Generative AI, the buddy that OCBC taps to deliver better experiences for customers and employees
Investing in technology is but one piece of the puzzle to create more value for customers and enrich the employee experience, says OCBC group chief operating officer Lim Khiang Tong
OCBC Group, the second-largest financial services group in South-east Asia by total assets, employs approximately 30,000 employees. Behind the scenes, the bank has a team of unsung heroes working tirelessly in the background, alleviating its staff's workload and performing many of the mundane tasks for them.
OCBC Buddy and Wingman are two of the behind-the-scenes assistants well tapped by OCBC staff.
Used more than 30,000 times on average a month, OCBC Buddy is their generative artificial intelligence (AI)-powered chatbot for employees that is able to trawl through hundreds of thousands of internal documents and intranet pages to offer clear answers to queries ranging from Who's Who in the group to block leave compliance.
OCBC Wingman, on the other hand, serves as a coding assistant for the 2,000 developers across the Singapore-listed OCBC Group in their work, helping them to auto-generate, debug and improve their computer code. Productivity of the developers has improved by 20 per cent as a result.
Buddy and Wingman are two of the automated software applications created in-house by OCBC and powered by generative AI to enhance employee experience, boost productivity and improve quality, says OCBC group chief operating officer (COO) Lim Khiang Tong.
Generative AI is also used to power a multilingual minute-taker, leveraging speech to text capabilities to automatically create transcripts and summaries of meetings in different languages.
BT in your inbox

Start and end each day with the latest news stories and analyses delivered straight to your inbox.
The technology - Generative AI - was popularised in 2022 by the success of OpenAI's ChatGPT (Generative Pre-trained Transformer) chatbot, a language model trained to generate conversational text based on context and past data.
OCBC started going big on generative AI around that time and has also rolled out its own OCBC GPT as a pilot trial from April to September 2023, involving close to 1,000 employees across both front office and backend operations.
Staff were excited about the use of the tool and started to become innovative as they became more familiar with GPT, reveals Lim.
OCBC GPT can be used to help with writing, research and ideation. Some examples are: "Help me write a professional speech with the following key points", and "Explain how a technology like blockchain works".
OCBC GPT is being made available to all 30,000 OCBC employees globally in November 2023.
"The feedback has been positive, with productivity boosts of about 50 per cent reported," says Lim, who oversees the transformation efforts in modernising the bank's technology architecture, harmonising processes and instituting a data-driven and customer-centric culture across the group.
The 62-year-old IT veteran with over 30 years of industry experience also oversees the group's operations and technology, technology architecture, operational excellence, customer experience, data and analytics, transformation, corporate services and property management functions.
"Investing in technology is but one piece of the puzzle to create more value for our customers and enrich the employee experience. The way we approach and implement these investments, and bring our people along the way, are equally important," says Lim.
Speed and scalability are essential, says Lim. The group has been able to scale up quickly when generative AI technologies emerged as it has built up its infrastructure and resources in this area over the years.
Document AI, a generative AI-powered tool that extracts key information from documents, was developed in less than three months and rolled out in August 2023. It is particularly useful for OCBC front-line employees who are required to read through multiple documents relating to finance, risk and ESG (environmental, social and governance).
Simply drag and drop and the time required to extract information is cut drastically from 30 minutes to one minute per document.
With a suite of generative AI-powered productivity tools launched in the bank within a year, Lim expects to reap more benefits from these emerging technologies.
"More importantly, as a result of using these tools at work, our staff can focus on higher-value tasks, translating to more productive and happier employees."
Meanwhile, OCBC's arsenal of generative AI-powered tools that are work-in-progress include post-sales surveillance that leverages voice-to-text capability to review all the sales closed over the phone.
As the process involves huge volumes, the group currently only conducts a sampling of such transactions.
With the tool, all telephone sales could be converted into text and any abnormality could be picked up early.
And with the mundane work taken over by generative AI-powered tools, work has become more interesting for OCBC staff.
"No one wants to do the multiple rounds of checking," says Lim, who was named Singapore's Digital Transformation Leader in 2020 by market intelligence firm International Data Corporation for helping OCBC to accelerate digital transformation efforts.
One would imagine that the staff, having benefitted from AI capabilities, would ask to have more bots - or co-pilots as they are more increasingly known - to help them do the heavy lifting. "My data team has received a long list of requests," laughs Lim.
And when it comes to ensuring that AI in general is fair, OCBC has systems and rules to govern the way it designs, builds, tests and deploys an AI model. "We continue to enhance this framework, to ensure it adheres to the FEAT (fair, ethical, accountable, transparent) principles," discloses Lim.
This means that all AI models are reviewed with the accuracy tracked over time. They are also tested to ensure that they are not biased against or unfair to any group, gender or nationality.
Apart from tapping generative AI, OCBC continues to leverage more "traditional" or predictive and analytical AI.
Powering decisions with AI
Now, some four million decisions are made with the help of AI across the group every day, including in risk management, optimising operational efficiency and improvement of customer service. This is projected to rise to 10 million by 2025.
AI is also used to personalise customer service with actionable insights.
OCBC uses AI to trawl through transactions to identify which of its over 120 possible actionable insights may be relevant to the customer. For example, if two transactions take place in rapid succession at the same merchant, the bank would send a prompt to the customer to check if there is potentially a double-charge.
For customers who are earning lower interest in an account, the bank may also suggest that they take advantage of its high-interest-yielding OCBC 360 account.
Every year, OCBC sends out about 250 million of these nudges to help customers to save time and hassle as well as improve their financial management.
Lim says that it is important that OCBC, being a financial institution, is able to apply technology to help the customers better manage their finances.
AI is featured in OCBC's anti-scam efforts as well, flagging suspicious transactions based on the behaviour of the customer that it has learned. The bank helped a customer from falling prey to a scam in April this year, averting a potential loss of S$28,000.
"We are very happy that we are able to help the customer in a situation like this," says the group COO.
Enhancing internal competencies
Technology and data investments are not cheap, but OCBC does not stint on investing in its in-house capabilities. It has been proven that an internal team is able to respond much faster to the needs of the employees and customers. For example, OCBC Wingman - which was based on an open-source model - was up and running in less than two weeks.
OCBC's team of 300 data professionals also work closely with the industry and share source code to improve transparency as well as to promote trust in AI and its appropriate use.
Doing so helps to raise the profile and reputation of the group as a quality employer. This has helped it to attract very good talent from outside the finance sector amid keen competition for this limited field of professionals.
Data analysts and scientists in OCBC are kept engaged and challenged with exciting projects that are not usually available elsewhere. They tend to take ownership and are proud of the applications they develop, Lim has observed.
Also, OCBC is responsive to their needs. For example, the technology professionals are now seated next to the data team, pursuant to such a request. This may be a small change but there has been a great impact because understanding and collaboration between the two teams have improved.
To groom data professionals internally, the Bank collaborated with Ngee Ann Polytechnic to launch a 12-to-18-month Data Certification Pathway programme in 2019. More than 300 staff have since completed this programme. A slew of training modules on generative AI have also been created to raise awareness of this emerging technology among staff.
Meanwhile, OCBC offers AI scholarships for postgraduate students to deepen its pool of AI talent. It also takes in graduates from other disciplines for data roles, as the group sees value in leveraging their diverse backgrounds for different perspectives.
With a vision to become the most data-driven organisation in Asia, and with its investment in AI and people, OCBC looks on track on its journey to enhance the experiences of its customers and employees.
Brought to you by
Copyright SPH Media. All rights reserved.