ever need to make.” From electronic trading platforms to medical diagnosis, robot control, entertainment, education, health, and commerce, Artificial Intelligence (AI) and digital disruption have touched every field in the 21st century. AI has made its presence felt in all walks of life due to its ability to help the user innovate. It has also enabled users to make faster and more informed decisions with an increased amount of efficiency.
Of late, the banking sector is becoming an active adapter of artificial intelligence—exploring and implementing this technology in new ways. The penetration of artificial intelligence in the banking sector had been unnoticed and sluggish until the advent of the era of internet banking.
One of the first steps was taken in 2015 by Ally Bank (USA)—introducing Ally Assist—a chatbot that could respond to voice and text, make payments on behalf of the customer, give an account summary, monitor savings, spending patterns, and use natural language processing to understand and address customer queries.
Banks all over the world followed up with their best versions of chatbots: Erica to iPAL, Eva and the most famous one—SBI’s SIA. According to Payjo (the start-up which developed SIA), SIA can handle up to ten thousand inquiries per second and is one of the world’s largest deployments of artificial intelligence in consumer-facing banking. In this era of technological revolution, the banking sector has also witnessed a paradigm shift in its approach from brick and mortar branches to digital banks. Banks are increasingly spending on artificial intelligence and ML in data analytics for personalized and faster customer experiences to garner the interests of the tech-savvy and the millennial class.
According to the FinTech Trends India Report by PwC in 2017, the global spending in artificial intelligence has touched $5.1 billion. The IHS Markit’s “Artificial intelligence in Banking” report claims that this cost has grown up to $41.1 billion in 2018, and is expected to reach $300 billion by 2030. This shows that artificial intelligence has reached a stage where it has become affordable and efficient enough for implementation in financial services. The challenge now is in exploring more ways where the powers of artificial intelligence can be harnessed to streamline internal banking processes and improve customer experiences.
Front end operations of artificial intelligence are those that involve direct interaction with the clients. It includes applications and payment interfaces, digital wallets, chatbots, or interactive voice response systems. Back-end operations are more complicated as they involve the systematic processing of large chunks or terabytes of data to provide security to the system, analyse fraudulent transactions, and generate reports, improve compliance. We shall now discuss the future of artificial intelligence in each of these fields.
Customer being the key driver of a service industry, customer service is at the forefront of any business. AI can be used to derive a better understanding of customers spending patterns, which will help banks customize products by adding personalized features. This supplements meaningful customer engagement, building strong relationships, and growth of business for the bank.
For instance, SBI is working on a system ‘Automated Real-Time Customer Emotion Feedback’ (ARTCEF) using AI to study, in real-time, the facial expressions of customers. AI can also be used to offer personalised payment experience—like most suitable EMIs at checkout based on past payment patterns, offering multi-currency cards to customers who frequently travel abroad etc. AI can also help in setting up biometric face recognition ATMs which function without the need of a card, using real-time camera images and at the same time detect and prevent frauds.
Chatbots and interactive voice response systems which utilise Natural Language Processing are increasingly used by banks nowadays to increase the efficiency of services. It reduces the expenditure on human capital, thus resulting in savings for the bank. Customer satisfaction is also enhanced as they can avail the service at the comfort of their homes without having to visit branches —saving their time. Chatbots can be enhanced in future to announce new offers to the customer like loans, or to alert customers if they have any EMI payment due, suggest good discounts based on the tie-ups the bank has with e-commerce sites etc.
Data analytics to predict future outcomes and trends:
Effortless and swift processing of a large amount of data can help banks observe the patterns of customer behaviour, predict future outcomes and help them contact the right customer at the right time with the right product. It can also help in identifying frauds, fraudulent transactions, and simultaneously detect anti-money laundering pattern on a real-time basis.
Machine Learning and Cognition can be used to identify suspicious data patterns and convince banks if the actual source of money is legal or illegal. AI can also study the past consumer behaviour to predict future requirements, which helps banks to up-sell and cross-sell successfully.
