Investors are scrambling to figure out how to value the next generation of AI companies—but they don’t seem to be grasping what’s going on inside Thomson Reuters, at least in the eyes of CEO Steve Hasker. 

If the Nasdaq were a rollercoaster, Hasker’s ride has been nauseating. Shares plunged to a 52-week low to under US$80 in late February, after Anthropic’s new AI assistant triggered a sharp sell-off in legaltech stocks. Then sentiment flipped, with shares surging past US$110 by the first week of March, after Anthropic’s integration with Thomson Reuters inspired investors to bid up the firm’s shares, which remain majority-owned by the Canadian Thomson family empire. The market whiplash—dubbed the “SAASacre” by CIBC analysts—has renewed the debate over AI’s winners and losers. 

Meanwhile, Thomson Reuters has faced scrutiny, particularly from its workers in Minnesota, about a contract between its U.S. subsidiary and U.S. Immigration and Customs Enforcement, which is scheduled to end in May 2026.

It has been a test for Hasker, the Toronto-based firm’s turnaround man, who defended his strategy in an interview. 

It has been edited for both length and clarity.

Q: There’s been some market volatility that left you with a short end of the stick. Do you think the company was mispriced?

Yeah, I do.

Q: What do you feel like the market missed? 

I’d make a very clear distinction between three layers of sort of professionals and AI that’s suitable for each. For knowledge workers more broadly, there’s a set of AI tools that sync your email with your calendar and your to-do list: an executive assistant for office workers, like [Microsoft] Copilot and [Claude] Cowork.

There’s a next layer for professionals. A Thomson Reuters salesperson will have access to an AI-driven sales support tool. If you’re a journalist at Reuters, you’ll have access to what’s called Fact Genie which we’ve developed.

The third layer, which I think is very much missed by the markets recently, is fiduciary professions like legal, tax and audit, where you have to be right. You can’t rely on a foundation model that hallucinates three per cent of the time and is not obsessed with data privacy. You can get fined, you can get disbarred, you can end up in jail.

We serve fiduciaries, not a general set of knowledge workers.

Q: Which of your businesses is changing the fastest?

The legal profession is probably the most prominent. Tax accounting is not far behind.

Q: Is there a wave of consolidation coming in the legal and financial tech industry?

I think so. We’re optimistic we’ll see some opportunities. We’re not relying on it. But we’re entering a new phase in terms of the valuation expectations, or at least I’m hopeful of that from the venture capitalists and from the private equity sponsors.

We look for businesses that are not only financially accretive for our shareholders, but [add] to our culture. We’ve had a good track record of keeping the key talent within the half dozen meaningful acquisitions over the last couple of years. 

Q: [Former board member] Ed Clark has said that internally, Thomson Reuters is asking people which parts of their job they dislike, and is focusing on automating those tasks. Your headcount has stayed pretty steady over this past year, despite fears about knowledge worker layoffs. What’s your strategy on hiring and retraining workers?

In any job, there is a mundane part that people don’t relish. There’s a real opportunity to use AI  to free people up. We launched what we call a “Reimagine” effort. It’s enabled us to reach 85 per cent daily usage of our AI tools. 

Community and society-driven aspirations are very, very appealing to attract young, digital and AI talent. We are a company that helps provide access to justice and uphold the rule of law. We’re a company that helps create functioning taxation systems, which helps catch bad actors through the work that we do with law enforcement. 

We’ve got the Reuters news agency, reporting the news often where it’s difficult and dangerous. It’s not been more relevant and powerful than to see the photograph of Prince Andrew, taken in the back of a car by Phil Noble, one of the Reuters photographers. Or the coverage of the Iran conflict. 

Q: You did a human rights review a few years ago. Now the [U.S.] law enforcement business, and specifically immigration enforcement, is back in the news again. How has your view of that business evolved? What role will that business play in the future of the company?

A: Our government business is a small but important part. We’re very proud of the work that our government colleagues perform in serving law enforcement. It really is focused on drug trafficking, is focused on child sex exploitation, and it’s focused on data-driven support for law enforcement in tackling the worst of the worst, in terms of gangs and cartels. That work has been ongoing for multiple administrations, and it continues. That’s the real focus. 

Q: Is there anything Canada can do to better support the ethical advancement of AI?

We’ve got a front row seat to how powerful these models can be. We’ll launch it formally [this year], but we’ve developed our own foundation model specifically for legal. I think Canada’s role is as a country with tremendous talent, homegrown talent, immigrant talent, first and second generation; a country with tremendous universities. Canada can and should take a role in both promoting the development and the speed of innovation around AI, but also framing responsible AI. 

Editor’s note: After this story’s publication, a Thomson Reuters spokesperson said that while the company will launch its foundation model this year, it has not yet chosen a date. The story has been updated.