According to Global Market Insights, Artificial Intelligence (AI) in the Banking, Financial and Insurance (BFSI) Market is estimated to be worth over USD 2.5 billion in 2017 and is anticipated to grow at a Global CAGR of more than 30% from now through 2024. Not surprisingly, the Asia Pacific region driven by China is leading the way with an estimated CAGR of over 40%. AI has applications that vary widely in finance - from cost savings to improving customer experience and fraud detection - and right now there are already 2.5 million U.S. financial services workers whose jobs are directly impacted by AI.
Finance was one of the first sectors to embrace AI. Sydney Swaine-Simon and Abhishek Gupta write: “The financial sector is one of the first domains to drive interest in using artificial intelligence, even before high computing machines were available. In the 1960s, a lot of research focused on Bayesian statistics, a method used heavily in machine learning. Some of its use cases included stock market prediction and auditing. It wasn’t until the 1980s until the majority of commercialization opportunities were explored with expert systems. During that time, over two thirds of Fortune 1000 companies had at least one AI project being developed.”
However, with all the current enthusiasm about AI, one Bloomberg headline caught my eye this past September: “AI Hedge Fund Is Said to Liquidate After Less Than Two Years”. Although not confirmed by the hedge fund, this story nevertheless seems to create a crack in what seems to be an otherwise very positive slew of headlines on the “$1 Trillion Opportunity” of AI in finance with its data about AI hedge funds “disappointing investors this year”. I got in touch with several finance, fintech and digital transformation experts to ask them what they thought of the story including Christina Qi, Jim Marous, Kunal Patel, Spiros Margaris, Sally Eaves and Theodora Lau. Opinions varied and there were strong AI pro and con thoughts, but as emerging tech strategy expert Kunal Patel reminds us “there are always limitations, challenges and concerns with anything new.”
Christina Qi, Hedge Fund Partner, Domeyard LLC, Forbes 30 Under 30. @ChristinaQi
“There’s this huge revolution of AI in finance – everyone is so excited about it - but the truth is that very few firms are successful employing AI in finance. AI is exciting, and we've seen a lot of theoretical explorations of AI in finance – but we haven't seen many cases of AI being put in practice at a large scale in our industry.
AI usually depends on a large amount of data and that can be tricky.
A lot of hedge funds use ‘alternative data’ now – for example satellite data that can monitor Walmart parking lots to estimate sales. We can predict how much oil is in a country at a certain time by monitoring the height of oil rigs, or the amount of cargo in cargo ships. Then there are of course funds that monitor Twitter, Facebook or LinkedIn posts. There are vendors out there selling access to this data and there are funds that extract the data internally. Alternative data sounds appealing, but in practice it's difficult to employ these data sets in finance because their relationships to market returns are weak and noisy, and the quantity of data is very limited. We've seen very few funds finding success from monitoring social media, for example. The edge is slim.
Another problem that comes up, if vendors are selling this data to everyone, evening the playing field, the potential profits go down because everyone will have access to the same data.”
Jim Marous, Financial Industry Publisher, Advisor & Global Speaker; Co- Publisher, The Financial Brand; Publisher, Digital Banking Report. @JimMarous
“We have researched this subject extensively at the Digital Banking Report and we find that most organizations talk about the application of AI much more than actually doing it. While there are strong case studies around the ability to cut costs and reduce risks, this isn’t where the focus should be. Certainly not limited to the way banks are using it today.”
Kunal Patel, Product Manager Strategy & Emerging Technologies, 1E. @KunalPatel805
“Like it or not AI is a good thing. It’s worth noting that some of the advances in AI can help financial institutions be more efficient, provide better cost savings operationally and provide a better customer experience in the long term. However, there are always limitations/challenges and concerns with anything new.
The main limitation of adopting AI based technology is the cost. Whenever you venture in building something new and revolutionary within the tech space, there are a bunch of complexities which spring up. Moreover, the costs can spiral out of control given the need for continual maintenance in-life.
These complex programmes need regular tender loving care, upgrades and being able to adapt to business changes are a must, especially when financial intuitions are so obsessed with digital transformation. There is also the risk of failure from a system perspective, given the nature of the financial services market, this needs to be managed carefully.
Implementation times can be costly, integration and a lack of knowledge in this area can also be a hindrance, along with interoperability considering the plethora of systems that exist today.
Financial institutions need to consider privacy, transparency, tech complexity and potentially business decisions/strategy considerations with regards to loss and control.
