A friend once posited the idea of a time traveler taking his computer back to the 1940s to provide Sir Winston Churchill with the power of artificial intelligence so the Allies can wrap up the Second World War years ahead of schedule. Immediately recognizing the potential of AI, Sir Winston sets the time traveler to the enormous task of compiling his wartime memoirs.
The real-world use of artificial intelligence in investing circles today is similar. While all eyes are on the potential for AI to begin offering infallible investment advice, family offices and advisors are more likely to leverage its ability to organize records, produce portfolio reports and identify investment opportunities within narrow criteria.
While family offices are open to investing in companies leveraging both types of AI, they are conservative in using AI within their own operations.
A recent survey from Citi Private Bank reveals that more than half (53 per cent) of global family offices have sought portfolio exposure to AI technologies, and an additional 26 per cent are interested. At the same time, however, fewer than 15 per cent said they were using AI in their own operations to automate tasks, build presentations or make forecasts.
The gap between AI’s current capabilities and optimism on what it’s expected to deliver often results from a combination of media hype and a misunderstanding of different types of artificial intelligence.

Machine learning, one branch of AI, analyzes data, identifies patterns and quickly performs complex tasks. It can help advisors and family offices to organize records, produce portfolio reports, develop presentations, identify potential investment opportunities and check compliance.
Generative AI, which powers products such as ChatGPT, Copilot and DALL-E, builds on machine learning and generates content such as text, images, video, music and conversation.
While the overall adoption rate may be low, some family offices and advisors who have embraced AI technology are enthusiastic about its impact on their operations.
Bell Kearns & Associates Ltd., a Toronto multi-family office, recently worked with the Vector Institute, a non-profit that strives to accelerate the use of AI in business. As part of Vector’s FastLane project, Bell Kearns worked with two master’s students in computer engineering generative AI.
“Though we only have 45 clients, we have over 200 strategies that are up and running, so we have what I call the Niagara Falls of quarterly reporting,” said Helen Kearns, CEO of Bell Kearns. “It just pours in here, and we have to distill it. So we want to build a tool for how these different strategies are doing.”
Kearns says that the students are “close” to developing a solution for the office’s reporting needs that achieves compliance using capital markets information while putting a lock on sensitive client data.
Himanshu Joshi, senior project manager of applied AI Programs with the Vector Institute, notes that working with businesses requires an alignment of goals with an array of available AI technologies and then determining which problems AI is most capable of solving.
The institute also stresses that AI innovations must be developed within a responsible and ethical framework. That includes safeguarding sensitive information, applying appropriate government regulations and guardrails to these projects and ensuring protections remain in place down to the end user.

To those ends, Vector offers seminars and boot camps designed to make businesses “AI-ready.”
Because most small family offices have little tech expertise available, Vector will also place AI interns, typically PhD and master’s students, to help, Joshi says.
“In some cases, the collaboration results in new products which can be marketed to other companies,” he adds. “We find that 60 per cent of machine learning associates are hired by the company as permanent employees.”
Family offices are at an inflection point and can no longer afford to be agnostic about the potential of AI to transform their businesses, says Craig Stewart, executive director of applied AI programs with Vector.
“Learn as much as you can and make sure that it’s part of your senior management and board discussions,” Stewart advises. “From a family office perspective, you can’t be in the dark on this stuff any more. I think it’s fair to say that if family offices are not currently examining their options, they’re going to fall behind and lose their competitive advantage, both domestically and internationally.”
Arthur Salzer, CEO and CIO of Northland Wealth Management, says his office has been an early adopter of machine learning systems. He reckons that his company was the first Canadian client to use wealth management software Addepar following its 2009 launch. The product specializes in data aggregation, analytics and portfolio reporting. Northland is also working to automate third-party reporting.
He advises, however, that AI is not an out-of-the-box solution that will automatically deliver efficiency.
“It’s not simply flipping a switch and it’s working from Day One,” Salzer says. “There’s training to it. You have to mold it, shape it and direct it as to how it’s going to be used. It’s almost like directing an intern or a young associate on how to do a job.”

Human oversight and the ability to understand how machine learning software generates reports is also critical.
“We need to know whether there’s a data feed problem or whether something was inputted incorrectly on our end,” he adds. “Did a trade go wrong at a custodian? All sorts of things can happen, and it requires human oversight and direction on our part to monitor results and to keep improving the process.”
In early December, the Canadian Securities Administrators (CSA) published a notice and consultation on the use of AI for registrants with provincial securities regulators, which also stresses the importance of human oversight. The notice stresses: “At the current stage of development of AI systems, we do not believe it is possible to use an AI system as a substitute for an advising representative acting as decision-maker for clients’ investments and consistently satisfy regulatory requirements such as for making suitability determinations or reliably deliver the desired outcomes for clients.”
Managers of large investment funds have long relied on machine learning and algorithmic trading models to create competitive advantage, says Sidney Shapiro, assistant professor of business analytics at the Dhillon School of Business at Alberta’s University of Lethbridge. The CSA notice, he says, is more likely to result in guidelines to ensure advisors don’t rely entirely on AI-generated information to make investment decisions, and that they make explainable decisions that reflect the client’s best interests.
“Generative AI chat apps like ChatGPT will analyze numbers, stocks or securities based on patterns they see in language,” Shapiro says. “That doesn’t make sense, because the patterns built into language may not be the right pattern for what works with numbers. That’s taking a qualitative technology and trying to overlay it on top of quantitative analysis.”
For those who might be inclined to believe that readily available generative AI services can dispense valuable investment advice, Shapiro suggests posing a simple spelling question. For example, how many times does the letter “d” appear in the word “dividend”? It’s an easy question for humans, but ChatGPT has a hard time of it, guessing “two times.” That’s because it can’t find an online reference where that question has already been asked and answered.

“AI in this context isn’t thinking for itself—it’s just a massive recycling machine that looks to the past and to what’s already happened,” Shapiro adds. “As investment ads say in the disclaimer, ‘past performance doesn’t dictate future results.’”
He compares the use of ChatGPT to generate financial advice to a doctor using it to diagnose a patient.
“Generative AI will tell you what’s wrong with most people, not necessarily what’s going on with me,” he says. “Like a good doctor, a good financial advisor will offer advice based on a combination of reliable information, their life experience, all of the things they’ve learned, all the trades they’ve made, and the unique needs of their client. “That can’t be replaced by AI.”
The Canadian Family Offices newsletter comes out on Sundays and Wednesdays. If you are interested in stories about Canadian enterprising families, family offices and the professionals who work with them, but like your content aggregated, you can sign up for our free newsletter here.
Please visit here to see information about our standards of journalistic excellence.