AI in currency: How it is quietly rewriting the rules of modern banking

Artificial intelligence is quietly transforming consumer-facing industries

When Delhi-based self-employed professional Seema Sharma’s daughter turned 18 and needed to upgrade her bank account from minor to major, the two went to the branch armed with her Aadhaar, PAN, birth certificate, multiple photocopies, passport-sized photos — the works. Sharma braced herself for long queues and tedious paperwork. To her surprise, it was all done in a matter of minutes. The bank had anticipated the transition. None of the physical documents she carried was needed. 

Behind this seamless experience was artificial intelligence (AI) at work. It left Sharma impressed again when she closed the auto loan on her BMW X1. It was a breeze compared to the two-week ordeal she went through for getting a no objection certificate (NOC) on closing a loan against her Maruti 800, nearly 20 years back. The 15-odd backend steps were handled by AI. 

“Previously, repetitive but critical processes were done manually. Now AI does it,” says Rohit Pandharkar, technology consulting partner, EY India. Of course, the final go-ahead is given by a human, after all checks. 

Real-time service delivery in a tightly regulated sector like banking was once unthinkable. Not anymore. From onboarding new customers, cross-selling, or up-selling products to fraud detection, compliance, and internal workflows, banks are embedding AI into various aspects of their architecture. While full AI-driven autonomy remains a distant goal still, due to regulatory and operational constraints, the pace of transformation is accelerating. 

“From 24x7 conversational agent and real-time fraud detection to automated compliance, AI is becoming central to banking operations,” says Sonali Kulkarni, country head, BFSI (banking, financial services, and insurance), Microsoft India and South Asia. 

ALSO READ PART II: From cancer detection to treatment, how AI is transforming health care

Intelligent banking 

AI’s most visible impact is in customer-facing areas. 

At HDFC Bank, it is powering customer service interactions across WhatsApp, SMS, email, and voice. The bank is exploring AI-powered co-pilots for its relationship managers to assist in query resolution. 

Axis Bank has been running AI-led experiments for two years now, with a focus on enhancing employee productivity. It has deployed Axis Deep Intelligence (ADI), a GenAI-powered chatbot across 5,500 branches supporting more than 100,000 employees to resolve queries. The focus at the bank is to improve turnaround and average query handling time.  

“The goal is to adopt AI in every facet of work we do. AI will make banking hyper-personalised,” says Avinash Raghavendra, president and head, information technology, Axis Bank. 

The Aluva (Kerala)-headquartered Federal Bank has introduced Feddy, a multilingual AI assistant that answers customer queries. And Yes Bank is using AI for document processing and automating loan disbursals.  

Across banks, the goals are the same: Faster response, lower operational costs, and better customer engagement. 

At Barclays, too, AI is being embedded across functions: Fraud detection, personalising customer experiences, to automate route processes, providing quick document summaries etc. Recently, the bank deployed tools like Microsoft 365 Copilot, which acts as an intelligent assistant, helping with tasks such as brainstorming, writing, coding, and searching, and GitLab Duo, which uses GenAI to make employees more productive, automate content creation and streamline operations. 

Praveen Kumar, CEO, Barclays Global Services India, is of the view that AI is making banking intelligent, efficient and intuitive. “It’s shifting our focus from mere transactions to personalised, predictive services.” 

AI, adds Anil Kanwar, India regional director of data streaming platform Confluent India, thrives on real-time signals. “And banks are finally adapting their systems to respond the moment an event occurs — be it a login, know your customer (KYC) update, a transaction, or a compliance flag.” 

ALSO READ PART III: Buy-buy AI: Why India's retailers are betting big on AI-powered growth

Automating complexity 

Much of AI’s impact is in areas previously mired in human-driven grunt work. 

Consider the NOC issued after closing a car loan. Earlier, it involved 15 to 17 back-end steps: checking dues, penalties, liens, repossession status, and cross-verifying credit history. Today, AI can automate many parts of this chain and reduce processing time. 

Similarly, credit memos that once took bankers several days to prepare are now being auto-generated in minutes. AI reads financial documents, fetches data from APIs, and presents a risk sum­m­ary that a human analyst then reviews and approves.

Pandharkar of EY India explains with an example: Suppose an auto ancillary company asks for a ₹12 crore inventory loan.  They would need to analyse three years of profit and loss account, balance sheet, goods and services tax (GST) filing, bank statements, Udyog Aadhaar (for small businesses), credit scores, fraud checks, PAN verifications, and so on. Earlier, at least two analysts spent five working days putting this together. Now using GenAI, the system carries out the necessary checks, goes through 300 pages of source material, all in less than 10 minutes, and generates a five-page credit memo. 

