With some of the initial AI hype fading, experts say firms are finding more practical uses for the technology.PhonlamaiPhoto/iStockPhoto / Getty Images
Artificial intelligence has captured the public’s imagination and led to huge stock market gains for the big technology companies set to benefit from its growth, but there’s plenty of room for adoption in the investment industry itself.
A January report from the CFA Institute revealed that investment firms still favour legacy technologies compared with AI and big data. However, firms’ appetite for AI is growing.
At the same time, some of the hype around AI is softening. Jason Pereira, senior partner at Woodgate Financial Inc. in Toronto, says perceptions of AI have passed the early stage at which expectations are often unrealistically high.
“Then, we discover where the limits of a technology are,” he says. “Anyone who thought they’re going to replace advisors with this is being proven dead wrong.”
In a 2024 report on emerging technologies, research company Gartner Inc. acknowledged this shift in perception, arguing that generative AI has already begun slipping into the “trough of disillusionment” in which people abandon their initial high expectations. After this point, people find more realistic uses for the technology.
Here are some of those uses advisors could see this year:
Beyond the front office
Investment firms will be looking to move beyond useful but limited front-office productivity applications – such as transcribing meetings, summarizing documents and drafting e-mails – and begin integrating AI more deeply into back-office workflows.
The CFA report found much work still to be done here. Just 9 per cent of survey participants from this industry had integrated AI and big data technology extensively across core business and mid- and back-office functions.
“AI technologies are still evolving in the industry, and we have seen a fragmented approach to implementing AI at firms,” says Cheryll-Ann Wilson, senior affiliate to the CFA’s research and policy center and the report’s author.
Deeper integration could span multiple uses, including generative AI-powered enterprise search. This uses deep learning to probe huge mountains of unstructured information in ways traditional databases can’t. These data are buried in documents such as reports and news articles from many different sources, including third-party services and a company’s own data.
“This is about bringing data closer to the advisor or branch manager,” says Joseph Lo, head of enterprise platforms at Broadridge Financial Solutions Inc. in Vancouver.
AI as an investment tool
Broadridge Financial Solutions Inc.’s BondGPT feature already uses OpenAI’s ChatGPT 4 generative AI to search many data sets for bond information. Traders can query those data conversationally to get new insights, opening up more possibilities for non-technical staff in investment companies.
Others are exploring AI for portfolio construction. Last year, Evolve Funds Group Inc. launched an exchange-traded fund that used generative AI from Gradient Boosted Investments Inc. (also known as Boosted.ai) to help curate a managed portfolio of AI companies.
Raj Lala, Evolve’s chief executive officer, says this trend will grow, with AI-driven models incorporating macroeconomic indicators and investor risk tolerance to rebalance portfolios automatically.
He also expects more AI-assisted automated trading algorithms. “AI can move beyond predefined algorithms and create trading strategies dynamically in real-time, adjusting based on changing market conditions.”
An army of AI agents
AI systems will become more autonomous, experts say, as companies such as OpenAI and Anthropic PBC expand into “reasoning” models. These can break down complex requests into multiple tasks for applications such as complex calculations and deeper research.
According to Zac Maufe, global head of regulated industries at Google’s cloud business, the next step is to allow AI to execute these tasks itself by talking directly to other software applications.
For consumer users, that means AI agents might move beyond suggesting holiday trip activities to actually finding flights, hotels and restaurants, and booking them on your behalf.
Investment industry applications will focus on internal tasks, Mr. Maufe says. Initial uses could be simple, such as collating, categorizing and forwarding expense receipts for an advisor.
However, applications for AI agents could evolve quickly. After a reasoning AI model researches a market and builds an investment thesis, Mr. Maufe believes an AI agent could take action on it, perhaps changing asset allocation with a human’s guidance and approval.
“It’s not happening yet, but that’s where we’re headed,” he says.
Multimodal AI
The way people interact with AI will also evolve, Mr. Maufe says, pointing to multimodal systems as a key enterprise feature for 2025.
Multimodal AI understands multiple types of data simultaneously. Its ability to digest and distill multiple document types –including audio transcripts, videos, slide images and text-based reports – will carry big benefits for investment companies, Mr. Maufe says.
“Imagine the ability to create a podcast from all that dense information that someone can listen to on their commute,” he says. Google’s NotebookLM AI product is designed to do that.
Not all at once
One challenge for investment companies, Mr. Maufe says, is refining their data governance practices and data quality to ensure appropriate AI outcomes.
Evolve’s Mr. Lala also expects increased regulatory scrutiny.
“Regulators are still figuring out how to govern AI-based decision-making in finance, particularly around transparency, explainability and bias in AI models,” he says. “The regulators may demand greater auditability of AI-driven strategies, slowing adoption.”
Many AI trends will take time to catch on in the traditionally conservative investment sector.
“It will become more integrated into the back-office process, but this will be gradual,” Mr. Lala says.