Like healthcare and hospitality, banking is also largely dependent on human interactions and emotions. While the potential for generative AI in banking is remarkable, it cannot judge and falter human behavior, especially when offering customized financial products and services to customers.
The McKinsey Global Institute (M.G.I) projects that GenAI could generate an additional $200 billion to $340 billion annually in the global banking sector, increasing industry revenues by 2.8 to 4.7 percent, primarily through improved productivity.
How are banks embracing AI?
In 2023, Accenture highlighted a transformative trend in Indian banking. It noted that with AI integration, banks can serve many customers simultaneously, increasing both transaction volume and digital interactions.
Back in 2017, HDFC Bank Ltd introduced India’s first AI-powered banking chatbot, Eva (electronic virtual assistant). Powered by Bengaluru-based startup Senseforth, Eva was designed as a breakthrough in customer service, capable of handling millions of queries instantly across multiple platforms. Eva set the standard for AI-enabled customer interactions with the ability to pull information from thousands of sources and respond in 0.4 seconds.
By 2020, ICICI Bank expanded this concept with its chatbot, iPal, and integrated it with Amazon Alexa and Google Assistant. This allowed retail banking customers to perform various transactions through simple voice commands. However, the service was discontinued in August 2021.
More recently, in 2023, the State Bank of India (SBI) announced a strategic AI-driven initiative aimed at improving decision-making and operational efficiency. With plans to deploy advanced data warehouses and data lakes, SBI is also exploring partnerships with fintech companies and non-banking financial companies (NBFCs) to revolutionize co-lending practices and drive a more connected financial ecosystem .
In the meantime, Deutsche Bank‘s innovative AI journey is powered by data approach, control and talent. It has collaborated with Google Cloud (since 2020) and NVIDIA (as of 2022), accelerating cloud transformation and AI adoption within the bank.
In 2023, it launched a bank-wide AI program with practical applications such as AI chatbots, developer support tools and unstructured data analytics, positioning it as an early adopter of generative AI and LLMs.
But is everything safe?
According to Kroll’s 2023 Fraud and Financial Crime ReportA survey of 400 senior executives on three continents found that 67% expect financial crime to increase in the coming year, with 57% considering external gatekeepers as a key risk factor. There is a link between money laundering and organized crime
significant, with up to $2 trillion (2-5% of global GDP) laundered annually as criminals try to mask illicit profits.
In response, banks are increasingly looking to AI.
For example, HSBC previously relied on rules-based systems to detect potential money laundering, often leading to numerous “false positives” that had to be manually reviewed. Now, in partnership with Google Cloud, it uses advanced AI trained on extensive data to autonomously recognize suspicious activity.
This AI-powered anti-money laundering (AML) solution is more accurate, reduces false positives and improves detection capabilities without preset parameters.
As banks evolve their defense mechanisms, fraudsters become more adept at evading them. Speaking at the DECODE webinar, Sahil Anejavice president at HDFC Bank, pointed out that traditional rules-based monitoring, while fundamental, is rigid and struggles to keep up with the fraudsters’ methods.
“For example, when UPI was first launched, there was a significant spike in fraud incidents. Banks responded by setting thresholds for unusual transactions, temporarily reducing the size and frequency of fraud. However, fraudsters quickly adapted to these rules, necessitating a shift to AI-driven platforms for better fraud detection,” he said.
Aneja spoke about a common benchmark for fraud prevention, indicating that within 30 to 45 days, fraudsters adapt and develop new methods to bypass financial institution platforms. This means that institutions must continually refine their systems and move to machine learning models.
What’s next?
Infosys Finacle, part of EdgeVerve Systems, a wholly owned subsidiary of Infosys, recently introduced the Finacle Data and AI Suite, a powerful toolkit designed to seamlessly integrate AI into banks’ digital operations and accelerate their AI journey. This suite offers a collection of platforms that allow banks to build low-code, predictive and generative AI solutions from scratch, with an emphasis on transparency and explainability.
This allows banks to increase their data readiness, standardize AI model development, leverage generative AI technologies, and deliver actionable insights across their ecosystem.
Meanwhile, Axis Bank believes that AI will not change the nature of work in India. However, the Mumbai-based company has expanded its technology team to 800 employees, up from around 60 almost five years ago.
The bank currently employs around 70 people working exclusively on AI and plans to further expand its team by 10% annually. This shows that AI adoption in the banking sector is not slowing down.
s-10:21 -d4:11:2024x