Lexis+ with Protégé vs VitalLaw Expert AI: how they compare in 2026

Lexis+ with Protégé is a generative layer over a research platform that is decades older and works without it. Without the models, Lexis+ remains a working research service, with Shepard's, case law, statutes and Practical Guidance intact; what goes is conversational search, drafting and summarization. LexisNexis says the AI was built internally with technology partners rather than acquired, and describes the retrieval platform as proprietary.

The models are the engine of the capability sold as VitalLaw Expert AI, layered on a research platform whose value stands without them. What the subscriber pays for here is model work: conversational answers, summaries that surface key points and risks, and drafted checklists, memos and policy language, all generated by retrieval over VitalLaw content. The product page describes it as integrated into the VitalLaw platform to enhance an existing workflow, and that platform, with its treatises, reporters, editorial analysis and practice tools such as SmartCharts, was sold for years before the AI arrived in September 2024. Remove the models and a subscriber still has the research platform.

The product was marketed on legal citations free of hallucination, and quoted elsewhere as 100 percent free of hallucination; LexisNexis has not withdrawn the claim. It was tested in Magesh, Surani, Dahl, Suzgun, Manning and Ho, Hallucination-Free? Assessing the Reliability of Leading AI Legal Research Tools, 22 J. Empirical Legal Stud. 216 (2025), from Stanford RegLab and Stanford HAI, the first preregistered evaluation of commercial legal AI tools. It used 202 handbuilt queries across general research, jurisdiction and time specific, false premise and factual recall categories, scored by experts, with the dataset published. The product hallucinated on more than 17 percent of queries, was accurate on 65 percent, and gave incomplete answers on 18 percent, the best result among the systems tested. Grounding runs through a proprietary retrieval platform, sources reviewed by Shepard's editors, linked citations, a citation check tool and feedback in the product. LexisNexis disputed the figure, saying its internal data showed lower rates, and has not published that data. The study's first method drew objections from both vendors tested and was revised, access limits constrained the design, and the systems were tested in May 2024 and have changed since.

Grounding is real and documented, with linked sources and a described retrieval method, and no accuracy measure is published. The vendor says the product uses retrieval augmented generation that pulls only from repositories inside VitalLaw rather than public data. Footnotes and PinPoint Citation link every answer to its original sources, and Thinking Mode shows how a query was broken down. Wolters Kluwer's editors write vetted question and answer pairs for complex topics, and the launch release says outputs and user feedback are reviewed. No accuracy figure or evaluation is published, no failure mode is named, and nothing states what the system does when the content does not answer the question.

The RELX Responsible AI Principles name human oversight, and LexisNexis says human experts review model output for legal accuracy and that review and monitoring run across products. Citations are linked so a user can open the authority behind any point, a user can submit a citation to be checked and told when it may be wrong, and feedback in the product is said to improve accuracy. No autonomy threshold is set, nothing states what the system does without review, nothing describes what expert review covers or how often, and no escalation behavior is described. The independent study found incomplete answers on 18 percent of queries, so the system declines or under answers at a measurable rate, and LexisNexis documents no refusal behavior.

A stated commitment to human oversight comes with real review surfaces and no control structure. The vendor's April 2026 article lists human oversight among the responsible AI principles the product follows, and the product page presents Thinking Mode, footnotes and pinpoint citations as tools to verify facts and understand context before relying on an answer. Chat history lets a user return to prior work. The product answers and drafts on request and acts on nothing by itself. No mode, threshold or categorical limit on use of an output is published, and nothing addresses what happens when an answer is wrong.

The preview from May 2023 is described as covering global law firms, corporate legal departments, small law firms and US courts, and LexisNexis says preview feedback shaped specific changes before general availability in October 2023. The Stanford RegLab and HAI study found 65 percent accuracy across 202 expert scored queries, an outcome measured by a third party rather than claimed by the vendor. No customer, firm or court is named, no usage figure is published, and LexisNexis publishes no outcome measure of its own. The preview group is described only by segment.

No production evidence is published for the AI product. No customer is named, no case study or testimonial is published, and no figure for time saved or quality is attributed to any user. The materials describe the product as used within a globally trusted platform and claim productivity gains without a basis; the linked Future Ready Lawyer survey describes the market, not this product. The product and law firm pages, the April 2026 article and the launch release were searched.

