The value of AI in asset management

Executive summary:

  • AI can be used by investment managers to refine the analysis of buy or sell signals and to extract signals from large quantities of data via pattern analysis. It can also be used to improve trading efficiencies.
  • AI's ability to connect different types of data can help uncover hidden insights and unlock investor sentiment.
  • At Russell Investments, we're leveraging the technology to streamline the quantitative portion of our equity manager research process. This is allowing us to devote more time to our qualitative research efforts.
  • AI is a valuable tool for the investment community, but it’s by no means a substitute for human thinking.

It’s been nearly two years since generative artificial intelligence (AI) took the world by storm, with the release of large language models like ChatGPT, Copilot, and Gemini dazzling humankind with their ability to interpret human requests and respond with the desired output—be it a summarized article, an in-depth data analysis, coding for an app, or the creation of images or videos. The possibilities of what AI can do today appear just as vast and endless as they did in late 2022—especially considering that the technology is still in its early days.

This is why, at Russell Investments, we believe AI has the potential to be economically transformative by augmenting human productivity through efficiency, personalization, knowledge sharing and information access, and advanced data and analysis. Two years on, the technology has largely met our operational expectations, boosting efficiencies in processes like coding and the extraction of insights from data. However, we are cautious about the hype, as AI and the regulations around it are still evolving. Ultimately, we believe that human oversight remains critical for truly effective implementation—especially in the financial services sector, where the context and relationships are deeply important, and where we see the technology as a way of boosting human productivity rather than replacing it.

So, how might investment managers use AI in their strategies today? What are some specific ways it can be used to help streamline operational efficiencies and broaden access to investment insights?

Let’s dig in and take a look.

What tasks can AI automate for investment managers?

Fundamentally, AI’s ability to analyze and draw inferences from patterns in data without following explicit instructions offers a multitude of potential investment applications. These include potentially using the technology to refine the analysis of buy or sell signals, as well as using pattern analysis to extract signals from large quantities of data. AI can also be used to improve trading efficiencies.

Generative AI in particular can be extremely valuable in helping identify themes and emerging trends in investing by scraping information from earnings reports, calls, and news articles. In the past, doing so was a tedious, laborious task carried out by investment analysts who spent hours each day pulling and manually compiling data from various sources. Generative AI eliminates this process entirely by gathering and synthesizing information instantly.

The list of repetitive tasks in the investing world like this that AI can help with—which are still mostly being done by humans—is expansive. It includes, among other things, the generation of investment reports, the automation of compliance and control-related tasks, and assistance with email and communication management. Simply put, AI’s ability to process human language enables the transformation of textual data into actionable insights, offering rapid analysis that can inform investment decision-making —in barely more than the blink of an eye.

Perhaps one of the most exciting and innovative uses of AI in research is how it can connect different types of data. It can combine structured data, like financial fundamentals, with unstructured data, like news articles or social media, to give a much fuller picture. This ability to pull together various data streams helps uncover hidden insights and unlock investor sentiment. Crucially, it also helps spot patterns or connections that, in the past, could have only been found by analysts with deep expertise and plenty of time on their hands. With AI, this kind of advanced analysis is rapidly becoming democratized, allowing more individuals access to key investment insights in a quick and easy fashion.

How we’re using AI to improve our operational efficiencies

At Russell Investments, we’ve leaning into AI to help streamline the quantitative part of our equity manager research process. Specifically, we’ve leveraged ensemble machine learning models in order to analyze large datasets and improve our investment product discovery process. AI has also helped us streamline routine tasks, like report summarization and code writing, allowing our teams to spend more time on other value-adding tasks.

Chart: How we’re using AI at Russell Investments
AI manager researchIn addition, we’re continuing to experiment with additional ways to use AI to increase efficiencies. As part of this effort, we recently concluded a project to build an industry news bot that helps filter out the most relevant insights on what is happening in our industry. Additionally, we have built a tool that creates presentation slides directly from investment memos, streamlining the process and saving valuable time.

It’s important to stress that while we see AI as a valuable tool to improve our operational efficiency and services, we believe it only works if it’s used in a responsible and ethical manner. As such, we’ve established an AI ethics and governance board to create guidelines, policies, training, and oversight for the responsible implementation and use of artificial intelligence within Russell Investments.

How AI enhances our manager research process

The function of manager research is a key component of the investment process for many asset managers and asset owners, with in-depth quantitative and qualitative analysis essential to understanding a money manager’s ability to outperform. Quantitative research, because its very nature centers around gathering and analyzing large amounts of data related to the performance and holdings of a manager’s portfolio, has traditionally been a highly labor-intensive process.

Enter AI. Thanks to the many different flavors of the technology, at Russell Investments, we’ve been able to automate much of the quantitative part of the manager research process, including generating analysis and reports. One of the most innovative components was the development of a quantitative manager screen that uses machine learning to help us discover new manager products and analyze over 10,000 equity investment products—something we could never have done manually. This capability significantly enhances the breadth of our analysis, while the rest of our automation efforts drive greater speed and efficiency in our processes.

Given the vast amount of data on manager products that we possess at Russell Investments—such as portfolio characteristics, firm and team characteristics, exposures, risks, and past performance—machine learning can enhance the manager research process at both the discovery stage and during ongoing maintenance. However, we can’t stress enough the importance of balancing quantitative/AI methods like these with qualitative approaches.

This is because while quantitative methods—including those powered by AI—excel at quickly analyzing large datasets, identifying patterns and similarities, and making predictions based on historical data, they lack the context, intuition, and nuanced understanding that human judgment provides. Simply put, an access-based qualitative approach offers in-depth insights, contextual understanding, and the ability to interpret complex situations and nuances in investment processes or team dynamics that cannot be captured by data. Can AI enhance human judgment? Certainly. But can it replace it? Absolutely not.

That’s why, at Russell Investments, our research analysts are continuing to conduct in-depth qualitative assessments to ensure that the summarizations and insights provided by AI are contextually relevant and accurate. What’s more, allowing AI to sift through all the manager research we’ve compiled over the decades actually allows us to devote more time to our qualitative research efforts. For instance, our analysts can now spend more of their time in face-to-face meetings with managers, exploring the differentiating aspects of their investment philosophies, rather than manually building charts and tables or poring over mounds of data to extract investment themes. Ultimately, we believe that combining the strengths of both our qualitative and quantitative research approaches will lead to a well-rounded evaluation of investment products.

The bottom line

AI’s ability to efficiently extract from and analyze both structured and unstructured datasets has the potential to vastly enhance the qualitative portion of the manager research process. In a nutshell, it provides more information more quickly—allowing today’s research analysts to have more investment insights at their fingertips than ever before.

Critically, though, all of this information must still be digested, processed, and weighed against a host of other factors on the qualitative side before a decision is made. This is where the value of human oversight comes into play. While AI is unquestionably a valuable tool for the investment community, it’s by no means a substitute for human thinking. Use it wisely.