Banks are also deploying AI to read and interpret complex PDFs, summarise reports, and even conduct sentiment analysis of customer conversations. 

“The real gain is in the time saved,” says Pandharkar. “AI can read 300 pages in seconds. Humans can’t compete with that — so the model becomes augmentation, not replacement.” 

Armed with intelligence provided by AI, human workload is reduced, shifting from the drudgery of going through multiple documents and several hundred pages, to, say, issuing loans or taking the final call. As Kanwar of Confluent India puts it: “We are moving from delayed intelligence to always-on intelligence.” 

In customer service functions, banks are deploying AI-powered voice bots that replace clunky IVRs. Instead of navigating phone menus, a customer simply says, “I need my tax certificate,” and gets it. 

When digital banking started over a decade back, the experience was patchy, but it reduced the need to go to a bank branch. Over the years, it improved. Now, AI has taken it to another level.  

“AI-driven customer services and real-time analytics are the norm now. It’s leading to at least a 30 per cent boost in human agent efficiency, and a two-minute reduction in call waiting time,” says Ganesh Gopalan, CEO and co-founder, Gnani.ai, which provides conversational AI and other services to multiple sectors, including banks. 

 

Proof of concept to production 

Much of it started with proof of concept (PoC) but is fast shifting to actual roll out. 

Banks are now deploying production-grade solutions at scale. According to Kan­war, 65–70 per cent of BFSI IT budgets now touch AI in some way — through data streaming, fraud analytics, or hyper-personalisation. 

Microsoft says that partners like State Bank of India have cut query turnaround time by up to 60 per cent using Azure-based GenAI tools. At Axis Bank, GenAI is being applied not just in customer contact centres but also in software development lifecycles — helping generate test cases, write code, and automate documentation. 

“The mindset has changed,” says Gopalan. “AI isn’t a side project anymore. It’s being treated as a strategic layer.” That AI infrastructure layer is being built by hyperscalers like Microsoft Azure, AWS, and Google Cloud. 

AI adoption is also being shaped by regulation. The RBI set up a committee in November 2024 called FREE-AI, or Framework for Responsible and Ethical Enablement of AI, headed by Pushpak Bhattacharya from the computer science and engineering department at IIT Bombay. The aim is to ensure banks remain compliant, ethical, and risk-aware as they scale AI depl­­o­­­yment. The committee is expected to give its report soon. 

Despite automation, AI still functions with human oversight. For example, generative models may draft responses or flag fraud, but a person typically validates the outcome — especially in lending or regulatory tasks. 

Humans remain the final call in decision-making, thanks, in large part, to regulations. That, however, has not stopped innovation. 

“In the near future, banks will deliver deeply personalised services where AI anticipates customer needs and offers real-time contextual recommendations,” says Mahesh Ramamoorthy, chief information officer (CIO) of Yes Bank. “Customer interactions will evolve from app-centric interfaces to intuitive, conversational experiences powered by GenAI and natural language understanding.” 

So, in the future, armed with AI, your bank won’t just respond to your needs — it will anticipate them. Imagine waking up to a gentle nudge: “Your rent’s due in two days. Shall I make the payment?” Or getting an alert while shopping for an air conditioner online: “Want to split this into three payments?” No digging through apps, no scrolling through menus — just seamless, intuitive assistance exactly when you need it. This isn’t science fiction. It’s the new baseline of financial engagement: Hyper-personalised, predictive, and embedded in your everyday digital life. 

  The road ahead 

Accuracy, however, remains a major concern. Even small errors in GenAI systems can prove costly in banking. This is why AI is still used cautiously in critical workflows. Explainability, transparency, and auditability are top priorities for CIOs. Budgets, too, are still modest. 

Most banks are investing in the single-digit crore range for GenAI, focusing first on high-impact but low-risk areas like customer support, workflow automation, and document summarisation. 

Still, the benefits are already showing. According to EY estimates, operational efficiency gains of 15–20 per cent are now common in AI-enabled functions. 

As AI matures, the future of banking will become more invisible, intuitive, and intelligent. Apps will be replaced by conversations. Workflows will be executed in the background. And services will be tailored in real time to individual financial 

journeys. 

As Puneet Chandok, president of Microsoft India and South Asia, recently said at a session in Mumbai: “AI is undoubtedly the most transformative technology of our time and its impact is particularly profound in India.” 

In the coming years, the most successful banks won’t be the ones with the biggest branches or flashiest apps. They’ll be the ones with the smartest algorithms — silently working behind the scenes to make finance faster, safer, and human-like. 

The writer is a Delhi-based journalist