AI-based systems help potential investors by analysing their salary and spending patterns. They can also predict market trends and choose the right funds for their portfolio by determining the adequate sum of money they should invest every month in realising their dreams. All this can be done without visiting branches or hiring experts. In the world of ‘Banking at your fingertips’, mutual funds, fixed deposits can be created at home, and the money is redeemed when necessary.
AI can further be leveraged to notify customers instantly for any suspicious transaction beyond their usual patterns.
Improved operations, efficient cost management vs. focus on profitability:
Banks essentially have to make a profit to survive, and today, banks face significant pressure on their margins. Regulators and their persistent focus on transparency make several businesses unprofitable.
AI technologies enable banks to bring more efficiency to their operations and manage costs. Robotic Process Automation (RPA) and Intelligent Process Automation (IPA) are immensely helpful here. Parsing of financial deals is just a matter of a few seconds, thanks to AI. They can also help manage contracts and act as brokers, simultaneously taking over routine tasks, thus improving productivity and efficiency. All this transforms to increased revenue, reduced costs, and a boost in profits. Robotic automation of processes can reconfigure the financial sector and make it more humane and intelligent. Automation of about 80% of repetitive work processes helps officers dedicate their time in value-added operations that require a high level of human intervention like product marketing.
What we need now is not just empowering of banks by automation, but making the entire system intelligent enough to beat the newly emerging FinTech players. This has prompted a lot of banks to use software robotics to ease the back-end process and achieve a better functional design. SBI plans to institute an ‘Innovation Centre’ to explore RPA, which can help in making internal banking processes more efficient.
This system has been used by some foreign banks to recognize, extract important information from old loan applications, lease agreements and feed it to a central database which can be accessed by everyone. It can help in costly and error-prone banking services like claims management by drastically reducing the time spent in reading or recording client information.
For instance, JPMorgan Chase’s COiN reviews documents and extracts data from 12,000 documents (which, without automation, would require more than 360,000 hours of work) in just seconds.
A minuscule percentage of the Indian population has an idea of credit. Even to this day, applying for loans is considered a hassled process. It is also an annoying task for banks to analyse an individual’s creditworthiness due to the lack of credit history.
The use of Big Data and Machine Learning to analyse spending patterns and behavioural data of a customer over 10,000+ data points can help banks have an insight into the customer’s creditworthiness. This also helps in giving pre-approved loans to a huge range of customers without the need for paperwork and allows self-employed and students (as they are out of the financial fold) to obtain credit. In case of SME and corporate loans, AI simplifies complex and critical borrowing process, identify the potential risks in giving the loan by analysing market trends, prospect’s behaviour and identifies even the slightest probability of fraud.
The Punjab National Bank scam exposed the banking sector to an enormous amount of risk and shook the regulators, financial and stock markets, and the banking industry. AI and due diligence can monitor such potential threats and help banks install fool-proof surveillance and fraud detection systems. Surveillance in banks has been through audits and sampling. Some data sets and files that are capable of causing huge risks may not be covered in these samples. The algorithmic rules-based approach can help in monitoring of each and every file, and machine learning techniques can keep a database of all such files which pose a risk to the bank.
Banks, while providing secure and swift transactions, can use AI to detect the fraud in the transactions or find out any suspicious activity in the customer’s account on the basis of behaviour analysis. With an increasing percentage of cybercrimes in the recent years, AI can be used to maintain cyber-security and most importantly, in safeguarding personal data. Citibank has invested over $11 million in a new anti-money laundering structure already using machine learning and big data.
AI-based systems can help in compliance by judging the functionality of the internal control systems. AI can also be a game-changer by detecting insider trading that leads to market abuse.
