Those seen as drivers include increasing Internet penetration, exponential growth in computer processing capability, the open source revolution, Big Data analytics capabilities, machine learning capabilities, and scalability of cloud platforms. While these drivers spur the development and growth of AI, a key enabler of AI in the financial services sector is digitization.
Digitization- an Enabler of AI
Digitization includes integration of sensor enabled data acquisition across siloed modules and personalization/customization of digital fulfilment via online and cloud. In the past, limited progress in digitization translated into limited impact of AI on the financial services sector. Realization of this bottleneck has resulted in increased focus on the digitization of financial services. Digitization is now seen as necessary by companies to ensure overall success of an AI strategy and financial institutions are at different stages of completion. Once fully digitized, the financial services sector will reap full benefit from the implementation of intelligent machine learning systems or AI.
Artificial Narrow Intelligence (ANI) is today’s reality
Recent research by Frost & Sullivan uncovered that on the overall scheme of things, majority of today’s AI software resemble Artificial Narrow Intelligence (ANI). Industry progress so far has been within this narrow intelligence scope and is still far away from Artificial General Intelligence (AGI), which is when the intelligence of a machine can successfully perform any human intellectual task. It is estimated that there is a 42% chance of achieving AGI by 2030. The future of AI is said to be one where Artificial Super Intelligence (ASI) is achieved. It is at this point when the capability of machines will surpass humans.
Contrast this with where we are today. Solutions in the market today fall into 3 broad categories namely assisted, augmented, and autonomous. Assisted refers to automation of basic tasks to achieve timeliness and low cost, augmented refers to machines helping humans make more effective contextual decisions, and autonomous intelligence refers to machine taking full control and relieving humans such as in driverless cars. While there are already some applications of autonomous intelligence, they are relatively few. It is hard to justify that we are anywhere beyond the ANI stage.
Challenges Faced in Moving beyond ANI
Within financial services, AI from a technology perspective is an extension of current big data analytics capabilities. Today’s software is built based on machine learning and predictive data analytics. AI within the insurance context refers to software that not only needs to automate, for example, autonomously download and store reports, but also needs to be able to find anomalies in reports and form benchmarks. AI within the banking context refers to software that not only needs to automate transactional roles but also needs to take on advisory and consultation roles. Based on these requirements, the pathway to AI in the financial services sector might be long.
First of all, today’s AI software has not met the cut yet due to lack of soft skills and responsiveness as well as privacy, trust, and regulatory concerns. There might still be instances where there would be need to fall back to a human. For example, live support will still be required if an AI interaction or transaction falters. Second, some tasks still either require a human touch or an actual human to handle. While AI has the potential to replace humans in some functions, for example, as personal assistants and in machine learning, some other functions will still require human intervention such as robo-advice. For now at least, the financial services sector will be served by assisted and augmented intelligence.
Investment in AI and its Impact on the Financial Services Sector
Investment in AI by financial institutions has seen Big Data analytics and machine learning feed the data intensive AI to enable automation and a personalized experience. In the near term, there will be increased use of AI for businesses to aid decision making and mitigate risks with predictive insights. There will also be increasing use of behavioural pattern analysis for identification of anomalies and fraud attempts. In the longer term, Neuro Linguistic Programming (NLP) is expected to go mainstream, as financial institutions increasingly use voice processing to offer services and assistance.
The biggest impact of investment in AI by financial institutions was highlighted by Citigroup in a recent study. According to Citigroup, automated banking will lead to the replacement of an estimated 30% of bank job over the next decade. While financial institutions are investing to save cost as revenues shrink, Fintech companies have been creatively carving out niches to increase revenue. Finie and Trov are some examples. Finie uses data driven AI to further the chatbot concept beyond the traditional rule based approach. Trov introduced the concept of on-demand insurance. As a reactionary measure, financial institutions have established in-house venture capital arms to invest in startups. The way forward seems to be one of either complementing or collaborating.
AI in Regulatory Compliance
From the regulatory compliance standpoint, AI and machine learning are expected to be mainstream technologies for regulation by 2020. In line with these expectations, regulatory bodies have started to make the necessary changes. For example, the CFA Institute recently shared that it is updating its Certified Financial Analyst (CFA) exam so that questions from 2019 will cover AI, Big Data, and robo-advice. This change reflects the growing impact of intelligent machine learning systems on the financial services sector. Governments and policymakers in the Asia-Pacific region are supportive of technology innovation. Regulatory sandboxes have been set up across Asia-Pacific to cultivate growth of Fintech solutions.
From the financial institution’s perspective, use of Big Data in RegTech solutions will allow for more efficient and better audit trails and cloud computing will enable agility in a RegTech solution as updates can be recorded easily. However, there is hesitation amongst financial institutions to invest in a particular RegTech solution, as the regulations for which it provides compliance could change and result in low returns on investments.
Key Takeaway
The necessary key enabler for AI in the financial services sector is currently falling into place, with financial institutions still at different stages of the digitization process. Based on end requirements, the pathway to AI might be long. Overall, it seems certain that AI will play a key role in the future of the financial services sector but for now, it is still assisted and augmented intelligence actively transforming the sector.
