Among Canada’s lawyers, the jury is still out on AI

Nearly 90% of Canadian lawyers said their firm is piloting or integrating AI tools, with some tasks reportedly cut from 10 hours to 1 hour and administrative work reduced by AI. The article highlights potential fee compression through fixed-fee pricing, but also significant risks around confidentiality, hallucinated legal research, and errors in client work. Near-term market impact appears limited, though the piece suggests AI could materially reshape legal-services economics over time.

Analysis

The key equity implication is not “AI helps lawyers,” but that it compresses the economic moat of labor-heavy legal services and shifts value toward software, workflow, and data-governance providers. If routine research/drafting becomes a near-zero marginal-cost function, pricing power migrates away from firms that sell hours and toward platforms that own the workflow layer, compliance rails, and secure document infrastructure. That is structurally constructive for legal-tech enablers, but it is also a margin headwind for incumbents that depend on junior associate leverage to scale revenue.

The second-order loser is the training pipeline: if entry-level work is automated, law firms may face a slower decline in headcount than a sharper decline in associate productivity per seat. That creates a longer-dated operating risk, because partner revenue can hold up while the internal talent funnel erodes; the downside shows up later as weaker succession quality, higher partner comp pressure, and potentially more lateral hiring costs. In other words, the near-term P&L impact can look benign even as the business model degrades underneath.

For Thomson Reuters specifically, the market likely underestimates the two-sided nature of AI adoption. Near term, AI raises the value of trusted legal content, citations, and embedded workflow because firms want auditable tools after hallucination incidents; that should support retention and upsell. Over 12-24 months, though, if large language models commoditize research interfaces, TRI’s moat shifts from content access to distribution and verification, and any miss on product monetization could re-rate the multiple.

The contrarian view is that fee compression may be slower than bulls expect because legal work is still trust-constrained, not speed-constrained. Clients may accept AI internally at firms without seeing lower prices, which means the first beneficiaries are margins, not consumers; that delays the demand elasticity thesis. The real catalyst to watch is regulation and malpractice litigation: one prominent AI error case can accelerate mandatory disclosure, logging, and human-review requirements, which increases spend on compliant, enterprise-grade tools and penalizes casual usage.

Key Decisions for Investors

• Long TRI on a 6-12 month horizon if the market is pricing AI as a moat-destroyer rather than a workflow-winner; use pullbacks as entry, target upside from higher attach of verified/legal-grade AI products, with downside capped by recurring content subscriptions.
• Buy a basket of enterprise legal-tech / workflow beneficiaries vs. short labor-arbitrage legal services exposure over 3-9 months; thesis is that AI value accrues to audited infrastructure, not billable-hour firms.
• Consider a medium-dated TRI call spread into the next product cycle; risk/reward favors upside if the market keeps underestimating monetization from AI-assisted research and compliance tooling, while premium is limited if adoption stalls.
• Avoid outright long positions in hourly-billing legal-service models; if AI adoption persists for 6-18 months, margin pressure should gradually surface before revenue growth does, creating a delayed but meaningful multiple compression risk.
• Set a catalyst watch on any publicized AI hallucination or disciplinary case in legal; that would be a near-term positive for trusted vendors like TRI and a negative for generic AI tool adoption in the profession.