The use of artificial intelligence (AI) by professional investors has come a long way since its early applications in algorithmic trading in the 1990s. To date, its biggest impact has been in operations – specifically enabling slicker, sharper processes across the back and middle offices.
And yet, as AI’s strength lies in collecting, classifying, simplifying and rationalising information that is growing exponentially in size, diversity and complexity, a natural evolution has underpinned its rise across the front-office – or client-facing – functions.
While investment banks, hedge funds and quantitative strategies have been harnessing the power of AI for years – through the use of machine-learning, natural-language processing and trading-pattern recognition, its use today reaches far beyond these traditional domains. According to a Mercer survey, nine out of 10 investment managers (91%) are currently using or planning to use AI in the next two years as part of their investment strategy or asset class research.
In today’s volatile, cost-conscious environment, professional investors are challenged more than ever to generate alpha and mitigate risk using creative, expansive and effective methods.
Given the inherent uncertainty in the market and the speed with which new information is generated and then incorporated in trading decisions, the ability to accurately capture and model a very large number of scenarios, before extracting the relevant information upon which they are based, is critical to the survival first, and then the success, of a multitude of investment strategies and portfolios.
Increasingly, AI is becoming essential to investment management to process vast and complex data sets with lightning speed and precision. Unlike traditional modelling tools, AI can analyse not just historical financial data – the usual yet still very relevant fundamental and technical aspects of investment research – but also more unconventional sources such as satellite imagery, social media sentiment and video content which, with the right technology, can often turn into ‘data treasure troves’ to generate comprehensive investment insights and forecasts.
By continuously and accurately monitoring and interpreting this array of information, AI-driven systems respond to changing market conditions, spotting subtle patterns and identifying latent risks that might elude even experienced human analysts.
“AI can analyse not just historical financial data but also more unconventional sources such as satellite imagery, social media sentiment and video content to generate comprehensive investment insights.
It is not just about stock-selection – AI could feature heavily in bespoke or managed portfolio construction, which is potentially a next wave of adoption.”
Will AI then replace the human touch? Not necessarily, as we do not think that replacing humans is its best use – indeed, far from it. AI is a very powerful enabler of more, sharper and timelier investment decisions.
AI’s role can be compared to that of an extremely bright, technically capable graduate – they have all the latest knowledge of technique and analysis but lack experience of contextual application. AI will not make decisions but it will alert the team to anomalies, ideas or opportunities that warrant further investigation – supported by a robust rationale.
Where AI really comes into its own is modelling with a huge number of moving parts. For a human, analysis can get out of hand very quickly – even with the cleverest of formulae.
AI, on the other hand, can process billions of alternatives with precision – at a latency that matches and is often lower than requested by the user – and then make sense out of them to offer detailed forecasts and risk assessments to inform smarter investment decisions … all on continually shifting sands.
Advances in computational power and the capacity to integrate both structured and unstructured data bring the edge, enabling investment managers to build and rebalance portfolios, identify emerging risks and improve forecasting accuracy over and above traditional techniques.
It is not just about stock-selection either – AI could feature heavily in bespoke or managed portfolio construction, which is potentially a next wave of adoption. It can examine and sort thousands of funds – both active and passive – and apply classifications that are more objective and more nuanced than a simple look at their KIID or prospectus might allow.
Want to assess the underlying risk/reward trade-offs and rank investments accordingly? Are the active funds genuinely delivering alpha? AI can incorporate assessments around fees and liquidity and suggest overlays based on consistency of process and performance, service levels or existing relationships.
Trying to fully understand all AI’s variants and potential uses at once could be quite overwhelming - instead, begin with small steps and build on those over time.”
AI can also be used to harness information to compare and construct bespoke portfolios that suit a client’s age, risk appetite, time horizon, investment objectives and other relevant criteria. The future then starts to look genuinely ‘digital’ and personalised, rather than relying on pre-set buckets that are, by necessity and by construction, restricted and move quickly out of date, given levels of correlation of certain asset classes and the increasing use of alternatives.
As trailed above, then, while some use cases have been around for decades – albeit in more primitive forms – many are only just discovering AI’s potential.
We see its application in asset and wealth management as sitting along a spectrum. At one end this might be simply monitoring performance or setting rule-based alerts; further along, it might be assisting portfolio construction with optimal weightings against a rapidly moving backdrop; or at the more advanced end, it could mean running deep analysis on and simulating or stress-testing various scenarios and predicting outcomes based on dynamic current market data.
As such, while trying to fully understand all AI’s variants and potential uses at once from a standing start would be quite overwhelming, beginning with small steps and building on those over time, will still lead to meaningful gains.
We see the next surge of activity as moving along investment categories – moving out of the institutional domain and more rapidly towards wholesale, wealth, private client and retail.
What will drive that trend? For one, the amount of wealth changing hands into those of younger, more tech-native clients and an accelerating level of familiarity – largely thanks to the rise of large language models such as ChatGPT that have introduced AI into our everyday lives.
Also though – and very importantly – there is the prospect of wider accessibility, acceptance and superior results. One would be tempted to warn that professional investors ignore AI at their peril.
Giovanni Beliossi is head of investment strategies at Axyon AI
