Thomson Reuters to roll out in-house AI model as quarterly revenue rises
Thomson Reuters (TRI-T) said it will deploy its in-house AI model, “Thomson,” for legal professionals after benchmarking showed strong performance versus models from OpenAI, Anthropic and Google DeepMind, according to the company. Second-quarter revenue rose 9% to $1.95B, profit was $448M ($1.02/share), and adjusted EPS was 99 cents. It raised its full-year revenue growth outlook to about 8%.
How this was made
The 30-second read
Why it matters
This is a combined earnings and product roadmap update. The raised revenue growth target and the planned AI integration into CoCounsel provide near-term fundamental support, while the stock’s large same-day decline shows the market is still pricing disruption risk and uncertain AI monetization.
Market read
Traders get a fresh earnings print, a guidance raise, and a specific AI rollout timeline, which together can drive positioning in enterprise software and legal-tech exposure.
What to watch
The article notes only limited training on legal content (about 8%) and a retraining plan; traders may need to watch whether performance improvements and customer adoption materialize fast enough to justify the AI spend.
Background
Thomson Reuters is building an in-house large language model (“Thomson”) with Safe Sign Technologies, acquired last year, and is preparing to deploy it for legal professionals.
Ticker impact
Thomson Reuters reported Q2 revenue up 9% to $1.95B, raised its full-year revenue growth target to about 8%, and detailed deployment of its in-house “Thomson” AI model.
Bias modestly positive over days to weeks if investors focus on guidance and AI productization; near-term volatility likely given disruption fears.
The article provides fresh, decision-relevant datapoints: Q2 results, adjusted EPS beat, raised revenue growth target, and a scheduled AI integration (connect to CoCounsel later this month). The 10.2% drop suggests the market is discounting disruption risk, but the disclosed fundamentals and roadmap are new.
Market effects
Reinforces the legal-tech and enterprise software narrative that proprietary AI models can be productized, potentially raising competitive pressure on AI-agnostic vendors.
Supports Canadian large-cap software sentiment via a major earnings and guidance update from a Toronto-listed name.
Highlights a global AI build-versus-buy strategy in regulated knowledge work, relevant to other legal, tax, and compliance platforms worldwide.
Counterpoint
The market selloff implies investors may doubt monetization timelines or worry that frontier-model benchmarking does not translate into durable pricing power.
Key entities
- companyThomson Reuters
TRI, preparing to deploy its proprietary “Thomson” AI model and reporting Q2 results with raised full-year revenue growth guidance.
- companySafe Sign Technologies
British startup Thomson Reuters acquired last year to help develop the in-house AI model.
- companyKKR & Co. Inc.
Buyer of a majority stake in Thomson Reuters’ global print division for $500 million, freeing cash for AI-enabled products.
