The firm shaped agentic AI before adopting it – now the platform runs across nearly 1,000 lawyers
In less than two years, Fasken's clients have gone from telling the firm not to use AI on their work to expecting it and demanding better outcomes as a result. The technology driving that shift – agentic AI – is also what Clarke Barnes, Fasken's firm managing partner, says the firm has been building toward for longer than its July Legora partnership announcement suggests.
Fasken's relationship with the technology behind Legora predates that announcement by more than a year. Through a joint innovation program with Walter AI that began in 2024, Fasken lawyers contributed directly to the development of the Walter platform and helped shape its emerging agentic capabilities before Legora acquired the company in March 2026. Other major Canadian firms have since moved in the same direction – McMillan completed its own firmwide Legora adoption earlier this year – but Fasken's path was different: it built before it adopted.
"We really see Legora as a continuation of that AI strategy and journey," Barnes says. "Prior to the announcement of adopting Legora, we had already developed and deployed several legal AI tools along with Walter. We were experimenting with bespoke use cases and monitoring developments of several AI legal platforms."
Starting where automation made sense
The entry point for the Walter AI work was Fasken's emerging technology and venture capital group. The choice was deliberate: that work is structured and repetitive enough to suit early-stage automation without the complexity that would have made agentic tools unmanageable at the outset.
"The work itself was really suited to that concept of automation and how you could use an agent to improve that automation with the functionality that agentic AI brings to the table," says Andrea Alliston, who leads knowledge and practice innovation at Fasken.
"The effort has to be on both sides with a common goal," she says. "We're very fortunate with the Legora acquisition to be able to continue that partnership, because innovation and being creative and being thoughtful about different ways of practising is very much in the culture of Fasken."
What agentic AI changes
With the Legora rollout now underway across nearly 1,000 lawyers in Fasken's corporate commercial and litigation practices, priority use cases developed with practice group leaders include due diligence review and examination for discovery transcripts – tasks chosen for their potential for high-volume, high-impact deployment.
What changes with agentic AI, Alliston says, is the question itself: not what a system can answer, but what it can execute – and where it cannot act without human review. "You don't want to be sending a communication outside the firm without having someone look at that and be comfortable with the advice," she says. "It is a new way of thinking about the way work is done and making sure that it still resonates with the lawyers, because they are the ones [who] are using the tools."
What interests Alliston more is what the tool does to thinking itself. "It doesn't take away the judgment or expertise or experience of a lawyer, but it just helps them have access to a broad range of perspectives if you know how to use the AI to get those to surface."
Choosing the platform
Fasken began its formal evaluation of the legal AI platform market in late 2025, running candidates through a requirements-based assessment followed by a large pilot. Legora's full bilingualism was a decisive factor at the time of selection. "We are a bilingual firm, and it is really important that all of our lawyers be able to work in the language of their choice," Alliston says, noting that that competitive advantage has since levelled across platforms.
She adds that the pace of change in the space made the decision genuinely difficult. "In my time in this space, in the innovation and KM space, I've never seen applications evolve so quickly," she says. "You're constantly going to feel like, am I making the right decision?"
The Walter acquisition, announced while the evaluation was in progress, provided added confidence by clarifying Legora's agentic development roadmap. "We could see that roadmap a bit more clearly given our experience with Walter, and that was really important to us," Alliston says. Harvey, Legora's best-funded rival in the legal AI space, has been pursuing a parallel strategy of aggressive expansion in the Canadian market.
Making adoption the priority
Both Barnes and Alliston are emphatic that the technology is the easier part. Adoption – changing how lawyers work at the practice level – is where firms consistently underestimate the investment. Barnes says it comes up at every managing partner roundtable he attends.
"You have to be willing to commit the training time personally, almost sometimes on an office-to-office visit basis," he says, "to make sure that more and more lawyers and more and more of our staff are comfortable with it and really can explore what it can do."
Two structural programs support that effort: a secondment program placing lawyers into Alliston's team for a year at 50 percent of their time to build practice-specific Legora workflows, and a new cohort of AI engagement lawyers providing hands-on support across the firm.
Barnes pushes back on the industry's prevailing anxiety about AI displacing legal work. "You can doom scroll quite a lot about the legal industry, and management is understandably worried about efficiencies and loss of hours and that sort of thing. I think that's pessimistic, too pessimistic," he says. "We're busy, we're busier than we ever have been."
Alliston places the technology's real promise beyond the work firms are already doing. "We will start to see more interesting opportunities to deliver services to clients, not just through the process, but actually being able to do things for clients that we can't actually do today because it's just not cost [effective] with humans doing the work, or it involves data that needs to be processed, and you can gain insights from that data."
This article is based on an episode of CL Talk, which can also be found here:
The episode is also available on our CL Talk podcast homepage, which includes links to follow CL Talk on all major podcast platforms.
