The adoption of artificial intelligence (AI) and data science applications in the investment management industry globally has risen rapidly over the past three years. The democratisation of data, enabled by technology, has given rise to this proliferation, strengthened by the growing sophistication of behavioural analytics.

Applications in asset management rely on several major techniques. In the case of natural language processing (NLP), computer vision and voice recognition programmes, AI is used to capture text, audio and imagery from a variety of public sources and internal/vendor databases. This includes transcribing analyst conference calls and extracting data from issuer filings for valuation models. In recent years, the popularity of AI in general – and that of machine learning (ML) specifically – has surged in both practice and academia. Machine learning is concerned with using data progressively to adapt the parameters of statistical, probabilistic and other computing methods and algorithms that allow machines to uncover patterns without explicit programming instructions. Deep learning (DL) is a type of ML that is based on artificial neural networks (a type of learning modelled on the human brain). DL algorithms are often applied to improve the results of NLP, computer vision and voice recognition programmes. They can also help extract useful information from large piles of data for analysts and portfolio managers to process. For example, these algorithms can infer certain key words from conference call transcripts or identify sentiment from unstructured data, such as social media. AI techniques can be used in fundamental bottom-up research to identify key themes driving narratives and potentially the current market valuation or identifying changes in the tone and strategy of company management, for example. However, ML and DL applications tend to be popular with quantitative (systematic) managers who often find it helpful to apply these techniques to improve the effectiveness of their quantitative processes.

Using big data for industry and company analysis is more popular than using AI or ML techniques among investment professionals. Asset managers are looking to big data to help them drive high-quality outcomes so that they can invest more effectively in their core activities. Data scientists define big data with four Vs: volume, variety, veracity and velocity. The terms often used in the investment circles are alternative data or unstructured data. Alternative data refers to data from sources that are not currently used or not yet mainstream. In comparison with structured data, which is data that is digitised and stored in relational databases, unstructured data refers to data that is often in text, image or voice formats and are not readily processable. Examples of alternative data and unstructured data often used in the investment world today, include satellite images, earnings conference call recordings and transcripts, social media postings, consumer credit and debit card data, and eCommerce transactions. In a way, extracting signals from big data is simply an extension of what a retail analyst would do – visiting stores to analyse customer traffic, merchandise and so forth. Now asset managers use satellite images or sensor information of the parking lot to infer the same information. The new techniques offer efficiency gains; an analyst can cover a lot more stores in much less time using satellite imagery or sensor data. Alternative data tends to be niche and is more popular with fundamental managers running discretionary portfolios who use these data as one input in the investment decision-making process.

According to research from the CFA Institute, investment professionals globally are using AI and data science techniques for industry and company analyses in the following ways:

Artificial intelligence

Source: CFA Institute, AI Pioneers in Investment Management Survey, 2020
Note: Survey participation (N = 159)

Data science

Source: CFA Institute, AI Pioneers in Investment Management Survey, 2020
Note: Survey participation (N = 159)

One caveat is that these results do not allow us to infer how often or how extensively these data sources are being used in industry and company analysis. However, these results do suggest that the investment industry is generally in the early stages of adopting AI techniques and related technologies, and a few investment professionals currently use big data applications in their daily investment processes. Overall, given the low usage of AI and big data techniques, coupled with the large number of data scientists entering the investment industry, it is reasonable for asset allocators to expect to embrace a growth in modern approaches to investment decision-making.

The effects of the COVID-19 pandemic ushered in a new era in March 2020, as traditional models failed in the systemic economic shifts caused by the pandemic, and alternative data came into play as a provider of real-time data to derive insights for immediate action. This change triggered an interest in using AI and alternative data to both increase signal accuracy and generate a competitive edge as asset managers reset their strategies. Asset managers realise that inputs and data should not include only structured data like the clean time series such as prices, PE’s or economic indicators. To gain an edge, the current development or implementation in the investment industry has increasingly been predictive models using unsupervised learning to understand data structure, supervised learning algorithms for prediction and reinforcement learning to learn actions. Another popular technique is the use of neural network architecture taking in price data such as FX, equity indices, yields, commodity prices and so forth. Using the most efficient search technologies such as cloud and/or distributed scaling, global traffic congestion monitoring, real-time inflation trackers and using deep learning to gauge sentiment on social media and the news are also some techniques that are gaining popularity in increasing the predictive power of models.

As Absa Multi Management, we are always looking for innovative ways to combine the collective power of both artificial and human intelligence. The path of adoption begins with routine, rudimentary tasks that analysts perform, that AI and big data have some advantage over human beings in the breadth of information that they can process at high speeds. Incorporating AI and data science techniques can augment human intelligence to enable investment professionals to reach a higher level of efficiency, minimising behavioural bias and making investment decisions that leverages the collective intelligence of machines and humans. The differentiating factor among investment teams of the future will excel in collective intelligence through cognitive diversity (artificial and human) to achieve a competitive edge.

Disclaimer: This article has been produced by Absa Multi Management (“AMM”) a division of Absa Investment Management services (Pty) Ltd (“AIMS”), Reg. No. 1980/002425/07, an Authorised Financial Services Provider, License No. 524. The information contained in this article does not constitute an offer or solicitation to enter into any transaction, nor does it constitute any recommendation, guidance or proposal to enter into any transaction. While every effort has been made to ensure the accuracy of the above information, AMM does not accept any liability or responsibility for any loss, damage or expense incurred neither in relying on the above information nor in the use thereof, nor makes any representation as to the accuracy or completeness of the above information. Anyone making use of the information contained in this article does so entirely at their own risk.