Thomson Reuters CTO Joel Hron argues that curiosity, adaptability and rapid learning will help employees navigate AI-driven change because neither companies nor workers can reliably predict which products and workflows will matter next. His March 16, 2026 interview with ITPro offers a view of a professional-information company trying to pair generative AI with trusted legal, tax and compliance expertise. It describes a strategy and ambition, not proof of improved productivity, product accuracy or commercial success.
Why Thomson Reuters is a revealing AI case
Thomson Reuters operates in legal, tax and accounting, compliance and risk, and news and media. Those fields rely on specialized information and consequential decisions. An AI system that produces a confident but incorrect answer can cause more than inconvenience: it may omit relevant authority, misread a document, overlook jurisdictional context or expose confidential information.
That makes professional AI different from a general-purpose chatbot. A capable language model is only one component. The quality and provenance of the information it uses, how it retrieves and cites sources, whether it respects permissions, and how its output fits into an accountable professional workflow can matter just as much.
Hron’s challenge, as presented in the interview, is to move quickly enough to respond to generative AI while protecting the trust customers associate with Thomson Reuters’ content and products. Speed can help a company keep pace with changing expectations; in high-stakes work, however, novelty is not a substitute for dependable results.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problems#1 Best Overall
Who is Joel Hron?
ITPro reports that Hron became Thomson Reuters’ CTO in July 2024. Before that, he led AI at Thomson Reuters Labs and served as a vice president of technology. He previously was CTO of ThoughtTrace, a company Thomson Reuters acquired in 2022. His current remit includes product engineering, AI and research and development; he reports to Kirsty Roth, chief operations and technology officer, according to the interview.
That career puts Hron between startup-style experimentation and a large professional-services technology organization. It does not mean he alone created Thomson Reuters’ AI products: the interview describes the work of teams and a broader technology organization, which Hron puts at approximately 5,000 people.
How generative AI changed the company’s priorities
Hron says Thomson Reuters’ priorities shifted during and after the completion of the ThoughtTrace integration in late 2022, as generative-AI products began reaching the market. He describes Thomson Reuters Labs as a strategic center for shaping the company’s AI approach and early products.
The business problem is not simply how to add a chat box. Professional customers need tools that can help with research, analysis and document-heavy tasks while making it possible to check the supporting material and retain human responsibility for decisions. A superficial AI feature may signal that a company is keeping up without materially improving a customer’s work. The harder test is whether a product makes a defined workflow more useful, reliable or efficient.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
What AI products does Thomson Reuters identify?
In the interview, Hron names Westlaw Advantage and Deep Research as notable achievements, and refers more broadly to AI-enabled products across legal, tax and compliance. The interview gives limited technical detail about Deep Research; Hron describes it as a system that reviews and strategizes in a manner similar to a researcher. That description is an executive characterization, not evidence that it performs at the level of a human researcher.
Thomson Reuters’ AI portfolio page presents products for distinct professional settings, including CoCounsel Legal, Westlaw Advantage, Westlaw Edge, CoCounsel Tax, CoCounsel Audit, CLEAR Investigate and Global Classification AI, alongside other tax, legal, compliance and risk offerings. These are not one unified system: they target different users, content, workflows and buying decisions.
Rank #2
The company’s Westlaw Advantage product page describes the product as using agentic AI and verified Westlaw content. That is Thomson Reuters’ product positioning; it does not by itself establish accuracy, adoption or customer returns. The interview supplies no usage, revenue, retention, productivity or comparative-performance figures for Westlaw Advantage or Deep Research.
What “model agnostic” means—and what it does not prove
Hron says Thomson Reuters combines internally developed models with off-the-shelf tools and uses internal specialists to manage and control that combination. He describes the company as taking a model-agnostic approach to large language models. In practical terms, the strategy is presented as avoiding dependence on a single model provider and leaving room to choose tools for different tasks.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchThat flexibility could make it possible to weigh performance, cost, latency, privacy, jurisdiction and task requirements when selecting a model. But the interview does not disclose model names, supplier contracts, routing architecture, benchmarks, error rates, security controls or data-retention practices. It therefore cannot establish that model agnosticism has reduced costs or improved accuracy.
Nor does model choice alone make a system dependable. Retrieval and grounding, source citations, permissions, evaluation, workflow controls and human review remain important. Proprietary legal, tax, compliance and news content may be a strategic asset, but only if a product uses it appropriately and lets professionals assess the basis for its output.
Why legal and tax attract AI investment—and carry risk
Hron points to legal and tax as focal areas for disruption because they involve large volumes of structured and unstructured information, research, drafting and repeatable professional work. AI may help retrieve information, analyze documents, prepare first drafts, synthesize research and reduce administrative effort. It may also help firms with limited staff handle more work or make specialized knowledge easier to access.
