FinTech & Artificial Intelligence: Fiction or Becoming Fact?

ASEAN countries capitals

By Edwin Yapp, TRA ASEAN Executive Consultant

In 2015, investment banker Jamie Dimon famously wrote in his letter to shareholders that "Silicon Valley is coming." The CEO of JPMorgan was referring to how innovative startups, particularly in the field of FinTech, are charging hard towards what the financial service industry (FSI) considers its core business, warning that “they plan to eat traditional banks’ revenue stream.”

A year later, Dimon dialled back his doom-and-gloom expectations, saying that his bank was more ready to compete with Silicon Valley FinTech startups, noting that they are on equal footing with them because JPMorgan has heavily invested in FinTech.

So which is it? Is FinTech as disruptive as people make it out to be? Or is this trend just overhyped? The truth is probably somewhere in between. Based on Tech Research Asia’s (TRA) research into this area this past year, it is our belief that FinTech is not yet about to bring the curtain down on traditional banking as we know it. But at the same time, the FinTech revolution isn’t all hype either. There are clear areas in which FinTech can help liberalise a traditional business sector that is notoriously famed for being slow, risk averse and overtly process and bureaucratically driven.

In a study on FinTech we conducted in 2016 of 3 Southeast Asian nations, the key takeaways were that while FSIs were once worried about the rise of FinTech startups (about 5 years ago), they have begun to realise it's not really about disruption. Conversely, FinTech startups will not necessarily eviscerate many banks’ core businesses too. Case in point: Our survey indicated that, 71% of organisations in Indonesia intend to invest in online payments this year while approximately 40% will deploy mobile wallets. In Malaysia, 30% of organisations are the most likely to adopt crowd funding this year. And in Singapore, 33% of firms expect to adopt contactless payments in 2017.

What we find is that these two groups are not necessarily diametrically opposite each other as the “disruption mantra” of the last few years would have you believe. Rather, they are collaborating and must continue to do so in order for FinTech solutions to achieve widespread impact. In other words, FinTech in 2017-18 is very much a status quo component of contemporary financial services. The reason for this lies in the fact that while FinTech players may be nimbler, more innovative and form ecosystem communities easily, they are small, do not have scale, and struggle to build reputation and customer bases like the FSI organisations traditionally have. Put simply, FSIs and FinTech startups can co-exist together, and this is validated by other industry research studies.

While FinTech in 2017 may still be considered nascent especially in Southeast Asia, the sector is picking up fast. Driving this change is the emerging use of artificial intelligence (AI), which when working alongside FinTech is expected to create even more powerful solutions.

For this article, TRA defines FinTech loosely as either existing players in the FSI branching out into new services as well as new providers, often start-ups, which use technology to offer financial services-related technology either to end customers or to FSI providers

Drivers of FinTech

There are many factors driving FinTech especially in ASEAN (The Association of Southeast Asian Nations). While not exhaustive, here are some factors:

  • Home to more than 600 million people and with a combined GDP (gross domestic product) of about US$2.6 trillion, ASEAN continues to be one of the most significant areas of the world in terms of economic activity and trade between nations. This is especially true with the forming of the ASEAN Economic Community (AEC), as the region will comprise a single market and production base, a highly competitive economic region, a region of equitable economic development that is fully integrated into the global economy – all of which will help the FinTech industry to grow;

  • There are millions of untapped unbanked in ASEAN. The World Bank notes that there are 264 million adults who are still unbanked, which is a potential harvest field for FinTech initiatives;

  • The exponential growth of smartphones and mobility in Southeast Asia, which according to the ASEAN Secretariat, means that the region has over 100 million units of smartphones and growing. Of this about 30 million are capable of operating on advanced 4G, which bodes well for FinTech companies:

  • Savvy millennials – defined as those born between 1990s and the new millennium – dominate the population in ASEAN with over half the 629 million people (and counting) being under the age of 30. This represents a market filled with young people who prefer to interact with the FSI digitally rather than physically; and

  • Consumers today are much more focused on brands, including banks due to the continuing digitisation of business . With the rise of FinTech companies and new consumer expectations, customer centricity will be the key differentiator.

Artificial Intelligence in FinTech

So where do artificial intelligence and FinTech intersect? First coined by computer scientist John McCarthy in 1956 at The Dartmouth Conference, the term AI has only experienced significant progress in the last two decades or so. But what is AI? And how is it different from a few other connected terms, such as machine learning, deep learning and neural networks? Below is a general definition of these terms.

