How the Most Powerful Financial Institutions in the APAC Region Are Applying Artificial Intelligence

AI could produce economic value – created from the introduction of new product services and categories, cost savings arising from better products, lower overall prices, and improvements in lifestyles – between $1.8 – $3 trillion a year by 2030 in Asia.

AI will have a disproportionally high impact on the industries of financial services, healthcare, manufacturing, retail, and transportation, which combine to contribute around two-thirds of Asia’s GDP currently. – Artificial intelligence and Asia, UBS

While the two largest economies in the world are dominating global research and development in the artificial intelligence (AI) field, Credit Suisse expects China to emerge the winner.

AI is going to redefine all aspects of banking – from the way we interact with our customers digitally, to the way we help them make faster and better financial decisions. The impact cannot be ignored. The time to act on AI is now! It is integral to digital banking, or what we call ‘the new digital.’ We have already enjoyed tremendous success with AI in the past two years – working with FinTechs to commercialize AI solutions for transaction monitoring and chatbots – and I believe there is a massive competitive advantage to be gained from being a first mover in this space. AI is going to break the Internet and be bigger than the mobile revolution. We want to be at the forefront of this new age of banking. – Pranav Seth, SVP, Digital & Innovation (E-business, Business Transformation, and FinTech & Innovation Group), The Open Vault@OCBC

Financial institutions around the world seem to be most quick in bringing AI into the front-office, with such deployments as credit scoring, insurance, and client-facing chatbots. A MEDICI study has found that virtual assistants/chatbots are the most ubiquitously explored across regions. Among other use cases of AI in Asian banking sector are process automation, fraud prevention, compliance, advisory, and underwriting. Let’s look at some examples:

Virtual assistants, chatbots

The low-hanging fruit for AI within the Asian financial services industry is to leverage virtual assistants, chatbots or speech recognition software for regular customer interactions, thereby lowering the dependence on traditional banking channels like branches. Asian banks should get a major boost to risk management, which is traditionally a weak spot, as AI can help better manage credit-risk assessments and anti-money laundering programs. In the longer term, as robot advisors become more sophisticated, Asian banks can further utilize the technology in product marketing and after-sales. – Shifting Asia: Artificial intelligence, UBS

One of the largest banks in the Pacific has launched a chatbot app that helps customers with more than 200 banking tasks such as activating their card, checking the account balance, making payments, or getting cardless cash.

Another institution launched a digital virtual banker specifically for business customers, enabling them to receive instant answers and assistance with common banking questions and tasks. The banks’ virtual banker is in pilot mode and available 24/7, providing help with more than 200 common questions related to the servicing of business banking accounts.

The virtual assistant’s AI is derived from thousands of real-life customer inquiries. There are more than 13,000 variants of the 200 questions the virtual banker can answer; if the question can’t be answered, the customer will then be directed to a human banker. What’s important is that customers were involved in the testing and development phase.

Meanwhile, one of the largest banks in Southeast Asia is using an AI-powered virtual assistant to enhance the experience at its mobile-only bank in India. An assistant that continuously learns can understand language the way humans speak it, helping the mobile-only bank to anticipate and reply to thousands of customer queries, and customers to fulfill banking transactions in real-time, at any time, anywhere.

Furthermore, one of the largest institutions in Southeast Asia has launched AI-powered voice banking in collaboration with Google. Anyone can use the service by speaking to the Google Assistant – on a smartphone or a Google Home device – to initiate a conversation about the banks’ services. These services range from planning for retirement or a new home to saving for a child’s education, getting the latest financial market updates, and more.

While some institutions use own resources to develop assistants, others find tech partners to power the function. YES BANK & PayJo, HDFC Bank &, HDFC Bank & OnChat, Axis Bank &, and OCBC & CogniCor, for example.

YES BANK has also partnered with Gupshup to launch YES mPower, a banking chatbot that helps customers get information about loan products offered by YES BANK and instantly gives loan eligibility. YES BANK plans to incrementally release more services through the chatbot, making YES mPower the one-stop shop for all loan requirements.

