How we’re using AI to augment our manager research capabilities
Executive summary:
- Artificial intelligence and machine learning technologies are powering a multitude of enhancements and unlocking new capabilities across Russell Investments.
- Manager research is benefitting from increased quantitative insights and automation efficiencies, enabling both greater breadth and depth of coverage.
- We expect ongoing technological advancements from our in-progress development efforts to continue to position Russell Investments at the forefront of innovation, enabling us to generate unique, actionable insights for the benefit of our clients.
Introduction
We’ve all heard terms like artificial intelligence, machine learning, and process automation bandied about in recent months and years as a way of collectively grouping a diverse range of technologies and capabilities into more easily identified trends. But how do those enabling technologies translate into real world applications? How can process-oriented financial services firms harness these new capabilities for their clients?
At Russell Investments, we’re constantly evolving our technology to enhance our quantitative capabilities for the benefit of our clients. The embedding of artificial intelligence and machine learning into our processes is only the latest step on that continuing journey. We believe these capabilities, when paired with the access-driven qualitative insights gleaned from our investment teams, enable our firm to be able to deliver robust and repeatable beneficial outcomes for our clients. We’re excited to share some of these innovations below.
Background
Before going any further, it might be useful to provide more specific definitions for some of the computational terminology used in this article. While the field is still rapidly evolving (and with the resultant hype there is an increased potential for the misapplication of terms), we will strive for consistency in this article as delineated below.
- Artificial intelligence (AI) is an umbrella term that, in its broadest sense, encompasses an array of underlying technologies, techniques, and processes that allow trained computer systems to autonomously perceive and react to inputs and their operating environment. Example applications of AI include game engines like AlphaGo, autonomous driving systems, and generative tools like ChatGPT.
- Machine learning (ML) is a subset of AI that focuses on the development and deployment of statistical relationships and algorithms to train and ‘teach’ systems using existing data that can then be applied to new datasets to form more generalised and predictive outputs.
- Process automation is a more general term that can incorporate multiple technologies (including artificial intelligence and machine learning) to increase workflow efficiency by drawing on multiple data sources and applications to prepopulate data within reports, suggest differentiated and defining characteristics, and provide ongoing monitoring of risks.
Adding to our capabilities
Russell Investments has a long history of technological innovation, from our RADAR rank and content management system to our Explorer analytical engine to our METRiQ suite of portfolio management tools. The addition of artificial intelligence and machine learning enhancements to these systems is only the latest step in our technology- and data-driven journey to enhance outcomes for our clients.
The world is awash in data and being able to sift through that flood for what is truly important is a key component of what separates successful research and portfolio management teams from also-rans. Based on our extensive combined record of qualitative research, ranks, and quantitative analytics, Russell Investments associates set out to harness the power of AI and machine learning to augment our research processes.
The result was what we call our Manager Research Automation (MRA) initiative, which sets out to innovate and scale our manager research process through automating manual tasks, providing advanced analytics and visualisations, and aggregating disparate data sources (both internal and external) to contextualise and benchmark thousands of managers across consistent evaluation points via an intuitive, user-friendly interface.
Enhancing manager research
While Russell Investments already maintained one of the most extensive repositories of proprietary manager research and analytics data, the MRA effort enables a still-greater breadth and depth of coverage by systematically applying and integrating those capabilities into our Explorer analytics engine. Today, we are highlighting three key facets of that enhancement: discovery, quant, and monitoring.
Within our manager discovery process, our MRA initiative focused on efficiency and automation. Russell Investments analysts are now able to quickly and effectively identify new products within their coverage universes and to create detailed, standardised reports to facilitate informed comparisons with existing ranked strategies.
MRA also augmented our quantitative analysis. While analysts were long able to specify discrete peer universes for comparison, our AI-driven quantitative tools now automatically suggest appropriate peers for managers on both outcome- and fundamentally-based criteria. Our systems then generate quantitative reports, enabling our analysts to conduct detailed evaluations of the relative merits of these close peer strategies.
Technology has also bolstered our monitoring of managers. In assigning ranks and recommending strategies, analysts at Russell Investments are required to identify ‘watchpoints’ (essentially risks to the current rank). Our MRA initiative advances now allow not only for automated monitoring of these analyst-specified watchpoints but will also proactively suggest potential watchpoints based on the observed (and expected) performance pattern of the strategy, enabling analysts to more quickly and readily identify red flags.
In conjunction with the face-to-face meetings and qualitative assessments continually undertaken by Russell Investments analysts, we believe these enhancements enable greater efficiency in our processes, add to our discovery capabilities to help uncover up-and-coming managers beyond naïve filtering, and increase our ability to identify early warning signs of potential deterioration in efficacy among ranked managers.
What’s next?
Building on our heritage of innovation, Russell Investments is continuing to develop proprietary AI and ML capabilities to further leverage our technology platform for the benefit of our internal teams and end clients. Though this brief update will (hopefully understandably) be somewhat light on detail, we can reveal that our research teams have developed a proprietary tool (QMS) that uses machine learning to predict manager performance and risk profile relative to peers on both a short-term and long-term basis.
QMS aims to capture complex relationships and dynamics within the manger universe, utilising a diverse set of features that cover product and firm characteristics, holding fundamentals, and financial metrics to provide a unique, quantitative view of the manager. Over time, we expect to further integrate the QMS tool into our manager research process to enhance both the manager discovery process and risk detection among existing funded managers. Stay tuned.