  • Artificial intelligence: This term broadly describes how a machine, specifically a computer in concert with advanced software algorithms, is programmed to behave and react the way humans do. The goal of AI is to understand tasks and apply solutions that involve self-directed learning, reasoning, perception, and problem-solving, just like how humans are able to. The areas that AI intersects with involve a cross-disciplinary approach that is based on mathematics, computer science, linguistics, psychology, and other behavioural disciplines.

  • Machine learning: This is a subset of AI and focuses on the ability of computers and software that is able to learn without being explicitly programmed, or in other words, self-adaptive learning. Machine learning gives computers the ability to change their algorithm automatically when exposed to new data, and is particularly good at being used to handle repetitive functions, crunch large data sets and spot anomalies in patterns.

  • Deep learning: Yet another subset of machine learning, deep learning is an advanced approach to machine learning. As opposed to normal computer programming which treats problem solving in a linear fashion, deep learning treats problems in a multi-layer, hierarchical method. When a problem is fed with raw data in deep learning systems, it goes through a process that creates certain “answers.” These answers are then passed on to another layer for analysis by a second layer in the hierarchy, which then yields another set of answers to be processed by the next layer, and so on. The multi-layer processing continues across the network until the best output is determined. Because this essentially mimics how humans think, deep learning is also known as neural network learning or deep neural networks.

As far as FinTech is concerned, the field of AI is still pretty nascent. TRA believes that there isn’t a universal way of classifying the use of AI in FinTech as yet - things are moving too quickly (although algorithmic trading is one established area that uses forms of AI, most other areas are still developing and branching into new fields). Perhaps a better way is to look at how AI is being used to perform certain tasks. For the purpose of this discussion, TRA would like to consider how AI is used in four areas, which are: Enhancing customer experience; collating and analysing customer behaviour; detecting fraudulent financial transactions & improving regulatory compliance; and improving efficiencies and reducing operational costs of financial services companies.

Enhancing Customer Experience

The FSI has begun deploying automation in customer service in earnest in the past 2 years or so. More commonly known as chatbots, these are systems that utilise software that have been coded to simulate human conversations or interactions that take place on devices and robots. Chatbots can deliver human-like responses and can typically simulate basic conversations through automated and structured responses that is triggered by an input when someone “talks” to it. They can also be interfaces (physical or virtual) that provide intelligent interactions and continue to improve based on more experience and better data input.

Chatbots differ from the age-old IVR (interactive voice response) systems used by many banks today in that customers can interact with the former using natural language much like how one interacts with Apple’s Siri or Google’s Assistant without having to go through the tedious options menu which IVRs force users to go through. The goal of chatbots are usually two-fold, that is to generate sales funnels and to enhance customer satisfaction. Examples of financial companies that have deployed such systems include Capital One, Bank Of America, Mastercard, Royal Bank of Scotland and DBS Bank in Singapore.

Capital One example of a chatbot.

Some of the functions customers can expect using such chatbots are: Generating of customers’ notifications such as bank balance, recent transaction, payment history and credit limits; identifying areas where customers can save money through personalised advice; facilitating bill payment, checking of interest rates; and conducting money transfers, to name a few.

A close connection to chatbots are robo advisors or virtual financial assistants. Essentially an extension of the basic aforementioned chatbot functions, robo advisors go further by actually providing financial advice using complex software algorithms that are automated on a digital platform with little human intervention. Some common advice robo advisors dispense include investment planning, retirement planning and tax-loss harvesting. Some advantages of robo advisors include lower commission fees and 24/7/365 capabilities compared to human advisors; lower capital requirement to invest thereby making investment more accessible to the masses; and transactional efficiency via digital platforms.

Some better known examples of robo advisors are Betterment, Wealthfront, Chloe, Smartly, and Theo

Collating and analysing customer behaviour

One of the most frustrating challenges in the FSI world today is to know what their true customers’ true needs are and how to meet each individual’s needs in a more intelligent way. For instance, old methods employed by most of the FSIs such as electronic data mailer (EDM), outbound telemarketing calls, and snail mail mailers just aren’t effective anymore and also incur costs to banks. Unfortunately many FSIs still use these methods due to the fact that the incremental costs of doing so is still low compared to the potential of winning over a customer. To compound this problem, many FSIs aren’t using real time data but instead use batch data. Also, much of their customer data is in disparate siloed locations instead of being accessible in an open way.

