You are here

UOB is banking on data

The bank is ramping up efforts to prepare its entire 26,000-strong workforce for a digital future.

BT_20191111_UOBJI6G_3935794.jpg
UOB's data drive will not be at the expense of its commitment to ethics.

ALMOST all banks like to boast their data capabilities, but not many have managed to go beyond acquiring insights to delivering results.

At UOB, its emphasis on data analytics has translated into very real benefits for the bank and its customers, driving business performance and enabling the bank to innovate at an unprecedented pace.

More than just a buzzword, it is fast transforming the bank's approach to its people, process and technology, said Richard Lowe, chief data officer of UOB.

For instance, one area that the bank has reaped the rewards from data analytics is in its anti-money laundering (AML) monitoring, which is known to be one of the most tedious and expensive undertakings for a bank.

sentifi.com

Market voices on:

UOB is said to be one of the earlier banks to introduce artificial intelligence to its compliance function, with the help of a partnership with regulatory technology company Tookitaki. It helped to flag suspicious activity that would not have been spotted with existing technology while bringing down the rate of "false positives", or false alarms set off by the stringent AML checks.

Based on numbers from a six-month pilot, the rate of false positives in name screening has gone down by 60 per cent for individuals and 50 per cent for companies. Meanwhile, there was a 5 per cent rise in alerts identified correctly as a defined AML risk that would not have been picked up otherwise, said Mr Lowe.

This adds to efficiency, as well as a real means of value-adding as the AI was able to discover things that were not possible before, he added.

But UOB's developments in the data space would not have been possible without its current architecture, a result of planning and commitment of management to put their money where their mouth is.

Mr Lowe, who happens to be UOB's first chief data officer, said that this came about when he joined back in 2015 to create a centralised data team that focused on business needs.

"It translated into my vision around how we use data to enhance customer experience and drive business performance, as opposed to the pure technical aspect of data," he explained.

That was the start of UOB's data transformation journey. The team decided to go ahead with a "data lake" approach - consolidating data from all around the bank's different units into a central depository.

"This sounds very easy, but with organisations with so many years of history, their data is everywhere and it is a difficult task," he said.

At the same time as the technology was being built, UOB concurrently started to set up its data governance framework as well as to work on the people component of the equation by hiring talented data analysts and scientists.

Within 18 months, the enterprise data lake was completed, where almost all of the bank's business systems data are loaded inside. While the idea of it might sound daunting to the ordinary layperson, the bank actually intended for staff to access and benefit from the data available.

For them to do so safely, more than 150 data discovery "playpens" - think data sandboxes - have since been created across the group for employees to "play" with the data to innovate or come up with products in the comfort that the data is governed, said Mr Lowe.

It is not just the central data team that is able to understand and tap the data - these days, almost every business function has employees who are able to perform data analytics, he added.

This came on the back of efforts made by the company to groom existing employees, aside from the hiring of data talent. He estimated that more than 500 staff were trained since 2015 on the various data visualisation tools. With the launch of UOB's latest "Better U" training programme, the bank is further ramping up efforts to prepare its entire 26,000-strong workforce for a digital future. With the competition for scarce data talent, the bank is also partnering local institutions such as the National University of Singapore and the Singapore Management University to grant data analytics scholarships to secure a pipeline of talent.

UOB's investments in its data capabilities have paid off in numerous ways. For one, it has enabled UOB to launch its Asean-focused digital bank TMRW in Thailand in just 14 months - quite a feat in the banking industry.

"As the digital bank is mainly driven through data analytics, we were able to deploy so quickly because we already had the data in one place," said Mr Lowe.

The digital bank also uses data analytics to derive insights from user behaviour to enhance customers' experience while using the app, such as sending alerts to customers when their spending behaviour diverges or if a payment that usually comes in does not.

Customers are able to rate how useful the prompts are, which gives the bank continuous feedback on how to refine its algorithms and also enable customers to have peace of mind without being too intrusive, he added.

Even as the bank continues to step up its data efforts, Mr Lowe emphasises that one thing which will remain non-negotiable is its commitment to ethics.

The responsible use of artificial intelligence, in particular, is one grey area that many banks and institutions are still grappling with. On UOB's end, its current position is not to use blackbox technology - when there is no clarity to how an AI-generated solution is derived - even if there is potential that it could result in more profits for the bank, he said.

"If we can't explain how the AI makes its decisions, we don't use it," he said.

"I think there's a balance between innovating and UOB's values of being a trusted bankā€¦ there's no need for us to take these sorts of risks."

Going forward, there are several data-related initiatives that Mr Lowe will continue to press ahead with. Firstly, he intends to make data even more accessible and more understandable for users.

"We are still exploring the mountains of data in our data lake," he said. "What I want to do is actually try and make it easier to access and interpret some of these data for users to consume."

Secondly, the bank will look to implement more AI and machine learning solutions, such as in bread-and-butter areas like credit decisions. This can be done via many different ways, such as through partnerships, where the bank co-develops solutions with like-minded partners.

"What I'm trying to say is that we now have the platform to be able to do all these things," he added. "The opportunities are pretty much endless."