The five superpowers of data

• DBS believes banking needs to disappear out of people's lives to provide them with better customer experience. 
• Big data has helped the bank strengthen its relationship with the customers
• DBS has learnt to design products and services based on existing data

If you could be a superhero, which of the following superpowers would you pick: flying or invisibility? This, I heard recently, is an interview question asked by one of the big technology companies. A question no doubt designed to test a candidate's ability to think on their feet and demonstrate reasoning. However with the accelerating pace of technology advancement, navigating the future of business is like choosing superpowers. The emerging superheroes: blockchain, APIs, cloud, big data, machine learning, internet of things (IoT), and biometrics all claim to have superpowers that enable you to save the world.

So which one to pick? Where to start? It seems obvious but many seem to get it wrong. You select a superpower that is going to solve a business problem. Tinkering around with technology because it is the flavour of the month is likely to get you nowhere. Which problems to pick should be driven by vision and strategy – no change there.

At DBS, we have set ourselves the vision of "making banking joyful" - which usually brings a cynical smile to most people we tell, which is exactly why we like it – it is ambitious and differentiating. However, we clearly have a long way to go and it feels like we need the entire squad of technology superheroes to pull it off. We believe that to make banking joyful, banking needs to disappear out of people's lives. No one wakes up in the morning and says "hey let's go do some banking today." Therefore, to be joyful, banking needs to become invisible. When you get out of an Uber, you do not have to worry about the payment. In the same way we want to integrate banking into people's lives.

The magic of APIs are going to help us integrate into ecosystems; biometrics are going to solve the authentication challenge; IoT is going to allow payment between objects; and robots are going to automate operations to drive efficiencies. In banking, all these technologies are young and their superpowers are still developing. However, there is one superhero who is quietly delivering outcomes for us... and that is "big data" along with his sidekick "analytics" a.k.a. cognitive a.k.a. machine learning a.k.a. artificial intelligence (apologies to the precise definition police). And the cool thing about this dynamic duo is that between them they have a whole range of superpowers.

They can see into the future. At DBS, we now use machine learning to predict when a relationship manager is going to resign (and potentially take our clients with her). We can predict when an ATM is going to mechanically fail (important to us, as we have the busiest ATM network on the planet and we take downtime very seriously). Our audit team can predict which branch will have the next operational issue. We can predict branch and ATM queues and even the sales performance of candidates.

They are all seeing. Machine learning can help monitor the bad guys. We use data to detect rogue traders and fraudsters in the procurement and trade areas. We also use machine learning on video files to monitor the IT guys who have access to our production systems.

They are all knowing. We were an early adopter of IBM Watson and we use it to analyse the vast quantity of research material available to make investment recommendations.

They can speak any language. Recently, we launched a mobile-only bank in India, Digibank. In order to make it scaleable, we integrated a chatbot into the app to help answer customer questions. To do this, we partnered with a startup called Kasisto. The bot understands natural language and responds to a high percentage of customer queries.

And finally, they have X-ray vision - well not quite... but sometimes it seems that way as banks do know a potentially spooky amount about their customers. Your bank knows your age, gender, financial standing, address, where you work, and how much you earn. They also know what you spend, where you go on vacation, your favourite restaurants, and what you do in your spare time. They even know where you are and what you are doing. This is the ultimate data superpower, which has to be used responsibly. This is the one to worry about if it gets into the hands of the bad guys. However, it is a key in creating amazing customer experiences and, ultimately, in making banking disappear.

I have a lot of fun with the superpower metaphor but there is a serious side to it. It helps address one of the challenges we continue to face on our AI journey. People find it difficult to imagine the future and what problems a new technology can solve. People tend to think in increments. By asking the hypothetical question "how would you run your business if you could see into the future or if you knew absolutely everything about your customers?" you typically get more imaginative ideas that can be delivered using technology that exists today. For example, it was when our audit team asked themselves "what would change if you could see into the future?" that they said they would go and audit the branch that will have the next issue, which led to the development of their predictive model mentioned above. By the way, they now have their very own superhero catchphrase – "the future of audit is auditing the future."

We have just only started. We have learnt a lot on our cognitive journey. Here is a summary:

Always start with the question - unless you are clear about what problem you are trying to solve, you run the risk of running round in circles. Simply wallowing around in the data does not work. We tried it, don’t do it.

Start with your own data - we were tempted to supplement our own data with the beguiling world of social and IoT data. However, it is hard to get hold of and work with and we had so much of our own data that we were not using, we decided to focus internally initially. In a couple of cases we have supplemented solutions with external data but so far this has been the exception.

Don't work alone - we worked with IBM, Kasisto and ASTAR (the research arm of the Singapore government). We started with very limited capability, talent is hard to come by and each helped us accelerate through the learning curve and at the same time yield results.

Design for data - as we have learnt the potential of machine learning, we are starting to design our products with data in mind. We ask ourselves "What data should I produce from this product or service that will enhance customer offerings or drive efficiencies?" and then design accordingly.

And finally..

Grab your mask and cape and step into the nearest phone box - there are so many untapped opportunities in banking and I suspect in pretty much every other industry. It's time to save the planet - well a small part of it anyway.