The State Bank of India is another interesting case: The Silicon Valley-based startup Payjo has launched an Intelligent Assistant for SBI, one of the world’s largest banks with 420 million customers. SBI’s Intelligent Assistant or SIA, an AI-powered chat & voice assistant, answers customer inquiries instantly and helps customers with everyday banking tasks. SIA is set up to handle nearly 10,000 inquiries per second or 864 million per day – close to 20% of Google’s worldwide traffic. SIA continuously learns with each interaction and gets better over time. Currently, it can answer inquiries on banking products and services and will soon be able to process financial transactions in many Indian regional languages across multiple customer channels in both voice and text format.

At the end of 2017, United Overseas Bank (UOB) unveiled its first two robots, Amy and Eve, to be assigned to teams supporting its wholesale banking and retail arms. The AI-powered assistants were set to perform repetitive and time-consuming tasks so that it frees up human workers for more complicated functions. The company shared that what used to take up to 240 seconds, 12 screens and multiple cross-checking for accuracy took Amy only 40 seconds to complete. Meanwhile, Eve has the ability to handle up to 1,800 applications per day and does it 3.5X faster than others previously assigned this repetitive task.

Fraud, security

AI-powered risk management for financial institutions is embodied in advanced fraud prevention and AML solutions as well as more accurate customer assessment. However, AI has a more defining, and more rarely discussed impact – stress testing result submissions and the adjustment of capital requirements for institutions.

Regulatory bodies are equally interested in adopting advanced technologies to tackle the issue. The Financial Stability Board (FSB) recently published a report sharing that some regulators are using AI for fraud and AML/CFT detection.

The Australian Securities & Investments Commission (ASIC) has been exploring the quality of results and potential use of NLP technology to identify and extract entities of interest from evidentiary documents. ASIC is using NLP and other technology to visualize and explore the extracted entities and their relationships. In order to fight criminal activities carried out through the banking system (such as money laundering), BdI collects detailed information on bank transfers and correlates this information with information from newspaper articles. The correlation involves both structured and unstructured data for file sizes of more than 50 gigabytes.

ASIC has also used machine learning (ML) software to identify misleading marketing in a particular subsector, such as unlicensed accountants in the provision of financial advice.

FSB also shares an example of the Monetary Authority of Singapore (MAS) exploring the use of AI and ML in the analysis of suspicious transactions to identify those transactions that warrant further attention, allowing supervisors to focus their resources on higher-risk transactions.

Investigating suspicious transactions is time-consuming and often suffers from a high rate of false positives due to defensive filings by regulated entities. Machine learning is being used to identify complex patterns and highlight the suspicious transactions that are potentially more serious and warrant closer investigation. Coupled with machine learning methods to analyze the granular data from transactions, client profiles, and a variety of unstructured data, machine learning is being explored to uncover non-linear relationships among different attributes and entities, and to detect potentially complicated behavior patterns of money laundering and the financing of terrorism not directly observable through suspicious transactions filings from individual entities.

One of the largest institutions in Australia is developing AI technology to help its cybersecurity, fraud detection, and regulatory compliance functions.

At the end of January, HDFC Bank had completed a pilot for AI-based Cyber Security Operations Centre (CSOC) and the bank is slated to go live with it soon. The log data from CSOC is to be sent for processing on the AI solution having big data capabilities and it was done for around eight months on a cloud platform. The bank has close to 100,000 employees and the AI solution is expected to help in monitoring insider threats.

Partnerships are not unique to the goal of developing a virtual assistant. OCBC Bank, for example, partnered with ThetaRay to use its AI solution to identify potentially suspicious transactions. The solution has reduced the volume of transactions reviewed by AML compliance analysts by 35%, and increased the accuracy rate of identifying suspicious transactions by more than 4X.

