“Al”chemist in the making: Is a panacea in sight for the Financial Services industry

This article was originally published in Financial Express and has been reproduced here with permission

Artificial Intelligence (AI), the raison d’etre for cinematic bewilderment in sci-fi movies since ages, is now one of the most important fabricators of contemporary thinking. Several innovations pertaining to AI including chatbots, virtual assistants, machine learning, robotic process automation (RPA) and neuro-linguistic programming (NLP) are being liberally adopted in a crowded marketplace.

In the Financial Services (FS) industry, the flare-out has been quite phenomenal. Banks and insurers are looking at AI to provide meaningful resolutions to their perennial issues, such as cost, efficiency, profitability and staffing, among others. Goldman Sachs has forecast a global money pool worth £26 billion to £33 billion by 2025 in the FS industry alone, primarily attributed to cost savings and new revenue streams on account of the AI revolution.

In an internal benchmark developed by the Capgemini Invent India team, where more than 100 global banks and insurers were assessed on the adoption of AI and more than 220 use cases developed, there are several interesting insights. Leading players and vendors are busy designing a symbiotic ecosystem of platforms, solutions and illustrations that not only give assurance to Wall Street investors but also provide eclectic choices for the consumers.

Major technology providers like Apple (Siri), Facebook, Amazon (Alexa), Google (Home), IBM (Watson) as well as several fintechs are busy refining their AI solutions to forge stronger alliances in the industry. Noteworthy cases of applicability have been around chatbot-driven interactions, virtual assistants, fraud detection, algorithmic trading and AML pattern detection.

Chatbots and virtual assistants provide a human-like interactive personalization experience for consumers on a digital platform. They engage with the bank customers to provide advice on recurring themes and guide them around personal financial analyses and day-to-day transactions. Typically, query resolution, financial advisory, social engagement, quote generation and claims management are some of the most notable use cases gaining prominence, thanks to chatbots. Gartner states that by 2020, 85% of consumers around the world would be interacting with a non-human entity for their business transactions.

• Bank of America has deployed its Erica chatbot on the bank’s mobile app, which also doubles up as a digital assistant. Erica uses a combination of AI, predictive analytics and cognitive messaging to help customers make payments, check balance, save money and pay debt. Customers also get help on their FICO score and to view educational videos and other content.
• Progressive, a leading US insurer, provides auto and home insurance tips to its customers via the voice-enabled Google home devices.
• HDFC Bank (Eva) and SBI (SIA) have incorporated advanced chatbots to better respond to customer queries.

Machine learning too has fast emerged as one of the most invested-in areas in AI. Computers on self-learning and deep re-learning modes have been entrusted to perform banking services of varying complexities. Document analysis and data mining, trading assistance, risk profile assessment, compliance monitoring and fraud detection, and complex financial advisory are some of the well-adopted use cases in this category. Notable examples are:

• JP Morgan has launched the COIN (Contract Intelligence) program that has machines interpreting commercial loan agreements with reviews and audits. The firm has plans to extend this to other complex areas like credit-default swaps and custody agreements. Another program called LOXM is used in trade executions based on a deep reinforcement learning mechanism.
• Wells Fargo has rolled out “Intuitive Investors” to automate investment portfolios for its HNI clients.
• Deutsche Bank, with the help of machine learning, has been sifting through the annals of voice and video recordings of its employees to meet the demands of regulatory compliance. The bank has also deployed Alpha-Dig, with underlying NLP capability to unclutter financial jargons and discover hidden meanings in various company reports, which would otherwise have an impact on its stock price.
• Ping An, a leading Chinese insurer, has a slew of initiatives embedding machine learning in face recognition, voice biometrics and NLP to facilitate a seamless claims experience.

According to the Capgemini study, AI adoption globally can be categorized into three prominent clusters: nascent, prospective and advanced.

In the nascent cluster, the FS players are at the initial stage of exploring AI technologies to improve internal efficiencies and boost customer experience. The implementation of a chatbot technology is aimed at resolving basic queries about products and services, while RPA helps in automating repetitive tasks. Machine learning on the other hand is used to analyze customer data and derive key insights.

In the prospective cluster, the FS players are increasingly partnering with fintech startups to implement mature AI technologies to gain operational efficiency. They are capitalizing on the benefits of the initial AI implementations. The technologies are capable of performing more complex exercises, such as banking transactions, monitoring compliance and knowledge management.

The advanced cluster is where AI adoption reaches its pinnacle. The top FS players have made significant investments to build in-house AI innovation hubs, which serve as the focal points to assess, build, refine, test and monitor AI frameworks, algorithms, prototypes, solutions and devices. These companies regularly experiment with intuitive solutions to optimize organizational processes. AI systems here are calibrated to undertake pioneering work in AML / fraud mitigation and pre-emption, user behavior exploration and complex transactions involving biometrics.

A well-conceived AI ambition in the FS industry should begin with a rigorous maturity assessment (self-diagnosis) phase, where the incumbents need to evaluate areas within their value chain through an optimal mix of quantitative and qualitative probes to identify candidates ripe for an AI makeover. As the study foresees, the onus lies on the banks and insurers to have a tenacious approach towards evaluating the AI narratives that are more pragmatic than ornamental.

The evolving developments in AI will keep the momentum intact as stakeholders plan to assess, regulate and monetize their schemes to stay afloat in a competitive marketplace. In the end, it will be the consumers who may indirectly end up influencing the dynamics; the machines, then, would have come a full circle.