The Moat Built on Authority
RELX powers work most people never see: lawyers pulling case law, banks screening customers, insurers pricing risk, scientists sifting papers. For years, that mix of recurring subscriptions and workflow dependence made the company look like a steady compounding machine.
Generative AI has added a destabilizing thought experiment. If a model can draft a research memo, summarize literature, or turn a stack of contracts into an issues list in minutes, how much of the value sits in RELX’s databases—and how much sits in the interface that delivers answers?
RELX’s answer is that it can turn AI into an upgrade, not a threat. In its October 2025 trading update, it highlighted strong underlying growth and pointed to rapid usage growth for Protégé, its AI assistant inside Lexis+ AI, as proof that customers will pay for trusted, integrated tools.
The risk is that AI changes habits faster than pricing models can adapt. Search is becoming conversational, and increasingly embedded in other software. If the first stop for a junior associate or a researcher is a chatbot living inside a document system or collaboration suite, the owner of that surface can demand a share of the economics.
The Chatbot at the Gate
RELX’s legal franchise, LexisNexis, is both its prize asset and its most exposed battlefield. Lawyers pay for precision and defensible citation, which should favor incumbents with curated content and citation infrastructure. LexisNexis has leaned into that advantage by expanding Lexis+ AI and rolling out newer versions of Protégé, including a “General AI” workflow that combines LexisNexis content with a customer’s documents and selected web sources.
Yet the same technology that strengthens incumbents also lowers the barrier for challengers. The competitive set now includes not only Thomson Reuters but a fast-growing cast of AI-native tools trying to wrap research, drafting and review into one guided workflow. Thomson Reuters has been integrating its AI assistant, CoCounsel, more tightly with Westlaw and Practical Law, and has pushed toward meeting lawyers inside Microsoft 365 and Teams. If that distribution shift sticks, a research platform risks becoming the plumbing—essential, but priced like a utility.
RELX has tried to preempt that outcome by partnering as well as competing. In mid-2025, LexisNexis struck a strategic alliance with Harvey to integrate LexisNexis content and jointly develop workflows. The move widens distribution, but also concedes a key point: some of the most valuable user experiences may be built outside RELX’s own products.
For RELX, the monetization challenge is that AI features quickly become table stakes. If “AI-assisted drafting” is bundled into every major legal platform, differentiation shifts to trust, coverage, and workflow integration—an arms race measured in engineering execution and data rights as much as model quality.
The “Claude Shock” and the Angst Trade
The anxiety spiked in early February when Anthropic rolled out a legal-focused plug-in for its Claude “Cowork” agent—positioning it as a way to speed up contract review, NDA triage and compliance workflows. In markets that have spent the last year treating incumbents like RELX as the picks-and-shovels winners of the AI era, the announcement landed less like a product update and more like a warning shot: what if the shovel-maker gets disintermediated by the machine?
The reaction was visceral. Shares in the legal-information establishment sold off sharply as investors repriced the possibility that the interface—the place where legal work begins—might migrate away from the publishers’ platforms and into general-purpose AI tools with domain “skins.” Some coverage framed it as a broader “panic mode” moment for professional-information names, while legal-tech outlets dubbed it a “Claude Crash,” arguing the market was extrapolating too far from an early product release. Either way, the episode captured the core strategic risk for RELX: not that its content becomes worthless overnight, but that AI rewires user behavior and bargaining power before incumbents can fully repackage their authority into a premium, paid AI layer.
A New Kind of Competition
Elsevier, RELX’s scientific publishing arm, faces a different flavor of risk. AI can make the corpus more useful by helping researchers navigate it, which is one reason Elsevier launched ScienceDirect AI. But AI also accelerates the unbundling of discovery from the publisher’s platform. If researchers can get a competent synthesis elsewhere, library willingness to pay premium subscription prices becomes easier to challenge.
Those economics were already under strain from open-access pressure and budget politics. In early 2026, several U.K. universities that had cancelled Elsevier deals argued researchers could still obtain papers quickly via other channels such as interlibrary loans—a small but telling sign that workarounds are improving. AI could make those workarounds feel seamless by summarizing whatever arrives, wherever it arrives from.
Then there is the fight over data itself. As professional text becomes fuel for AI systems, licensing terms are turning into battlegrounds. A recent legal-tech dispute alleging that a database license was breached to train AI models highlights how valuable—and contested—these corpora have become. RELX’s trove is an asset, but it also puts the company in a squeeze: deny access too aggressively and risk being routed around; grant access too freely and risk commoditizing the very thing it sells.
None of this guarantees RELX “goes ex-growth.” In regulated, high-stakes domains, customers still care about provenance, auditability and governance—areas where general-purpose AI still struggles. But the competitive question is shifting from “who has the content” to “who owns the workflow and distribution surface.” RELX’s job is to make sure it remains both.