Automation, process optimization, predictive analytics

ICICI Bank has deployed software robotics functions across the organization including retail/wholesale/agri banking operations, treasury operations, human resources management, international remittances, and private banking. The software robots are configured to capture and interpret information from systems, recognize patterns, and run business processes across multiple applications to execute activities including demographic data validation & updation, data formatting & consistency verification, multi-format message creation, text mining, workflow acceleration, and reconciliations. A few of the AI technologies that the platform uses are natural language processing, ML, cognitive tools, optical character recognition, and data analytics.

New Zealand’s largest financial services group implemented RPA across its four global delivery hubs. Rapid implementation in smaller units by decentralizing the adoption process enabled them to achieve quick ROI while reducing costs by 40% and reducing end-to-end delivery time.

A large Singaporean institution is working with IBM to scale an enterprise-wide Centre of Excellence (COE) in Robotic Process Automation (RPA), making it the first-of-its-kind large-scale implementation in the financial services sector in Singapore and the region. IBM helped the bank set up a COE in RPA in June 2017; by December 2017, the bank had applied RPA to 50 complex business processes across its operations, freeing up 25,000 man hours. Now the RPA program will be progressively implemented in other markets, including Hong Kong, China, India, Indonesia, and Taiwan.

Axis Bank, India’s third-largest private sector bank, has implemented AI across 125+ processes and cognitive automation across 90 processes to reduce manual labor and turnaround time.

Another large private sector bank in India has a particularly interesting application: the bank has partnered with Anytime Loans, an automated P2P lending platform, which uses AI to read facial features and feed it into a predictive model that determines the borrower’s propensity to default. The company has disbursed $10.18 million over the last 30 months to 38,700 entities for personal loans, business loans, and education loans. Its gross defaults are at a low 0.6%.

Authentication, underwriting, advisory, and more

While at a lower magnitude and scale, financial institutions in the APAC region are also exploring the applications of AI beyond the front office and mentioned earlier areas. Those applications include customer authentication, underwriting, advisory services, compliance, and branch service assessment.

ANZ, for example, has added a voice ID function to its Grow by ANZ app, allowing customers to authorize fund transfers of more than $1000 and BPAY payments of more than $10,000 without a PIN or password. The bank intends to roll the technology out to other digital channels. The service uses technology from Nuance, which also provides voice biometrics technology to the Australian Taxation Office.

Speaking of biometrics, India’s largest bank has developed an AI-powered solution which scans cameras installed in branches and captures the facial expressions of customers and immediately reports whether the customer is happy or sad. The bank plans to build a dashboard using the solution that will gauge the effectiveness of representatives.

Meanwhile, the banks’ wealth division is exploring AI opportunities to improve its underwriting processes. The bank has collaborated with University of Technology Sydney’s Advanced Analytics Institute to develop underwriting model that will harness ML to provide an opportunity for insurers to develop more efficient and reliable assessment processes.

In the advisory function, ANZ used IBM Watson to help its financial advisors understand its clients better by providing tailor-made services. The bank is now expanding its use into areas such as risk and back-office automation.

As for compliance applications, an Australian institution has used software that uses NLP and AI to convert 1.5 million paragraphs of MiFID regulation into actionable tasks as a pilot program.

With exponential growth in computing power and favorable supply-side factors, like the advent of advanced algorithms, a vast pool of indigenous AI-related talent and massive government-funded infrastructure development, both the breadth and depth of AI adoption in Asia is set to sharply accelerate in the coming decade. – Min Lan Tan, Group Managing Director, Head APAC Investment Office, CIO Wealth Management, UBS

I think Asia is extraordinarily well-prepared to lead in the Cognitive Era, and here’s why: the pioneers of this era, the people that are actually bringing cognitive solutions to different industries, are the app developers, the data engineers, and scientists, and Asia has these skills in spades. India is now home to one of the largest populations of developers in the world.

Just as important, many countries in the region, like Singapore, Australia, and Korea, are investing in initiatives to support their developer communities and the broader cognitive economy. In fact, last year, the Korean government announced it would spend KRW 1 trillion (about USD 884 million) by 2020 to boost the artificial intelligence industry, which is part of the reason why Watson is now ‘learning’ Korean. – Dr. John E. Kelly III, SVP, Cognitive Solutions & Research, IBM