Other potential roadblocks and challenges will be the risk to back and front office workforce, given the advantages of automation, which is a threat across all industries. The knock-on effect is a plan to retrain and redeploy employees and the negativity which won’t go away so easily.”
Spiros Margaris, Venture Capitalist, Founder, Margaris Ventures; Ranked Global No. 1 Fintech, AI, Blockchain & No. 2 InsurTech Influencer by Onalytica. @SpirosMargaris
“AI and machine learning in banking and the financial industry as a whole is the only way forward for any financial institutions that wants to stay competitive. How the financial companies mix the human advisor in the overall service package will be at the end just a choice of how they want to differentiate themselves from their competition. For me, any negative news about AI and machine learning is not stopping me believing in the AI technology as one of the key business drivers for the finance industry and the trend towards even more personalised customer offerings.”
Theodora Lau, Founder, Unconventional Ventures; Advisory Board Member, Envel. @psb_dc
“Indeed AI is not a magic bullet as it has been hyped up to be. It will not solve all of the world’s most difficult problems – nor will it become sentient and destroy humanity. However, that is not to dismiss the promise of AI.
To be effective, however, new technology innovation has to move beyond efficiency and incremental value creation. It is not about being able to pay for your personalized latte 15 seconds faster or get historical reference on your spending via a smart speaker; nor should it be about machines replacing humans. The true value of AI and other emerging technologies in financial services lie in their ability to create new customer value. How can we propel the next generation to greater financial security and help them make the right decisions at the right time? How can financial institutions build trust and empathy to create a deeper, more emotional connection between brand and customer? And how can technology be leveraged to create a more sustainable and equitable society for all?
The future is full of unknowns; the technology is there to help us solve some of them. The question is – are we willing to move beyond the low-hanging fruits and take on the responsibility for our future selves?”
Sally Eaves, CEO Sustainable Asset Exchange (SAX), Forbes Technology Council and Professor of Advanced Technologies. @SallyEaves
"Within the Finance function and beyond, AI technology and related disciplines (machine learning, predictive analytics, natural language processing, and interactive bots) can be leveraged to add greater value to the business. Benefits span the back office through to front office; including exceptional improvements and impactful increases in operational efficiency, supervisory effectiveness and fraud identification through to customer engagement. This is most evident in personalization of service opportunities through tailored, timely, and tantalizing service offers and product recommendations.
Through the application of assisted, augmented, automated, and autonomous intelligence, each use of AI within Finance creates different types of risks in addition to the positive benefits that result from the application of these capabilities. Innovation must always be considered in tandem with the new risks that accompany the opportunities created. Lack of awareness is one key factor to consider – how many consumers will fully understand what AI is and how it effects their everyday lives? Further, the effectiveness and efficiency of AI in any field is dependent upon the volume, quality, and timely data upon which the AI models are trained.
More specific risks are related to performance, control, security, ethics, economic, legal, and societal risks. These have the potential for significant organizational impact - both financial and reputational. Additionally, the talent gap with respect to AI skills availability vs. the growing demand for these skills represents another risk. Achieving a diversity of perspectives within AI development teams remains both a roadblock and concern. As adoption becomes more mainstream, these gaps can only be expected to widen.
Indeed, it is this human perspective that will come to the forefront, especially within Finance. When AI is used to create or support a decision, for example a decision to accept or reject a mortgage or loan application, the end user will not know how this decision has come about. These can be decisions that affect the welfare of individual clients – decisions that can change and affect their lives and livelihoods. Further, from an EU regulatory lens, the General Data Protection Regulation (GDPR) provides the user or client the ‘right to explanation’ around such decisions. How many organizations are in the position to readily comply? So, whilst AI can be employed to accelerate the production of quality nuanced insights and informed decision contributing to greater business success, the use and application of these insights and decisions must be explainable in plain English and must be fully justifiable to those it effects the most."
If I tally up the pro and con opinions here, I’d have to say AI pros win with this group. So make sure you are ready for the AI disruption in the financial services with our Machine Learning and AI in Finance course taught by Petros Geroulanos happening this December in London! Geroulanos is a visiting lecturer at SKEMA Business School in Paris and a FinTech consultant with over 25 years of professional experience in trading, product development and sales. As former head of VEGA Structured Finance GmbH in Stuttgart, Germany, he introduced default-free ABS to medium-sized corporates using data analytics. Geroulanos has worked for Swiss Bank Corporation, Union Discount PLC in Zurich and London – where he was engaged in trading, selling and structuring FX, Equity Exotic and Fixed Income Derivative Products as well as their underlying cash instruments.