For decades, software steadily transformed nearly every facet of modern life, reshaping global industries and redefining how people work and interact. That era, however, is evolving once again. Today, artificial intelligence—the newest and most formidable technological wave—is beginning to consume the very software that once consumed the world. Nowhere is this transformation more evident than in the field of legal technology, where Thomson Reuters, a colossus of legal information and research tools, is striving with determination to ensure it remains a dominant player rather than becoming just another meal on AI’s ever-expanding menu of disruption.
Thomson Reuters commands one of the most comprehensive and meticulously curated repositories of legal data in existence. This immense archive spans millions of cases, statutes, rulings, and other primary sources of law, representing decades of editorial refinement. By merging this unparalleled data collection with the capabilities of generative AI models, the company is crafting a new generation of sophisticated products designed to make legal work faster, smarter, and increasingly automated. Two of its flagship offerings, Westlaw Advantage and CoCounsel, exemplify this shift by integrating natural language processing, semantic search, and advanced reasoning to handle tasks once performed exclusively by human lawyers—ranging from drafting assistance to the synthesis of complex case histories.
Earlier in the year, these ambitious products attracted considerable enthusiasm from investors who saw Thomson Reuters as a rare established player well-positioned to harness generative AI’s commercial potential. Yet, as the months passed, summer brought an influx of formidable new entrants into the field—ventures flush with capital and ambition. Their rapid growth rekindled doubts among observers about the durability of Thomson Reuters’ competitive edge and the sustainability of its AI-driven strategy.
A particularly thorny question lingers in the background: Can CoCounsel truly produce legal reasoning and insights that surpass what powerful general-purpose large language models, such as ChatGPT and its peers, can already deliver? Put differently, is Thomson Reuters building something more substantial and enduring than a simple interface or “wrapper” atop existing generative models—a layer of polish on someone else’s core invention—or is it genuinely creating proprietary intelligence that can stand the test of time?
The company’s leadership attempted to answer these concerns during its most recent earnings presentation on Tuesday. The outcome, however, elicited limited enthusiasm from Wall Street. Investor sentiment was muted, and the company’s shares declined by more than six percent following the announcement, exacerbating an already steep 30 percent drop since mid-July. Despite these setbacks, Thomson Reuters emphasized that its legal division continues to grow steadily, reporting a nine percent increase in organic revenue—to $700 million in the latest quarter—slightly outpacing the eight percent growth achieved in the first half of the year. According to management, features powered by AI within the CoCounsel suite have been key contributors to that momentum, delivering double-digit growth figures.
Chief Executive Officer Steve Hasker told analysts that AI-powered “agents” are now assuming an expanding portion of the tasks typically handled by attorneys. His argument centered on what he described as the company’s distinctive advantage: the fusion of an immense, authoritative law library with a layer of human expertise—its pool of editors, researchers, and legal analysts—ensuring accuracy, context, and reliability. This combination, he asserted, is “difficult, if not impossible, to replicate,” since it represents decades of editorial refinement coupled with hard-earned trust among legal professionals.
Nevertheless, Thomson Reuters now finds itself as much in a defensive stance as on the offensive front in this increasingly competitive industry. Historically a heavyweight in legal technology, the company is facing a transformed environment in which the rise of generative AI has dramatically lowered the barriers to entry. The innovation floodgates have opened, inviting ambitious startups—buoyed by abundant venture funding—to challenge long-standing incumbents.
In June, for instance, RELX’s LexisNexis division formed a strategic partnership with Harvey, a fast-growing AI startup focused exclusively on legal technology. The arrangement tightly integrates LexisNexis’s vast global library of legal data with Harvey’s technical prowess in generative AI, yielding a tool that produces outputs enriched with citations and context from LexisNexis’s corpus. Access to that content still requires a LexisNexis login, but the collaboration underscores how alliances between data-rich incumbents and agile AI startups are rapidly reshaping the sector.
