When Nimish Panchmatia, chief data and transformation officer at DBS talks about the Singapore Bank’s success he sounds like dozens of other banking technologists:
“Our focus around the customer and putting customer first is probably one of the biggest ingredient in our success,” he said.
Familiar talk, but the bank’s list of awards means you have to pay attention. Euromoney named it the world’s best digital bank in 2021 and the world’s best bank, the fourth year running that a leading publication has named it the world’s best bank.
The bank’s internal policy is named RED — respectful, easy to deal with and dependable, Panchmatia said. All very laudable — how does DBS make it happen?
“First our messaging and our vision are very clear and at the top of house there is complete alignment about what we want to do, in culture and technology, how we want to handle our customers, in sustainability, and what we do for society. Then we take this big vision and break it down to very understandable small nuggets, and this is how people get engaged. Lofty statements are great, but people have to believe it and know that as it is explained makes sense to them.”
The bank also has a strong innovation framework, he added.
“We allow ideas from outside, we allow for failure — every year we have a failure award for ideas we thought were good. We tried them, but we found out that they didn’t work. Maybe the technology wasn’t ready, or maybe we didn’t really understand customer need as as well as we thought. Then through testing we figured that maybe that’s not a good idea. We’re quite good at killing ideas within that innovation framework.”
DBS has been quite serious about being digital across its entire tech stack.
“Often the the technology at the front looks really good, but in the back there’s a lot of processes that are not in synch with the technology you see in front. Our approach has always been that whatever we digitize we have to be digital end to end.”
Ten or 11 years ago a lot of the bank’s technology was outsourced, mostly to IBM, he said.
“But we saw what was going on with Amazon and Apple and Facebook, and in China with Alibaba. Digital was starting to play a big role, but we didn’t really understand it. We decided that if we are going to compete, we need to understand it, so we in-sourced. Now 90% of our technology is in-sourced.”
That all paid off when Covid-19 hit.
“I was in Hong Kong at the time and saw banks that didn’t know what to do. We switched in no time, seamlessly. Many many banks, well known across the globe, didn’t have laptops (their staff worked on desktop computers) and didn’t know how to work from home. We were able to react quickly. Nobody had ever practiced 70% to 80% working from home for a long time. We had to adjust our IT stack but because we owned it and understood it, that made it a lot easier for us to react.”
The bank continues to aggressively investigate and experiment with technology. It tried IoT (Internet of Things) and decided to wait for further developments, Panchmatia said.
But Artificial Intelligence (AI) and Machine Learning (ML) are not futuristic, they are used now.
“We have been very conscious about data and applying AI models — it has to be core to our business and we have invested significantly in these capabilities.” The bank has over 700 data professionals and they are across the organization, not in a central pool separated from the lines of business.
“We are building scale that probably only technology companies have in that space.”
DBS holds onto its staff well, in part by offering growth opportunities and continuous challenges, and when it does lose technologists it is more likely to be to a Google or Amazon than to another bank, he said. The bank has worked with DeFi, Web3 , blockchain and custody of crypto. And it is experimenting with the Metaverse internally and with conversational AI to develop chatbots for customer service.
“We worked with a partner and found that with the pace of change and context it was very difficult to work with a generic chatbot, so we built our own.” In keeping with its policy of being respectful of customers it will never force them to use a chatbot, and it maintains a 24×7 call center with short wait times.
In consumer finance, the bank relies heavily on data.
“We are quite big on personalization and when you are hyper personalizing you don’t have a gender bias. We use our data to tell us, based on the individual profile, the best conversation and product suite a customer would be interested in. We don’t consciously sit down and segregate our data on a male/female basis.”
He is aware of studies that have shown algorithms and models can be biased.
“We are quite conscious of algo drift,” Panchmatia said, “and we are diligent about it.”
How models drift is not a science, he added, so the bank’s teams have data scientists plus what he describes as “people-aware” scientists and engineers and behavioral scientists looking at model drifts.
“Model drift requires some judgement, so we we look at outcomes from our models, not just from a data science perspective or business outcomes but also through behavioral science.”
AI and ML provide great insight, he added, although a bank needs a fine balance to avoid making customers feel creepy.
“When done correctly it can a very powerful tool. You have to have the ethics of it as well. We have a very structured approach, a Responsible Data Use Committee. We discuss use cases from all angles, not just can we make more money out of this.”
One project they worked on used machine learning to identify the right time and right insurance plans to engage customers. The protection needs of customers change as they go through different stages in their lives. The machine learning model predicts when customers require additional or different types of insurance and prescribes a suitable insurance product.
DBS uses continues discovery, connecting with groups of customer and non-customers either physically or digitally in deep two-way conversations that feed into the design of a new product or service. Once a new product is packaged, the bank tests it. This process led to a fixed rate foreign exchange (FX) product for SMEs (small and medium enterprises).
“SME customers said they are exporting and importing and need an FX rate,” said Panchmatia. “Spot rates don’t work from their planning perspective, they wanted a fixed rate that would last for seven to 10 days and it was not there in our bucket. They could buy forwards, but those are high value. This was a part of the market that banks hadn’t addressed — providing a fixed rate for smaller payments for a short time.”