Standard Chartered Bank’s slogan is ‘here for good’, but it might as well be ‘safety first’. That’s how the banking industry is likely to compete against the onslaught of technology disruption, says one of Singapore’s more senior bank data executives: by putting compliance and governance around newfangled uses of data.
Shameek Kundu, chief data officer at StanChart in Singapore, argues that once banks have learned to harness the vast reserves of data swirling around in various silos, they can actually outcompete pure tech plays. That’s because they’re built to be dull and bureaucratic – the very attributes anathema to Silicon Valley and Chinese internet types.
“In some ways we are ahead of tech companies,” he told DigFin. The bank has invested $100 million to get its arms around its data, initially in response to global regulators’ demands for safeguards and reporting around terrorism financing and money laundering.
The result is a data lake that the bank came to realize was useful as a means of serving customers and generating revenues. “Legacy banks can match big tech companies in leveraging data,” he said.
The difference is that the stringent regulation of banks means they are designed to be bureaucratic and diligent about cyber-security, fraud, and so on. They are supposed to take seriously conflicts of interest.
The reality may not always live up to what banks are meant to be – but their processes are designed with compliance in mind. As customer data becomes more commoditized, and as the integrity of data becomes more valuable, banks could be better positioned than freewheeling tech companies.
“Privacy, quality and model management are all established practices in banks, with senior people responsible for them,” Kundu said. “Tech companies need to build these controls.”
Aggregating data from all functions worldwide is only the starting point, says Kundu. Taking advantage of it means stealing a page from the fintech playbook.
First, the bank is going open source.
“We’re just about to go open source with code around legacy systems’ data,” he said. The reason is to attract outside programmers who might want to mimic some of the bank’s experience in building the data lake. For StanChart, the advantage is to benefit from how these outsiders improve the code.
Legacy banks can match big tech companies
Going open source is also meant to improve the bank’s profile when it looks to hire developers.
Thirdly, open source, along with open APIs with third parties, enables banks to move beyond transaction data. Banks have long histories as information processors, working with structured (well-defined) data such as financial records.
But as banking moves to become a platform serving an ecosystem (i.e., as an operating system that runs many apps, supporting financial of customers through third-party partners) institutions need access to different sources and types of data – and being open sourced makes partners more willing to share.
For example, banks are now keen to provide lending to, say, a Chinese retailer’s customers, or offer auto loans through a car maker’s app. Being open source with respect to developing the bank’s data lake, and of course adopting open APIs, makes it easier to integrate, Kundu says.
Working with fintechs
The second big change Standard Chartered has made in its attempt to leverage its new data lake is to work with third-party fintech companies.
The bank has conducted up to 15 pilots with fintechs, and a handful have moved to production. One involves a U.S. West Coast company that modeled credit scores for lending to companies in supply chains (Kundu declined to name the company).
Another is with Switzerland’s Squirro, which is helping the bank combine relevant information for capital-markets trading counterparties. “It gives us relevant information about out clients,” Kundu said. This includes structured financial information as well as unstructured data, such as calls on securities (see below).
Kundu says a few pilots turn out to be duds, and a few make it into production – often those that take the bank’s data and provide ways for staff to use it, visualize it, and gain an insight or otherwise apply it, via algos; or deliver a smooth user experience. Then the new product is integrated into the bank’s existing transaction systems.
Data is a source of competitive advantage
The majority of pilots remain just that – good ideas but not ready for the bank to roll into production; for example, the bank may not be sure the partner is robust enough to work with such a large organization. Bambu, a B2B robo-advisor shop in Singapore, and KYC Chain, a reg-tech company in Hong Kong, were both inducted into pilots last year, but have yet to be graduated to production. The bank is in commercial discussions with Squirro, which has been paid for its pilot work.
Kundu says the bank knows that there are certain areas where tech companies will always have an edge, because the bank can’t get the kind of data sourced from, say, a messaging app. But if banks are to remain relevant to customers, “Data is a source of competitive advantage.”
SQUIRRO EYES NEW USE CASES FOR STANCHART
Dorian Selz, CEO and co-founder of Squirro, says he hopes working with Standard Chartered Bank to help it make sense of its data will lead to a broader relationship.
“We began working with them on a specific use case, in terms of a workbench,” Selz told DigFin from Zurich. A workbench refers to a means of helping someone work with data: the bank developed its data lake, and a fintech designs a workbench for banks’ staff to “sit” and be able to see and manipulate the information.
Shameek Kundu, chief data officer at StanChart, says the project involved combining data about counterparties in capital markets.
From there, Selz says, new questions arise.
“A corporate customer may ask, where’s my payment?” he explained. “A banker will ask, what’s happened to my deal? A relationship manager will ask, is a counterparty in Singapore treated the same as their affiliate in Dubai?”
To answer these questions, he says, banks have many systems and many sources of data – but they need one, single view.
Data lakes aren't useful without an access layer
Squirro’s origins go back more than a dozen years, to when the founders created a search engine for Switzerland, helping locals look up a restaurant or a shop. The experience taught them how to work with unstructured data.
That company was then acquired by SwissCom to provide similar services at an enterprise level. Although the nature of the searches had changed, everything still involved looking for information in walled off places – a supply-chain system here, a customer relationship-management system there, or even email archives.
The founders left that venture to set up Squirro with the aim of using technology to integrate all of these sources and extract new insights from it, via machine learning and analytics.
“Data lakes are wonderful,” Selz said. “But they aren’t useful without an access layer, and a way to make the data relevant to people.”
Selz says the company now works for a large U.S. bank, an insurer and an industrial company in Switzerland, Investec in the U.K., and Standard Chartered in Singapore. The fintech doesn’t deal with all parts of a bank. It focuses on four areas: corporate and institutional banking, real estate finance, asset management, and investment banking.