A few months later, in late August, another major move came when Clio, an influential player in legal operations software, acquired vLex—a prized specialist in global legal data—for $1 billion. These and similar developments frame a competitive landscape in which Thomson Reuters must now constantly defend its market share and technological credibility. In response to the growing chorus of AI startups proclaiming sky-high valuations and dizzying growth rates, Hasker expressed skepticism, describing many of these claims as “squishy.” At the same time, he maintained that adoption of CoCounsel remains ahead of its rivals, signaling quiet confidence in the firm’s approach.
Central to that confidence is what Hasker called Thomson Reuters’ “moat.” This defensive barrier rests on the depth and precision of Westlaw’s legal data, combined with the company’s enduring commitment to human-in-the-loop editing and validation. In the highly sensitive world of litigation and compliance—where consequences for factual or interpretive errors can be severe—he contended that no amount of algorithmic sophistication can replace verified, well-contextualized information. “Litigation,” he observed, “is high-stakes work with no room for error and serious implications when one is wrong.”
Each year, Westlaw ingests an astonishing 300 million documents from hundreds of court systems, appellate bodies, and legislative databases around the world. Within the Thomson Reuters ecosystem, teams of qualified lawyers and editors meticulously transform this torrent of raw legal material into structured, court-ready intelligence. They write headnotes, verify facts, cross-reference precedents, and apply consistent taxonomies so that end users—attorneys, corporate counsel, judges, and students—can rapidly discover and rely on the information. Roughly 85 percent of all primary documents within Westlaw—such as court opinions and statutes—benefit from these editorial enhancements, turning chaotic data into refined, reliable content.
The company’s longer-term strategic objective now extends beyond traditional research and summarization. Thomson Reuters is developing more sophisticated generative tools and interactive workflows—known internally as “agentic” systems—capable of performing higher-value legal tasks, such as drafting analyses, preparing arguments, and mapping litigation strategy. These developments reflect a broader ambition to transform legal research platforms from passive repositories into dynamic, AI-augmented collaborators.
Hovering over the entire legal technology market, however, is one unpredictable factor: OpenAI. The question many in the industry ponder is whether OpenAI will ultimately decide to release a dedicated product for the legal sector. Such a move could upend the competitive balance overnight. Although no concrete indications exist at present, speculation intensified after OpenAI publicly demonstrated internal tools performing contract reviews—a core function of most legal AI providers. Should OpenAI pursue this direction, it could disrupt its own customer relationships; many current partners, including Thomson Reuters, rely on OpenAI’s large language models, such as GPT, to power their offerings. The tension is amplified by the fact that OpenAI is also an early investor in Harvey, one of Thomson Reuters’ most direct challengers.
Even without a specialized legal product, OpenAI remains the gravitational force around which much of the industry orbits. Lawyers across firms—whether through official integrations or informal use—regularly experiment with ChatGPT and other general AI assistants for research and writing. Consequently, any advancement OpenAI introduces to its general-purpose models can instantly reset the expectations for legal AI. Harvey’s own CEO has remarked that clients routinely compare his company’s capabilities against the newest releases of OpenAI’s models, effectively allowing OpenAI to define the market’s benchmark.
Hasker’s counterpoint is straightforward yet nuanced: depth and specialization ultimately prevail over generalization. He argues that tailored legal AI tools, grounded in proprietary, high-quality data and supported by experts who verify and annotate outputs, will outperform broad but shallow AI systems in areas that demand rigor, accountability, and precision. “Customers,” he noted, “are beginning to appreciate that difference.” Still, in a candid moment, he acknowledged the profound uncertainty that defines this fast-moving ecosystem, adding, “Anyone who claims to know exactly what will happen next is probably slightly deluded.”
For now, the race continues—a contest between technological ambition and professional responsibility, between vast general-purpose AI models and deeply specialized legal knowledge. In that space, Thomson Reuters seeks to prove that even as artificial intelligence ‘eats’ the world of software, there remains a crucial role for institutions that combine human judgment, trustworthy data, and carefully guided innovation.
Sourse: https://www.businessinsider.com/thomson-reuters-legal-tech-ai-chatgpt-test-2025-11