The same features create risks. A generated legal or tax answer can be wrong, incomplete, outdated or unsuitable for a particular jurisdiction. Users may over-trust fluent output, overlook missing sources, enter confidential material into an inappropriate system or fail to preserve a clear record of how a consequential answer was produced. Errors can create professional-liability and regulatory concerns.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Rank #3
For that reason, the near-term value proposition is better framed as assistance than as a replacement for professional judgment. A qualified person still needs to assess the context, verify important claims and take responsibility for decisions. The interview does not explain precisely how human review works in Thomson Reuters products, so it should not be read as evidence that every consequential output receives a particular form of review.
Adaptability as a talent strategy
Hron’s central workforce argument is that the next 12 months of product and workflow change are difficult to predict. He says curiosity, adaptability, rapid learning and iterative delivery will matter, and that this outlook is influencing hiring, recruiting and team organization. He also describes engineers sharing internally built experiments and prototypes.
Put into observable terms, adaptability can mean learning unfamiliar tools, working with incomplete information, testing an idea and revising it, asking where customers experience friction, and critically evaluating AI output rather than accepting it. It can also mean collaborating across engineering, product, research and professional teams, transferring knowledge between acquired organizations, and redesigning a workflow instead of merely automating its existing steps.
These are reported cultural examples and management beliefs, not independently audited measures of engagement or productivity. The interview does not show that adaptability predicts AI-era performance better than domain expertise, experience or management quality. Nor does it document specific changes to hiring criteria or employee training.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Why engineering is central to the transformation
Hron argues that strong software engineering will support changes to traditional legal and tax business models. The point extends beyond building a model-powered feature. AI can alter search, retrieval, drafting, document handling, review, collaboration and customer support; engineering determines whether those capabilities become part of a dependable product or remain an isolated demonstration.
That work may involve connecting models to proprietary content, permissions, citations, audit trails and interfaces professionals can use. The competitive distinction may therefore lie less in access to a language model than in the engineering needed to embed AI in a trusted system. Hron’s interview describes this direction but does not disclose a technical architecture or measured outcomes.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the company still needs to prove
Hron says Thomson Reuters wants the market to see it as broadly innovative and market-leading, not innovative only in particular product areas. He also acknowledges more work is needed and says the company wants to release products that make customers and competitors think differently. That is a stated competitive ambition, not independent confirmation that Thomson Reuters has achieved market leadership.
The company brings established content and customer relationships to the contest. Startups and technology companies may experiment quickly, while some professional customers may place greater weight on reliability than novelty. Thomson Reuters must show that its products improve real work and find a commercial model that makes sense without weakening the perceived value of existing subscriptions. The interview provides no customer-pricing, financial-return or adoption data to resolve those questions. Hron’s expectation of significant developments over the following 12–24 months is a forecast, not a confirmed product roadmap.
Useful evidence of progress would include customer adoption and retention, measurable workflow outcomes, accuracy and citation performance, disclosures about how errors are handled, and clear evidence that customers find the products worth paying for. For an organization deploying professional AI, other practical questions include who is accountable for outputs, how employees are trained, how experiments are evaluated and stopped, and whether data and access controls fit the work. The interview does not answer these questions in detail.
Where an adaptability-first approach can fail
Adaptability is valuable, but it is not a complete operating strategy. Without the conditions below, rapid experimentation can consume resources without producing safe, maintainable products:
- Experiments without customer value: prototypes generate interest but never become repeatable products that solve a defined problem.
- Model-centric decisions: teams debate which language model to use while neglecting data quality, retrieval, permissions, evaluation and source traceability.
- Unverifiable or unaccountable output: users cannot check relevant sources, identify a responsible person or understand how a consequential answer was produced.
- Change without support: employees face continuous shifts in tools and priorities without adequate training, clear success measures or time to learn.
- Enthusiasm mistaken for delivery: a high volume of prototypes is treated as proof of secure, useful, maintainable products.
- Expertise displaced by a vague hiring label: adaptability is valued at the expense of legal, tax, security, data or engineering fundamentals.
- Unmanaged integration: acquired teams bring conflicting processes, tools or incentives that make it harder to share knowledge and ship coherent products.
Strong teams are more likely to combine domain specialists, software and data professionals, product managers, security and privacy experts, user researchers, and people responsible for professional risk and governance. Adaptability can help those teams learn; it cannot replace the expertise and controls their work requires.
Quick Recap
What other enterprises can take from Thomson Reuters’ approach
- Start with a consequential workflow, not a model. Identify a specific customer or employee problem and define the work the system should support.
- Bring domain knowledge and data into the design. Determine what information the system may use, whether users can inspect its sources, and how access permissions apply.
- Choose tools for the task. Evaluate available models and other components against the workflow’s requirements rather than assuming one model fits every use.
- Set evaluation and oversight before scaling. Define how the system will be checked, what happens when it fails, and who remains accountable for consequential decisions.
- Hire for learning without compromising expertise. Pair curiosity and iteration with professional knowledge, technical fundamentals and sound judgment.
- Make experimentation measurable. Give teams room to test ideas, but use customer feedback, safety requirements and business outcomes to decide whether to continue, change or stop.
- Revisit the operating model. As tools and workflows change, review how roles, training, governance and product development need to adapt.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
