In March, two ETFs launched in Canada that use AI to analyze and select stocks.ookawa/iStockPhoto / Getty Images
Sign up for the Globe Advisor weekly newsletter for professional financial advisors on our sign-up page. Get exclusive investment industry news and insights, the week’s top headlines, and what you and your clients need to know. For more from Globe Advisor, visit our homepage.
Artificial intelligence (AI) is changing every industry, and investing is no exception.
Already, AI is making significant inroads in the asset management space. Recent examples include the launches in late March of two exchange-traded funds (ETFs) in Canada powered by the technology.
“It’s almost like you can download your brain into the machine so you can scale up, as a portfolio manager, your thinking,” says Josh Pantony, chief executive officer (CEO) and co-founder of Boosted.ai, based in Toronto.
Boosted.ai provides the AI technology used in Evolve Funds Group Inc.’s new Evolve Artificial Intelligence Fund ARTI-T. The ETF leverages generative AI, which can pore over vast amounts of data, including text, using large language models to synthesize insights.
Equally important, generative AI allows portfolio managers and research analysts to pose questions as they would to another person and receive potentially extensive, complex answers in text or voice format. Think ChatGPT but more focused on solving investment problems.
“For example, if you want to ask the machine what stocks will do well as global warming gets worse, it will provide meaningful answers backed with the sources it used to arrive at that conclusion,” Mr. Pantony says.
An Ernst & Young Global Ltd. report published on Oct. 31, 2023, that surveyed asset managers found the top use case for the industry was data analysis to drive alpha-generating strategies.
“If you have a system that’s able to read all the articles, companies’ filings, earnings reports and other related information – specifically through the lens of improved product sales and efficiencies – that’s pretty powerful,” says Raj Lala, CEO and president of Toronto-based Evolve.
Evolve Artificial Intelligence Fund uses AI to find companies that are potentially benefiting the most from AI itself. Meanwhile, another ETF that leverages AI and launched in March is focused on global equities.
Evovest Global Equity ETF EVO-T uses “machine learning to model excess return versus the benchmark using financial information at our disposal,” says Carl Dussault, CEO of Evovest Inc., a Montreal-based firm founded in 2017.
Evovest’s AI analyzes large sets of financial data – only numerical data – to determine which global stocks are likely to outperform on a weekly basis.
Evovest and its partner, National Bank Investments Inc., which supports the ETF’s back-end operations, see the new ETF as just the start for AI-driven, actively managed funds aimed at generating alpha.
Central to its potential is that AI can provide active management at a lower cost because it saves portfolio managers and analysts time researching securities to build unique portfolios.
“The whole goal is to obtain more information than a human can and interpret it in a better way,” Mr. Dussault says.
At the end of every week, Evovest’s AI analyzes and rates equities, which are then assessed by portfolio managers. On the following Monday, the ETF portfolio is reconstituted based on the AI insights.
“Conceptually, what we’re trying to do is put price targets on securities that broker research teams otherwise do with far more people,” Mr. Dussault says.
Rather than remove the human touch, analysts and portfolio managers still have the final say. Most importantly, they can see how the AI arrived at its conclusions.
Of course, it’s not just these two Canadian fund companies leveraging AI in this way. The technology is viewed as a game-changing tool that will “supercharge” research and decision-making, says Kristina Karnovsky, executive vice-president and chief product officer at market insights company FactSet Research Systems in New York.
“There is a lot of competition in the industry, and portfolio managers and research analysts are stretched to find timely intelligence and understand key drivers and risks – all while doing more with less,” she says.
FactSet’s new AI tool, known as Mercury, launched in December and uses generative AI to address all those needs.
Central to Mercury is its search tool allowing users to ask questions to the AI about securities and other market information FactSet has collected during its more than 40-year history and receive insightful answers.
“They can just pose their question without having to know where datasets can be found or what codes might be needed to request them,” Ms. Karnovsky says.
Previously, many investment research tools had steep learning curves before they could be used effectively.
“The whole idea is to give users the right information at the right time in the right format, whether they’re asking a simple question about a stock, wanting a chart built for a report, or a summary of a filing transcript,” she says.
FactSet is certainly not the only market research provider offering an AI-fuelled product. Bloomberg LP – considered the leading provider of market research via its widely used terminals – launched a generative AI tool, BloombergGPT, last year.
Still, the technology is not perfect. One particular concern about generative AI is “hallucinations,” whereby the software finds connections in data that are incorrect.
“Hallucinations often result from searching a vast universe of content, like the internet, for an answer,” Ms. Karnovsky says, and FactSet limits this risk by only allowing its generative AI to use FactSet’s database.
Hallucinations are also a concern for Evovest and Boosted.ai, both of which state they have methodologies to mitigate the risk.
“Once the AI suggests a thesis, a separate layer of the technology double-checks it,” Mr. Pantony of Boosted.ai says.
“The hallucination risk is also why it’s still important to have a human involved.”
That’s central to Evovest and FactSet’s tools too, where humans are the backstops.
Mr. Pantony can’t foresee asset management companies fielding AI-driven products without real professionals vetting the strategies first. Yet, it’s almost certain most asset management firms will soon rely on AI to build portfolios, he says.
“There’s going to be a great democratization of this technology over the next few years,” he says, noting it won’t be just asset managers using it but investment advisors and retail investors.
“What we’re seeing for capabilities today is just the start,” he says.
For more from Globe Advisor, visit our homepage.