Artificial intelligence (AI) can now write a contract, summarize a lawsuit, explain a tax question, and confidently invent something that doesn’t exist.
That final skill creates an opportunity.
The first phase of the AI boom rewarded companies supplying chips and computing power. The next phase could reward businesses owning information professionals are willing to trust. A lawyer doesn’t need another chatbot capable of sounding correct. The lawyer needs the correct case.
Source: Getty Images
Data becomes the scarce asset
Large AI models are becoming increasingly capable and increasingly available. That can make the model itself less differentiated. Professional data is different. Legal cases, tax rules, regulatory decisions, accounting standards, citations, and years of curated professional information are expensive to collect and maintain.
They also need to be current. A general-purpose AI answer may be useful when planning dinner. A hallucinated tax ruling can have more expensive consequences. That’s why I’d watch Thomson Reuters (TSX: TRI).
More than a news company
Thomson Reuters sells professional information and software to lawyers, corporations, accountants, tax professionals, and governments. Products such as Westlaw, CLEAR, Checkpoint, and CoCounsel sit directly inside professional workflows.
That positioning gives Thomson Reuters an interesting AI advantage. It doesn’t need to convince professionals to abandon trusted databases for AI. It can put AI inside the trusted databases they’re already paying for.
Second-quarter revenue increased 9% year over year. Organic growth was 8%. Its Legal Professionals, Corporates, and Tax, Audit & Accounting Professionals businesses, the company’s “Big 3,” delivered 10% organic growth. Management raised its full-year Big 3 organic-growth outlook to between 9.5% and 10%.
AI moves
Thomson Reuters has launched CoCounsel Legal and developed the first production-ready version of its own Thomson LLM. Management is increasingly describing its goal as “fiduciary-grade” AI. These are systems designed for professional work where accuracy and source quality carry consequences.
Recurring revenue represented 82% of company revenue during the second quarter, and that’s important. Thomson Reuters can fund AI development while customers continue paying subscriptions for the information and workflows underneath it.
The company is also separating its Global Print business through a joint venture with KKR, receiving approximately US$500 million in gross proceeds. That leaves the remaining business increasingly focused on higher-growth digital information, software, and AI. For investors searching for Canadian growth stocks, that’s a cleaner long-term story than printing legal textbooks.
Considerations
TRI recently traded at $136.61, yet its 52-week high is above $233. That’s a decline of roughly 42% from the peak. The stock still trades around 25 times trailing earnings, so “down a lot” doesn’t automatically mean cheap.
Competition is also serious. OpenAI, Google, Microsoft, Anthropic, and future model providers can keep improving professional products. Large law firms and corporations may build their own systems.
Thomson Reuters must prove customers will pay more for its AI capabilities rather than simply expect them as part of an existing subscription. That’s the metric I’d watch after buying stocks in Canada. Does AI increase contract value and retention, or merely increase Thomson Reuters’s development bill?
Bottom line
The cheapest part of AI may eventually be generating an answer. The expensive part could be proving the answer is right.
Thomson Reuters already owns decades of professional content, trusted brands, recurring customer relationships, and workflows where accuracy carries a price.
The stock has fallen enough to make that opportunity far more interesting. Therefore, the next AI winner doesn’t necessarily need the smartest model. Owning the information that keeps the model honest may be enough.
