How money managers are transforming AI solutions from experimental to practical

“It’s quite amazing how far the industry has come in 17 years.”

Read: 59% of global institutional investors haven’t discussed AI with investment managers: survey

Artificial intelligence solutions in money management are transforming from experimental to practical, with nine out of 10 managers confirming current or planned AI use, according to Mercer’s 2024 survey of its global investment manager database. Indeed, the survey found almost half of respondents were using large language models (44 per cent) or machine learning models (48 per cent).

The pursuit of investment strategy edge is one of the factors driving the adoption of AI in money management, says Alpesh Sethia, chief technology officer at the Healthcare of Ontario Pension Plan.

Money managers are at the forefront of the adoption of novel artificial intelligence solutions as they transform from experimental to practical

“I’ve talked to many asset managers and, often, that edge is something where they invest heavily in data science within their organization to make sure they’re looking at data the right way, processing it the right way and putting it into the lens of what their portfolio looks like to manage it.”

AI in practice

The novelty of AI may still surprise some people, but to O’Sullivan, the technology is an off-shoot of the ongoing work of active managers.

“In the fixed income world, [it] came as a logical progression from what we and some others [were] doing before, which is using computing power to sift through data to try to find signals before other people could find them.”

Read: Expert panel: Investment strategy, AI among key issues for DB pension plan sponsors in 2026

RPIA’s system makes suggestions to portfolio managers, highlights anomalies and shows new issues that could be potentially priced in for an asset.

At Mackenzie Investments, AI tools are most predominantly used by the quantitative equity team and the systematic team in its multi-strategy group, according to Christopher Boyle, senior vice-president and head of global institutional and partnerships.

“There’s a lot of programming that happens within that space and our teams are using AI as a programming tool. That means we’re able to move faster to implement new signals we’re identifying.”

The organization relies on a rigorous governance framework that examines its AI tools and works on model validation, he says, noting this is still an innovative tool.

Pictet Asset Management has been using AI capabilities for more than seven years, says François Forget, head of distribution in Canada. “We materialized it three years ago by launching strategies on it.”

When the firm hired a new head of quantitative, he adds, it decided to lean on the wealth of information from an internal research group and launch strategies dependant on AI. Its long-short strategy using AI is about $1.2 billion and its long-only is $2.5 billion, says Forget.

Demand fueled by mainstream effect

Michal Prywata, co-founder of the AI-powered institutional finance platform Vertus, has been busy with calls from money managers interested in how to use AI.

Read: 2025 Investment Innovation Conference: How will AI impact the future of investment leadership?

He’s on a campaign to raise awareness on what the technology can and can’t do. “The conversations we’re having are highly educational. . . . People are just very excited and want to be at the leading edge of these technologies.”

While he recognizes a natural interest from the money managers combined with a degree of demand from institutional asset owners, he also notes there’s a big lack of understanding of how it all works. “People think you’ll tell [OpenAI’s] ChatGPT to invest for you and you’ll do fine. You might get lucky, but it’s not purpose-built for that.”

For all AI’s bells and whistles, it may still be too early for the tools to sway pension plan sponsors when evaluating investing partners, says Genevieve Hayman, senior research affiliate in research, advocacy and standards at CFA Institute. “When [plan sponsors] look at external service providers, the question isn’t, ‘Do they use AI or not?’ The question is, ‘Do they have a capability to contribute something, to give that value-add and whether or not it uses AI is a secondary concern.’”

When Forget meets with investors, he’s asked to explain the performance of Pictet’s strategies and the thinking that goes into them. “We can explain what’s considered, what was built in, how it has evolved over time. We can provide a deepdown analysis of why those outputs were generated and we can provide . . . performance attribution.”

The interest from institutional asset owners for AI-driven solutions is thoughtful and measured, according to Jeff Shen, co-chief investment officer and co-head of systematic active equities in the systematic investment team at BlackRock Inc.

Read: Institutional investors urged to examine AI integration beyond productivity: CPPIB

Determining factors for investors, he adds, include consistent performance, robustness across market environments and how the strategy itself fits within broader portfolio objectives. “They ask how AI is used in the investment process, how models are governed, how risks are controlled and how outcomes are monitored,” he wrote in an email to Benefits Canada.

Colin Ripsman, president at Elegant Investment Consulting, is receiving questions from plan sponsors about how AI is being used to inform third-party investment strategies. Typically, he finds smaller firms aren’t very forthcoming with details while the larger managers are open to share more information. “I think that’s going to become more of an issue over time and it’s something we do evaluate.”

The HOOPP’s Sethia admits the use of AI tools requires more in-depth reviews by organizations and should fall within an acceptable use policy that’s created in-house. Its guideline includes regulations around how employees use consumer tools like ChatGPT, Anthropic’s Claude or Google’s Gemini.

“We have an AI acceptable use policy. . . . It’s quite easy to go to ChatGPT, without even logging into it and trying to put stuff in and get an answer back. But we put a policy in that you need to protect HOOPP’s data.”

Despite the untapped possibilities of AI in investment management, Boyle still expects returns to be a dominant factor in how asset owners pick managers. “The primary distinguishing characteristic — or the one that’s going to lead people to entrust money to us — is going to be our ability to generate an attractive, predictable return pattern.”

Evening the playing field

It’s likely that the adoption of AI as an investment tool will lead to a decline in hiring analysts because many tasks can be delegated to AI tools while retaining a human decision-making process, says Russ Goyenko, an associate professor of finance at McGill University’s Desautels Faculty of Management. “AI can summarize a lot of reports and summarize a lot of earnings calls. It can save a lot of time.”

Read: U.S. tariffs, AI impacting institutional investors in 2026: report

Alex Dameski, country head for Canada at Apex Group Ltd., isn’t convinced by the replacement theory. He sees AI tools as part of the human process instead of entirely taking over. “In a lot of cases, when [revolutionary technology] is deployed, certain jobs go away. There’s potential to retrain people, but deployment [should] theoretically be slow enough that we could adapt pretty effectively.”

Ripsman says larger, global investment management firms have a leg up in the AI race. “It’s not just a question of the reach; it’s also access to data. For these models to work, you [need] massive amounts of data and maybe that’s beyond the capabilities and the reach of small, local firms.”

Despite these advantages, he’s curious to see how the larger managers will evaluate AI tools because, in the past, quantitative models worked well in certain controlled environments but later struggled with inflecting markets. Using AI tools in an investment environment isn’t as simple as blindly using the data produced by the model, he adds.

On the other hand, Boyle considers AI to be an equalizer between the largest asset management firms and their smaller peers because of the efficiencies introduced by the technology. “I think it’s going to allow firms that don’t have as large teams to compete because the application of AI is going to be such a powerful driver of efficiency and innovation for firms that have had to rely on having large teams in order to do that.

“You can have a core team — a strong, nimble team — and then complement it with this kind of capability.”

Bryan McGovern is an associate editor at Benefits Canada and the Canadian Investment Review.