Thomson Reuters Grew 10% on AI, but the Stock Fell. Here Is the Trap
Thomson Reuters posted higher revenue and raised its full-year outlook, but shares declined. The result shows that AI software depends less on the technology than on what customers will pay and what profit remains.
The 20-second read
1. Why the stock fell on good results
Revenue rose. The full-year outlook went up. The stock still fell. That was the scene after Thomson Reuters (TRI) reported earnings on Nov. 3.
| Metric | Figure | Note |
|---|---|---|
| Third-quarter revenue | $1.83 billion | Up 10% from a year earlier |
| Adjusted EPS | $0.81 | Third quarter of 2026 |
| Full-year organic growth outlook | 8.5% to 9.0% | Raised from about 8% |
On the numbers, the business is heading the right way. The issue is the price of that growth. As the cost of building AI features and rolling them out across products rises, higher revenue may not translate directly into higher profit.
The key is not whether AI revenue exists, but how much of it is left as profit.
2. If AI cuts the hours worked, who makes the money?
Thomson Reuters owns the legal research service Westlaw and CoCounsel, an AI tool for legal work. The new version, CoCounsel Legal, aims at more than giving a lawyer the answer to a single question.
The workflow CoCounsel Legal targets: case research → drafting a brief → rechecking citations
The idea is to connect work that lawyers handle in several separate steps into one flow. For the client, a law firm, the value is saved working hours.
That raises an interesting question. If a law firm bills its clients for the hours it works, what does faster work do to the existing billing model? And if AI software is sold by the seat, a big jump in productivity may not lift the software company's revenue at the same pace.
Who captures the savings AI creates changes the economics for both the law firm and the software vendor.
AI productivity gains and AI vendor revenue growth are not the same thing.
3. Thomson Reuters' real weapon is not the AI model
The value of AI software for professionals is not set by how fluently it writes. Legal documents need exact precedents, verifiable citations and procedures that fit real work.
Thomson Reuters' strengths are the legal data it has built up over many years, specialist services such as Westlaw, and the work environment lawyers already use. Apart from the race to improve AI models, the ability to turn trusted material into finished work product can be a competitive edge.
High customer value does not automatically guarantee high profit, though. The cost of serving customers can also rise as AI use grows. If a pricier product brings equally higher costs, the gain in profitability is limited.
The moat in specialist software looks less like model performance alone and more like a mix of data, workflow integration and trust.
4. The same question for Microsoft, Salesforce and Adobe investors
The issue does not stop at legal software. It is an investing frame that also applies to software companies embedded in customers' daily work, such as Microsoft (MSFT), Salesforce (CRM) and Adobe (ADBE).
More AI features create chances to sell products at higher prices or offer existing customers add-ons. At the same time, development and service costs weigh on margins. Whether the traditional per-seat subscription still fits also deserves a look. The more real workload AI handles, the more customers may price a product on the value of completed work rather than on a simple right to use it.
So a company that owns a good AI product should be told apart from a company that runs a good AI business. Firms whose customers pay extra for AI, and whose revenue from it grows faster than costs, are likely better placed.
Software investors should look at pricing power and the ability to convert revenue to profit, not at the number of AI features.
5. Four numbers to watch next
- Not necessarily. If customers value the time saved and pay extra for it, AI providers can grow too. The key is how much of that value they can build into their prices.
- It is better not to assume it will disappear right away. The thing to watch is whether an AI premium, or pricing based on the volume of work processed, gets added to existing seat fees.
- Data and trust help with customer retention and pricing power. But if development costs and the cost of delivering AI services rise quickly, profitability can come under pressure.
- Do not judge the quality of growth by the revenue growth rate alone. You need to check AI monetization, revenue per customer, costs and operating margin together.
What to watch
Insight Times Editorial Desk
