TORONTO —Damien Steel was hired to bring more AI into one of Canada’s largest institutional investors. But in his first weeks at the Healthcare of Ontario Pension Plan (HOOPP), he’s spent less time thinking about technology tools than about the people who will use them.

Last month, the longtime startup investor joined the retirement savings organization as chief technology, data and AI officer—the only top-level leader at one of the so-called Maple 8 with artificial intelligence in his title. Unsurprisingly, Steel is bullish on what the technology could do for HOOPP. But first he’s got to get the firm’s internal systems ready, and get more employees using AI tools.

Many large organizations are looking to AI as a means to “increase efficiencies, drive down costs, [and] make people faster at what they do,” Steel said, in his first interview since taking the job. HOOPP wants those benefits, but it’s also trying to change how it operates so it can “quickly adopt and use any technology of the future.”

One early move is updating the way the firm brings in new tools. Staff asked for some 85 new pieces of technology in 2025, up 145 per cent year-over-year, and the pace of requests has already doubled this year. HOOPP plans to use AI agents to speed up a security and risk vetting process that currently takes about two months. 

Steel is also trying to ensure that employees can experiment with AI tools, and that they produce useful results. HOOPP sped up work to consolidate all of its data into a single software system, and now aims to have it online in one year instead of three. “Once you do that, it unlocks everything,” Steel said, because AI tools are only as good as the information to which they have access.

The firm has also launched an “AI sandbox,” a safe space in which staff can learn how to use chatbots and agents without damaging the firm’s software systems or operations. It’s more effective than top-down training, Steel said. “You allow your team to be creative.” Rather than tracking which employees use AI the most with a leaderboard, HOOPP has softer metrics like whether all workers have access to the technology and how many are trying it.

The 1,100 or so staff at HOOPP can currently choose between OpenAI’s ChatGPT, Anthropic’s Claude Cowork and Microsoft’s Copilot. Those are leading products, but all are sold by American tech giants. Canadian tech founders have long expressed frustration that corporate Canada doesn’t buy from them, and the country’s institutional investors don’t back them.

HOOPP’s primary responsibility is to its members, said Steel. “I’m not going to use a product at the risk of impacting our results in our business.” Still, when the firm is considering buying new technology, “I think we owe it to look in our own backyard,” he said, citing the benefits of having tool makers in close proximity. 

HOOPP is a direct investor in Toronto-based Cohere, which sells a tool called North with similar capabilities to Cowork or Copilot; financial services is among the startup’s target sectors. HOOPP is currently assessing whether Cohere’s technology could be useful to it. 

While Steel acknowledged that many of HOOPP’s AI efforts are similar to lots of large organizations, he said defined-benefit pension plans face a unique set of risks and opportunities. The firm must keep paying out to members no matter what; if AI-driven improvements in health care increase lifespans, its liabilities will grow. But AI can also help improve investment returns and cut costs, leaving HOOPP more money to pay pensions. 

So far, the firm’s deal making teams are employing AI to understand the downstream effects of major events and trends like geopolitical tensions or the data-centre buildout. For example, HOOPP used AI code generation to help build a tool that its portfolio managers used to monitor and analyze the impacts of the Iran war. Human investors still make all the decisions about capital allocation, Steel said. 

HOOPP also plans to leverage its relationships with the top-tier asset managers that help invest its money to find new uses for AI, and to get more data to inform its own deals. For example, Steel has been studying how Brookfield uses AI in its private equity portfolio, and what Bridgewater’s AI tools for active trading can do. 

Longer term, he hopes some of the fund managers that HOOPP backs will share more market data. Brookfield, for instance, has great insight into what’s happening with privately held energy assets around the world. A richer pool of information will expand what HOOPP can do with AI. For example, the firm could use agents to reconcile the data it gets from third-party fund managers—which often use different timespans and measurements—so it can better compare their performance. “The way we evaluate deals probably changes,” Steel said. 

He also hopes AI will eventually flatten the internal hierarchy at large organizations, making new recruits as effective as workers with decades on the job. The technology will let “somebody with less work experience to drive more value,” Steel said, adding that the firm will also need to adapt its norms around performance and seniority. HOOPP hasn’t yet worked out what it would mean for evaluations and compensation, though.

Steel readily acknowledged that his own experience is as a technology investor, not an engineer or developer. He’s perhaps best known for his long tenure as partner at OMERS Ventures, including five years as its leader. Steel was later CEO of Deep Sky, a Montreal-based carbon capture startup, for two and a half years, departing last spring for personal reasons.

He joined HOOPP—where he will also advise on tech investments—after a few months at Whitecap Venture Partners. 

Steel argued his background working with startups actually makes him well-suited to his new job. Newly minted, innovative firms are “very comfortable operating in an environment of constant change,” he said. As AI accelerates, every organization, he claimed, will need to learn to do the same.