India’s most valuable bank, HDFC Bank, is leveraging artificial intelligence (AI) at its core by offering conversational interfaces, such as humanoids and chatbots, to transform the customer experience.
In the six months since she joined HDFC Bank, Eva has had over 1.2 million conversations addressing nearly 2.7 million customer queries. In an environment where the country’s biggest and most competitive banks are jostling to acquire modern customers, Eva is just what the bank’s country head of digital banking, Nitin Chugh, had asked for.
Earlier this month, Eva also became India’s largest banking chatbot.
Indeed. Eva, which stands for electronic virtual assistant, combines natural language processing and artificial intelligence to offer an almost human-like customer experience for users. For now, she works out of the bank’s homepage.
The core idea is to use bots, AI, machine learning (ML), and humanoids to offer a differentiated customer experience. “Convenience is a given, it’s no more a new frontier,” says Chugh.
For Chugh and HDFC Bank, Eva is among several building blocks that will help shape “a sci-fi” future, aligned tightly with the mission of providing a differentiated customer experience. And the ongoing convergence of new technologies, including AI, ML, and so on, is only making this journey more grounded, and not just another hype cycle.
“That’s how it’s happened in the past also. The only good thing is that there are quite a few things happening together now, “says Chugh. “It’s not like there is an AI revolution happening and AR/VR (augmented reality/virtual reality) is not happening, or IoT (Internet of Things) and cloud are not happening. Everything is happening pretty much together. So, because these are happening together, their ability to converge and make a very powerful interface is possible in a short time.”
Eva is not the only conversational interface HFDC Bank is using; it’s also leveraging humanoid robot IRA (intelligent robotic assistant) and the Facebook messenger bot. All are powered by AI and ML and are evolving fast.
Eva engages with HDFC Bank customers across the website and mobile interfaces, providing product details, fees and charges for various products, application processes, branch codes, and a lot more information in few seconds, but IRA is a physical robot guiding customers to relevant counters in a branch.
HDFC Bank worked with partners including Senseforth AI Research and Niki.ai to create them.
“Eva currently handles 50,000-plus semantic variations for thousands of banking-related intents,” says Shridhar Marri, CEO and co-founder of Senseforth. “Eva’s everyday interactions with users go through a conceptual banking knowledge framework, which in turn enhances her ability to field more questions accurately. Eva never sleeps and her learning never stops!”
Digital is not skin deep
“It’s not enough being simply digital anymore,” Chugh says. “It has to go beyond making life convenient for customers and become a launch pad for starting a whole new level of customer experience.”
Indeed, with competitive pressure from new banks on one hand, and the battle to gain the mindshare and engagement of millennial customers on the other, India’s financial services industry is going through a massive disruption. The pace of innovation will determine the winners and the losers. And just being digital isn’t enough anymore.
For Chugh, there are important questions that technology can help answer.
“You can’t tell the customer, ‘You know, now I’m going to give you a convenient way of dealing with me.’ Convenience is a basic necessity now. So what is the next level of experience for the customer? What is the next level of ease for the customer? What is the overall depth of the interactions that the customer is having with the bank, and the overall width of the transactions she is carrying out with the bank?”
For the new experiments to make sense and enhance customer experience, it’s crucial to answer the questions above.
“A lot of people can be digitally active, but sooner or later you realize that they’re just logging in to check their balances,” says Chugh. “And everything else is still the same old way. They’re writing checks, they’re withdrawing cash, they’re doing everything else like they were doing in the old days – to them it seems like that is how banking is. And digital is just a way of accessing my account and getting whatever information I need.”
At the core, these new platforms, powered by ML and AI, have to learn continuously from each customer interaction.
“The learnings are really about things that [Eva] doesn’t know. If a customer asks a question and the bot doesn’t know the answer, it gets recorded and somebody can go and address that. With machine learning, this will evolve and over a period of time will become a full-fledged interface for customers,” says Chugh.
HDFC Bank’s technology team also keeps a close watch and churns out lines of code to ensure the overall interface evolves.
“There is software that allows the bot to capture the information, analyze it, do some bit of self training, which is the easier part, and the rest of it has obviously got to be coded.”
For its part, HDFC Bank is not shying away from new experiments and the failures they bring.
“In our case, we are running around 200 experiments (at any point in time). Out of these, even if half of them fail, it doesn’t make any difference, because there are a hundred that can be scaled up. We do expect a lot of them to fail. So many of these learnings have helped us identify the right bets,” says Chugh.
While the technology experiments have been ahead of business all these years at HDFC Bank, the pace of experimentation has become faster and more intense. “You see this happening a lot more now,” he says.
To make a substantial impact, new technology products, especially those powered by AI, need to be part of a broader, sustainable platform strategy.
“The aim is to run these as a completely new channel. I can’t put a number on them, but almost everything that the bank has to offer can be delivered using a conversational UI (user interface) powered by AI on the customer side,” he says.
“We are looking at AI and ML as platform capability … When you need an interface which is unmanned, you need a fairly sophisticated level of intelligence.”
Apart from transforming customer experience, HDFC Bank’s conversational interfaces are also being applied internally.
“Internally, we are already experimenting with quite a bit of those things, from leave processing to ticketing and so on,” he says. “We are trying out several use cases, and that’s why we are taking a platform-based approach.” Over the next few months and years, all this will help offer even more personalized customer experiences. “Now what is likely to happen is that, because we will have some layers of artificial intelligence and new sources of data, the algorithms will become sharper, will become more segmented. We will be able to drive a high level of personalization rather than being at a segment level.”
Sci-fi today, reality tomorrow
Yesterday’s science fiction is today’s reality. And today’s sci-fi will become tomorrow’s mainstream. “Everything we are imagining is ending up being possible and real. I remember when we launched [our] watch banking application two years ago, it looked like sci-fi. How could you talk to your watch, get a few things done?” says Chugh.
So what’s HDFC Bank’s view of sci-fi when it comes to customer experience? “The sci-fi view is that you could … have a face-to-face conversation without moving an inch from wherever you are based. All of that can be made possible through AR/VR, and everything else can be embedded inside using voice or gesture. At the moment, this seems possible,” he says. “In the long term, it is difficult to tell what will happen because something new is going to come up. Maybe you’ll be able to bank from your car, your home, or using other home appliances that are also connected. That’s why I am saying it will be a convergence of IoT, AR/VR, and maybe some other new platform capabilities such as blockchain, other voice interfaces. I would think, in the next 18-24 months, you will see a lot of these sci-fi kind of interfaces.”
Learn more about changing the basis of competition in your industry. Why Machine Learning and Why Now?