Insurance underwriting and claims:
In this era of bancassurance, customers are more likely to come to banks rather than visit insurance agencies. Insurance sector can reap the benefits of AI in underwriting, claim-handling procedures, and fraud detection. It can also help in identifying risky behaviour and charge higher premiums to those groups of customers. Insurance firms have an enormous amount of data which can help make mathematical models and predict risky behaviours accurately. Such data can also be lent to banks to be used in customer risk identification. This reduces the turn-around-time (TAT) for both loans and insurance. For example, to analyse the damage to a vehicle, deep learning techniques can analyse an image of the vehicle, and calculate repair cost using predictive models.
Jack Ma, the founder of Alibaba, warned the audience at the World Economic Forum 2018 at Davos, that AI and big data were a threat to humans and would disable people instead of empowering them. A massive deployment of AI in banks would come with its share of risks and opportunities. Banks increase their investment in AI every year, often at the risk of becoming obsolete. But what we also need to understand is the risks to the system that AI can pose.
1. Loss Of Jobs
Banks face the risk of backlash from their employees due to the potential automation of tasks, which can lead to job loss and job reassignments. AI, in the garb of increasing enterprise productivity, will reshape the way the employees perform their jobs. This could lead to possible dissatisfaction among employees, resulting in resignations or employees being fired due to inefficiency. AI can replace a teller, customer service executive, loan processing officer, compliance officer, and even finance managers.
2. The Opacity Of Processes:
While deep learning models and neural networks in AI have proven over time to be perfect than human decision-making, they are often not transparent in terms of revealing how they generated such conclusions. It then becomes a challenge for bankers to explain that to the regulators. Justice Srikrishna Committee has mentioned that the biggest challenge in using big data, artificial intelligence is that they operate outside the framework of traditional privacy principles. This could now act in a reverse way and expose banks to risks without their knowledge. It could also possibly give rise to hidden biases in decision making since AI has access to data of all the customers.
3. Reduced Customer Loyalty
There is also a fear of reduced customer loyalty due to less customer contact and the lack of essence of “human touch.” Banks, especially in India, have an emotional value as they help many in cherishing their long-standing dreams—be it a beautiful house or a good education for students. All this could be lost due to AI and automation. The socio-economically backward groups would be the biggest losers and most affected in such a scenario due to low levels of education and the digital divide.
Nick Bilton, tech columnist, wrote in the New York Times, “The upheavals [of artificial intelligence] can escalate quickly and become scarier and even cataclysmic. Imagine how a medical robot, originally programmed to rid cancer, could conclude that the best way to obliterate cancer is to exterminate humans who are genetically prone to the disease.” The message conveyed here is that banks have to develop an understanding of the effects of digitization and develop an expansive foresight into the prospects of AI—so that we as humans have control over AI and not the reverse. The area that banks should now focus on is Data Acquisition. The lack of proper customer records is the biggest hindrance to AI.
We should ensure that the data used by banks are KYC compliant clean data as these would be used in AI models. Massive data infrastructure is required to leverage AI. Proper inspection of data and checks of accuracy are also needed before using such technologies in the public domain.
Analysis and standardization of data:
The amount of data with banks is so enormous that Oracle and Accenture have entire departments storing bank data. What we need is a proper analysis of the data, and that requires a high level of leadership skill to bring together cross-functional teams—one with knowledge of financial business, and the other with requisite machine learning skills to formulate a plan and infrastructure across various departments for efficient usage of such data sets. AI remains a niche-oriented domain with a shortage of talent and expertise.
Several experts in the U.S. and the U.K. opine that cyber, political and physical threats arise with the growth in the capabilities and reach of AI. The recent Facebook scandal highlights the risk corrupt data practices can bring to a firm. Complete transparency while venturing into new AI projects also should be ensured so that banks don’t face reputation risks.
Banks should start building AI systems with a small set of complex data and add subsequent ones, thus creating a universal record of each client. Adequate investments should be done on the safe storage of data and prevent it from leakage. This will help the bank detect potential hazards in the implementation stage of the project and enable efficient identification—and then execution—of goals and priorities of the organization. Artificial intelligence will soon become the sole determinant of the competitive position of banks and a key element enhancing their competitive advantage.