Call the HDFC Bank call center.
Lady AI replies “Namaskar, welcome to HDFC Bank”.
You: Mera’s credit card balance batain. (How much is my credit card balance?)
Lady AI: You want to know the outstanding balance on your credit card, is that correct?
You: Hahn. (yes.)
After verifying your identity, Lady AI will inform you of your card balance. If the call hangs up and you redial, you are reminded of the last time you were interrupted. Want to pay your bill? Lady AI will immediately send the payment link via her SMS.
My job is done. No human intervention.
Bank customer service has evolved from traditional call centers to interactive voice recordings (IVR) to AI-powered bots. And the latter is getting smarter by the day. Lady AI (a fictional name) is part of a new generation of AI bots with natural language skills. They are trained on large datasets of customer interactions, run 24/7 and use ML algorithms to learn rapidly to recognize customer and banking needs. I’m here. “Currently, a small percentage of calls are exceptional and complex and are routed to human agents,” said Anjani Rathor, Chief Digital Officer at HDFC Bank.
It’s clear that great data is the foundation of any powerful AI system. As such, the banking industry is gearing up to consolidate data from disparate sources into one place. Balaji VV, his CTO at ICICI Bank, said: “Banks are now creating enterprise data lakes to integrate, process, and model data in real time and on the move.” It will be simplified.” Banks are already using AI for offer personalization, credit underwriting, risk management and analytics.
AI can help banks offer personalized offers, pre-approved and pre-approved loans to existing customers, but also offer similar offers to new customers and those with no banking history. Efforts are underway. How does that happen? “Currently, there are thousands of customer profile categories created by banks to identify customer segments with similar requirements when making offers, and specific customer profiles can be identified based on the data available to banks. We group them into buckets,” says Balaji. Accepting or rejecting offers provides additional data points to improve our system. HDFC Bank’s Rathor said it “continuously learns on the go,” adding that as AI tools evolve, they are likely to incorporate external data sources as well.
AI is also being used to make lending decisions, where algorithms combine banks’ traditional underwriting models with intelligence based on credit data to make decisions. “This technology reduces the risk of default and enables banks to make more informed lending,” said Sriram Srinivasan, chief digital officer at Ujvan Microfinance Bank. . At the same time, one private bank official points out that the automation of decision-making in private banks has reached about 50-60%. The rest are still based on traditional methods. Risk reports have also received attention. Suppose you typically swipe your credit card to buy food or groceries. One night, for example, when the system senses a late-night swipe to buy gold, something you’ve never done before, you’ll get an instant confirmation call.
AI is also helping insurance companies. AI tools, for example, use customer-uploaded images of cars to perform a first-stage assessment of an insurance claim. “We are working with a number of insurance companies and implementations are going well,” said Geeta Gurnani, her IBM Technology CTO and his leader in Technical Sales for the IBM India and South Asia region. . “The efficiency of claims for damages based on this AI model improved by 40-70% in terms of accuracy.”
Will there be an impact on employment? Regarding job requirements, Gulnani said only certain day-to-day tasks will move to AI. HDFC Bank’s Rathor added: “Low-skilled repetitive or very simple tasks will be advanced and completed by systems, computers and technology, and humans will evolve into something bigger.” Banks are already data engineers, data scientists , hiring UX designers, etc., which wasn’t the case a few years ago. “No technology can replace humans in that way, but the workforce may not increase proportionately,” says a technology chief at a commercial bank.
AI tools also have their own biases. They are like black boxes, no one knows how the machine decides. “Consistency [in outcomes] This is very important in building trust in AI technology as it helps ensure the reliability and accuracy of the tools,” says Balaji. An audit of AI models will be requested tomorrow. “We have launched several AI explainable products, which fall into the gamut of AI governance. We will be able to audit and investigate,” he says Gurnani.
There are still AI advances in banking. For example, you cannot enable transactions. Additionally, banking tools like ChatGPT that learn from everything available on the internet would help banks get to know their customers better. “The more data we have about our customers, the more intimacy we have as a bank,” says Raiser.
@anandahikari
