The investment-management industry is undergoing arguably its most disruptive period ever. Thanks to a new wave of disruptive technologies, the very concept of investing is being transformed from a practice that was relationship-driven and only accessed by the affluent few to a democratised, inclusive activity that serves a much broader customer base. Today, the demand for digitised, convenient financial services has soared, and, as such, the opportunities to participate in exciting investment opportunities can be achieved with just a few taps of a smartphone.
But of all these game-changing technologies, it is perhaps artificial intelligence (AI) that has the capacity to most transform—and most improve—the investing process. AI typically refers to using computers to simulate intelligent behaviour that’s comparable, or even superior, to that of human beings. And when one looks around, it’s clear that those computers have not only arrived but are now changing the world in dramatic ways. Whether it is through robots performing life-saving medical and surgical procedures, AI-powered chatbots resolving important personal-banking issues or self-driving cars minimising the number of accidents on the road, AI is increasingly impacting our everyday lives—and invariably doing so in a beneficial fashion.
The same goes for investing, too. Already we are seeing automated systems providing a more significant contribution within the investment process, particularly as their abilities to outperform humans in the markets keep growing. And now, investment firms are employing AI to make faster, smarter and more profitable investing decisions. “Investing is the ultimate numbers game, and smart number crunchers tend to be good at it,” observed Daniel Seiler, head of the Multi Asset Boutique at Vontobel Asset Management. “So, artificial intelligence as a high-capacity data processor stands a good chance at revolutionizing the investment industry.”
Machine learning (ML) is arguably the most exciting aspect of this revolution. An application of AI, ML involves utilising data to learn, adapt and improve investment decisions without needing to be explicitly programmed to do so. By formulating various algorithms and then exposing them to substantial volumes of relevant data—such as historic market prices and transactional data—ML systems can be trained to quickly identify security mispricings and market inefficiencies. And as a result, they can detect more potential opportunities to generate alpha, which ultimately translates into a higher chance of investors and fund managers making money.
Cultivating this automated environment is easier said than done, of course. For machine-learning techniques to be successful in predicting future market opportunities, they invariably require large amounts of data, or big data, alongside significant computing power and abundant data storage. But thanks to a gradual decline in the cost of storage, as well as the progress now being made in expanding processing power, investment managers increasingly have the capabilities required to explore data-driven strategies. ML can then be used to extract useful, actionable real-time information from often voluminous amounts of data to better inform investors of the appropriate decisions to take.
Truth be told, however, this practice is still in its infancy. While the exploitation of alternative-data sources—that is, non-financial data that is normally not considered mainstream, such as satellite imagery and geolocation information—has led to something of an arms race among fund managers; extracting meaningful signals from these large, typically unstructured datasets is rarely easy. Indeed, a September 2019 report by the CFA (Chartered Financial Analyst) Institute, which examined the trends and use cases of AI and big-data technologies in investments, included a survey designed to understand the state of adoption of different technologies in the workflows of analysts, portfolio managers and private-wealth managers. The results of the survey showed that “few investment professionals are currently using programs typically utilized in ML techniques, including coding languages such as Python, R, and MATLAB” and that most portfolio managers “continue to rely on Excel (indicated by 95 percent of portfolio manager respondents) and desktop market data tools (three-quarters of portfolio manager respondents) for their investment strategy and processes”. What’s more, only 10 percent of portfolio-manager respondents were recorded to “have used AI/ML techniques in the past 12 months, and the number of respondents using linear regression in investment strategy and process outnumbers those using AI/ML techniques by almost five to one”.
Nonetheless, as algorithms continue to become more sophisticated and as alternative datasets continue to prove their worth generating alpha, data-driven investing using AI/ML is set to become significantly more effective for investors during the coming decade.
But it’s not just through ML that investors are benefiting from AI. A whole range of applications are now being adopted that are helping both retail and institutional investors across various aspects of the investing process. For instance, some are even employing voice-activated trading to allow users to place trades and perform portfolio-related tasks. TD Ameritrade, for example, uses Amazon’s well-known cloud-based voice service, Alexa, to enable customers to monitor their investments, receive virtually instantaneous answers to questions about the market, check account balances and receive customised market updates from the bank.
“Just as smartphones went from novelty to necessity over the last decade, voice interfaces like Amazon Alexa are quickly becoming more pervasive as people grow more and more comfortable using them,” Sunayna Tuteja, head of strategic partnerships and emerging technologies at TD Ameritrade, stated upon Alexa’s introduction in October 2018. “Combining technologies like artificial intelligence (AI), machine learning and voice user interface (VUI), we continue to find new ways to make personal finance and investing simpler for people.” And being equipped with the Alexa-enabled device is even allowing investors to manage their investments remotely from anywhere—surely a major positive in this time of social distancing. This even means that one could check portfolio performance whilst driving, thanks to in-vehicle compatibility with Alexa.
AI is also helping to foster more meaningful relationships between investment managers and advisors and their clients. By automating certain aspects of the relationship—such as the initial communication, documentation handling and risk profiling—clients can spend more time focusing on the more important aspects of their relationships with advisors, namely how to optimise returns. Automating such processes, moreover, helps to lower costs for both the manager and the client. According to Gopal Appuswami, lead of payments, fintech, analytics, products and innovation at LatentView Analytics, using intelligent information-management solutions means that staff have “the means to simplify how they access, secure, process and collaborate on documentation”. The increasing use of AI chatbots also helps to quickly and efficiently address initial queries at the start of the investing relationship. “AI chatbots now serve as the first line of support for retail clients,” acknowledged Phil Andriyevsky, a data and analytics leader at EY (Ernst & Young). And unlike humans, chatbots can be accessed at any time and via a range of digital channels.
By using AI to make important investment decisions, moreover, the errors associated with human decision-making can be drastically minimised, if not completely eliminated, from the investing process. Even the best fund managers will invariably succumb to the emotional and cognitive biases that are inherent in all of us, whether that be confirmation bias, bandwagon effect, loss aversion or numerous other biases that have been formally identified by behavioural psychologists. Failure to acknowledge these problems can and often does lead to sub-optimal asset-allocation decisions. But implementing a system that omits these human errors allows investment strategies to be chosen that are significantly more objective, in both their formulation and execution.
Given these clear advantages, one might wonder whether human investors will be wholly replaced by these rapidly evolving machines. As renowned investor Paul Tudor Jones once remarked, “No human is better than a machine, but no machine is better than a human with a machine.” And this sentiment already seems to be playing out in the robo-advisory industry, in which most investors are expressing a preference for a hybrid model, one that combines the benefits of a digital algorithm with the reassurance and human touch that only a real investment advisor can offer.
With such a range of benefits to be reaped by the investment-management industry, AI can offer an almost entirely new perspective on the investing process. As with many cases related to such innovative technologies, the ones who make the transition towards AI adoption the soonest are likely to be the ones who will benefit most in the long run. And with AI having tremendous potential to bring sophisticated investing to the smartphone, it won’t just be the privileged few who will be able to enjoy those benefits.