European media: RELX and Wolters Kluwer respond to the backlash against artificial intelligence

RELX & Wolters Kluwer: Pushing back on the AI pushbacks Engaging in an AI debate comes with risks. So we found when our ever inventive Director ofResearch came to us with AI-generated pushbacks to our 27 May “Charting a 2027-29 pathto AI reality”. As it happens, they sensibly coincide with real world investors concerns, whichwe address in this more condensed report. We will be available over the next few weeks to Christophe Cherblanc+41 582 723 540christophe.cherblanc@bernsteinsg.com Annick Maas+44 20 7676 6683annick.maas@bernsteinsg.com “Your thesis overestimates incumbency and underestimates AI-drivencommoditization”We do not believe our assumptions are complacent. We estimate 50%and 90% of RELX & WKL portfolios are implicitly “challenged” at current share prices. Moresignificant, we conservatively assume a bear outcome probability at “clarity time” of 20% Christopher Pancur+44 20 7676 7280christopher.pancur@bernsteinsg.com Specialist Sales “Pricing power is fragile if LLM platforms or new interfaces capture value” +“Workflow stickiness doesn’t protect against a shift to outcome-based pricing,which could shrink revenue pools even as usage rises”1/ Anthropic recentannouncements include a growing list of commoditizable “skills” (build a DCF valuation,update portfolios etc…) but also a fast expanding list of MCP connectors to various dataproviders like Factset, Verisk Analytics, Moody’s, Thomson Reuters etc… Concerned “TAM assumptions (e.g. $5bn legal AI) rely on optimistic penetration and pricing thatcould compress”1/ $5bn is the TAM at 100% penetration, after carefully adjusting forlawyer density and gross national income levels ; 2/ available Harvey or LexisNexis clientpricing claim ROI of 10x to 20x, suggesting “fair” pricing points ; 3/ Incumbents initiallyfocused on content like RELX (and Thomson Reuters) have a “right to earn” in what is a “Disruption could hit sharply at renewal cycles post-2027”Better assessment of AIeffectivetotalcosts will be a key driver as full clarity on the subject has not been established yet. We expect sufficient visibility for users to make informed choices by end 2027. Witha sales cycle of typically of 2-3 years, we assume clarity by end 2029 for all segments, asthe post 2027 sales cycle will have fully played out. We acknowledge the speed at whichthis clarity will emerge is more of a moving target. We assume clarity will emerge faster for “WKL is most exposed, with real risk that high-margin assets get structurally dilutedby free or lower-cost AI-native competitors”1/ WKL risk exposure largely reflects thedifficulty of “grasping” a very diversified portfolio spread across fragmented small, medium BERNSTEIN TICKER TABLE INVESTMENT IMPLICATIONS We rate RELX Outperform with a 3,450p PT, Wolters Kluwer Outperform with a €102 PT DETAILS 27 May 2026 - RELX & Wolters Kluwer: Charting a 2027-29 path to AI clarity6 May 2026 - Quick Take: Wolters Kluwer 1Q revs tracking well vs reiterated FY outlook - UTD Expert AI adoption ramping up29 Apr 2026 - Quick Take: RELX French Legal Tech Doctrine acquisition underlines confidence in Legal AI prospects29 Jan 2026 - Quick Take: RELX - ChatGPT Prism... Should you be scared of a LaTeX-native editor?28 Apr 2026 - Quick Take: OpenEvidence no longer available in EU/UK - just a lobbying stance?23 Apr 2026 - Quick Take: RELX share price rally since FY results is not "too much too fast" “YOUR THESIS OVERESTIMATES INCUMBENCY AND UNDERESTIMATES AI-DRIVEN OVER ESTIMATING INCUMBENCY? We do not believe the assumptions taken to model the “path to clarity” detailed in our 27 May report are complacent. Weestimate at current share prices 50% and 90% of RELX & WKL portfolios are implicitly “challenged”. More significant, whenweighing the actual outcome at “clarity time”, ie whether various lines of businesses emerge unaffected, or benefit from TAMexpansions, or are damaged by loss of relevance or weaker pricing power,we conservatively assume a bear outcomeprobability of 20% for RELX and 40% for WKL,the gap reflecting more visible TAM expansions for the former. This reflects •Segments with the highest confidence/lowest risk profile remain in our view those embedded in proprietary content ieessentially within RELX Risk Insurance solutions and US Legal content. In the latter case, this reflects our analysis of nativeAI start-ups effective focus and hiring patterns as developed in this report. In the latter case, we reiterate Risk relianceon extractive AI (think large datasets crunching) applied to proprietary data (blending public, licensed and all important •Segments whereon we have improving confidence are enjoying incremental new TAM (Legal RELX, Legal & RegulatoryWKL). This also applies over a more extended timeline to STM Tools (RELX), or with no increase in value extraction (RiskBusiness services). While visibility is still low, we are directionally more confident on Clinical Solutions (WKL) after 1Q26release indication that UTD Expe