Dated methods as outlined aren’t helping and one area adjunct to improving customer satisfaction that AI promises is to better analyse current customer behaviour. In fact, these dated methods are harming the FSIs because often, they have not taken into account the frustrations faced by customers blasted with irrelevant advertisement of products which do not meet customers’ needs. The opportunity costs have not been taken into account. Instead of using such methods, forward thinking FSIs have begun to use real-time, advanced data analysis to analyse the thousands of customer data to assess their risk profiles, spending habits, investment and remuneration potential to come up with cross-selling opportunities.

Using machine and deep learning methods, banks and insurers, for instance, can constantly analyse thousands of customers’ information as they happen and be able to design personalised financial products and recommendations that are tailor-made to each customer’s profile to enhance its customer profiling instead of utilising the passe method of ‘hit-and-miss’. For example, through AI-powered analytics, a bank may notice that customers could get a higher return on their money by moving some cash to a different account. Or when a customer changes his job, the system is able to detect his change in income levels, spending pattern and cross-sell new financial products to him. Or when the said customer purchases a new car, AI analytics can recommend better insurance plans that are tailor-made to her new income levels. All of these moves will definitely add value to customers. On top of this, banks and insurers can use such information in conjunction with other technologies such as chatbots and virtual financial advisors (robo advisors) to create a market differentiator against their competitors.

Fraud detection and anomaly analysis

One of the strongest propositions for AI is that systems can be programmed to detect anomaly in on-going patterns, particularly when it comes to huge amount of raw data. For example, AI systems are being used today to detect anomalies in irregular and fraudulent spending in credit cards, as well as when customers have their credit cards stolen or lost; monitor financial markets malpractices (insider trading), spotting irregularities in market movements that affect stock portfolios, and detect money laundering activities. AI is perfect for this because of the nature of its processing speed and self-learning capabilities, which can adapt to spot undiscovered scenarios through pattern recognition over the course of time.

Connected to this is that compliance and regulatory framework can be enhanced because a well implemented AI system can reduce the probability in spotting errors or anomalies which would otherwise not be done by humans. Taking out humans from the process also helps prevent any potential unwanted corruption in regulatory and compliance frameworks.

Improving efficiencies and reducing operational costs

While the prospect of using AI systems in FinTech may be daunting from a cost perspective as financial companies would need to invest heavily upfront, the total cost of ownership is likely to make sense over the course of time. By leveraging AI in operational settings and passing the burden of handling huge amount of data in a small period of time in an automated manner, financial companies can speed up processing times and reduce the overall handling costs while enhancing customer experience.

According to Goldman Sachs, machine learning and AI will enable $34 billion to $43 billion in annual "cost savings and new revenue opportunities" within the financial sector by 2025, as institutions use technological advancements to maximise trading opportunities, reduce credit risk and lower compliance and regulatory costs. Separately, The Bank of New York (BNY) Mellon has developed and deployed automated computer programs, or more than 220 "bots", across its businesses over the past 15 months seeking more efficiency and lower costs, as the adoption of artificial intelligence technology in banking increases. BNY estimates that its funds transfer bots alone are saving it $300,000 annually, by cutting down the time its employees need to spend on identifying and dealing with data mistakes and accelerating payments processing.

While these figures are only indicators of the potential AI has for costs savings, the fact is that it’s difficult to estimate how much FinTech can save financial institutions as there are so many variables in play. What’s certain however is that the FSI cannot and should not ignore the impact of AI and its associated technologies as in time to come, the use of AI in FinTech will be the new normal rather than the exception.

Recommendations

Although AI in FinTech is relatively new, make no mistake about it: There are already use cases for the application of such technology in the FSI. Like many other technological developments, the gestation period for technology is pretty short and once start-ups and the FSI as a whole are able to provide differentiated offerings to the market, AI in FinTech will take off quickly. As such, TRA offers the following questions to ask as a possible business or government user:

1. Have you done a frank evaluation of firstly the role FinTech and AI could play in your organisation and the strategic ways you can leverage this trend?

2. Have you considered formulating strategies for collaboration in FinTech with start-ups that can help you achieve your goals?

3. Have you optimised the technology foundations on which your FinTech and AI strategy might be founded and ensured that it is adaptable to future opportunities or challenges?

4. Have you examined whether your FinTech and AI vision and strategy focuses on technology or the people it intends to support or the outcomes you want?

5. Have you calculated the risk to your organisation of not adopting emerging tech like those that come under FinTech?

6. Would you be able to fast track the outcomes of your FinTech deployments by leveraging the technology and experience of external partners?