Rani Piputri, Head of Automated Intelligence Investing, NN Investment Partners
Where are we with AI?
Artificial intelligence (AI) has become an ubiquitous term in this day and age. Advancements of AI in industries such as media, medicine and engineering will not have escaped anyone. In finance, however, the picture is fragmented. We hear about fintech companies offering electronic payments in more efficient ways, and robo-advisory firms offering automatic investment schemes. However, inasset management, AI manifests itself insubtler ways. Large and complex asset management companies typically deploy AI in automation of reporting and digitization of its client services. The smaller, more niche counterparts in the industry, such as hedge funds, tend to apply AI in their pursuit of alpha – the value added of investment beyond simply market exposure.
In this fragmentation, the perception of hype versus reality cannot be ignored. When we talk about AI in making forecasts, systematic investors have already been using predictive models for decades. These models contain intelligence that adjusts to existing and anticipated market environments. For others outside of this narrow group of investors, this may be experienced as a new innovative force. This is simply because the underlying technology is now easier and faster to access, the data has become more abundant, while at the same time, the interface has become more intuitive and user friendly.
Clearly, AI is not the silver bullet that will guarantee an investment return for an asset manager. It is merely a tool that could give us more accurate and timely information to answer the same old questions: what, where, when and by how much should I invest, so that I benefit my clients?
In answering such questions, the potential of AI stretches beyond process automation and digitization of client services. AI offers a large pool of possibilities and enables faster lifecycle of research. Specific methodologies within AI allow investors to find value in novel sources such as textual data. Not only can it extract sentiment and other information from regulatory filings, articles and social media outlets, the mere electronification of contractual documents massively enhances the speed of investment decision, especially in illiquid and less transparent asset classes.
"Just as insurance and credit card businesses deploy AI to detect fraud, asset managers utilize AI to uncover unusual behaviours, especially in transactions of financial assets, aggregate investor behaviours, and ultimately macro and micro market behaviours"
In addition to its potential to work out complex and interconnected financial markets, AI is used by asset managers, in combination with human input, as an enabler to help compliance departments and risk officers. Just as insurance and credit card businesses deploy AI to detect fraud, asset managers utilize AI to uncover unusual behaviours, especially in transactions of financial assets, aggregate investor behaviours, and ultimately macro and micro market behaviours. Outlier detection has broad applications in developing investment strategies as well as risk management in the industry.
The potentials in AI comes with its own particular challenges for the asset management industry. AI techniques are often associated with black box operations. For an industry that is highly regulated and relies heavily on the trust that society puts on it, limited or even non-conventional narratives pose a significant hurdle for acceptance. When an investment decision is based on models that self-learn, either supervised or unsupervised, the sense of control often needed by clients will be fundamentally challenged. In the event of an investment decision not panning out, who is accountable?
Given its tremendous potential and equally substantial challenges, adoption of AI in asset management will most likely be a bumpy ride. The hype around the theme will disappoint those who are looking for a major disruptive force in the industry. Regulatory environments will determine much of what can or cannot be done, as the long term impact of AIis yet to be seen. Structural impediments around transformation of skills and organizational culture will also take time to resolve in this industry that is often considered slow moving. In many aspects of life, we are embracing the progression of AI. It is a real prospective that in asset management too, AI will immerse and bring long term benefits to the society at large.