The LexisNexis General Terms, effective 5 June 2026, warrant that prompts and files submitted to the AI services will not be used to train any large language model and will be treated at least as confidential information. The confidentiality clause requires advance notice before any disclosure compelled by law, so the customer can seek a protective order. The annual SOC 2 Type 2 examination of Lexis and Lexis+ covers confidentiality and privacy as well as security. Legal professional privilege and attorney work product are not addressed by name, and no separation by matter is described. Document upload is a main feature, so client material enters the system by design.

Confidentiality is asserted in general terms with one specific statement on training, and no agreement can be read before buying. The April 2026 article says every response is generated within the secure walls of the VitalLaw platform and every interaction remains within it, and the launch release says answers come without compromising an organization's data. The training statement covers public models only and sits in a marketing article. No subscription agreement, data processing terms or product privacy notice for VitalLaw is published, and nothing addresses privilege or work product, retention and deletion of chats, segregation between users or organizations, or what any model provider may retain.

Lexis+ with Protégé is governed by the LexisNexis General Terms, effective 5 June 2026, which say the online services are not legal, accounting, financial or other professional advice or a substitute for an attorney. The AI services clause adds that AI is not a substitute for independent professional advice and that users must verify what it provides. Nothing addresses a lawyer's own professional duties when relying on the output, and no jurisdiction or ethics guidance is named.

Nothing published addresses the advice line for a product that produces legal answers and drafts. No statement that the output is not legal advice, no position on a lawyer's competence or supervision duties and no jurisdiction limits are published. The product is marketed to legal professionals and says it lets junior staff contribute meaningfully, which is where a supervision statement would matter. The VitalLaw subscription terms, where a disclaimer would usually sit, are not published.

The RELX Responsible AI Principles are named: transparency, fairness, human oversight and respect for privacy. LexisNexis says on a trust page that the product is built with them, that human experts review AI models for legal accuracy and ethical performance, and that continuous review and monitoring keep products accountable. No model card, bias or fairness testing method or result, evaluation output, accuracy monitoring figure, named governance body or review schedule is published, and there is no ISO 42001. The only published accuracy measurement comes from outside researchers.

Responsible AI principles at group level are published and tied to the product, without a mechanism anyone could audit. The product states it adheres to Wolters Kluwer's AI Principles, which cover privacy and security, transparency and explainability, governance and accountability, fairness and a human focus, and commit to address risk management and issue remediation during design and development and after deployment. Nothing names who owns model behavior for VitalLaw, describes what is tested before release, or reports anything found about uneven output across practice areas or jurisdictions.

LexisNexis publishes encryption, data minimization, pseudonymization, privacy built in at every stage of product development, secure cloud infrastructure, a set of information security policies, and incident response plans. Those plans are updated and tested periodically, with technical, administrative, business and executive escalation paths and outside firms on retainer. Retention follows where a document goes. Uploads of up to 10 documents outside a Vault are purged at the end of each session. A user can keep up to 50 encrypted Vaults of as many as 500 documents each until deleting them, and results from Vault work stay in the My Conversations history for 90 days. LexisNexis removes identifiable information from AI interactions and says data one user shares does not shape performance for other users. No retention period is stated for other prompts and research questions, no tenant separation is described, and no subprocessor list is available outside the trust center's registration step.

General security statements cover the product without saying what happens to prompts and documents. The vendor says the product operates within strict logical and physical security frameworks and that privacy and security are part of the design, development and deployment of its AI. Conversations are kept for the chat history feature, but no retention period, deletion right, access control description, subprocessor list or incident commitment is published for VitalLaw. No product privacy notice or data processing terms are published.

The LexisNexis General Terms and Conditions, effective 5 June 2026, govern its online services. LexisNexis defends and indemnifies the subscriber against third party patent, copyright, trademark and trade secret claims arising from use of the service, and that indemnity sits outside a twelve month liability cap. The AI Services clause says AI systems may not be accurate or error free, that users must verify information, and that AI is not a substitute for independent professional advice. No term covers a wrong answer or citation, though the product was earlier marketed on results free of hallucination.

Nothing published says who bears the loss when the system is wrong. No agreement, warranty, indemnity or limitation of liability for VitalLaw or its AI is published on the product page, the April 2026 article, the launch release or the AI Principles, or in a search for VitalLaw subscription or service terms. The corporate website terms govern wolterskluwer.com and are not connected to the product.

The Lexis+ with Protégé product page says the product connects to iManage, SharePoint, NetDocuments and other document management systems so users can ask and draft from their organization's documents, alongside secure Vaults for uploaded files and shared Workrooms. Nothing describes what each connection syncs, in which direction, or what a firm must configure, and no API or developer documentation is published.

No integration into practice systems is published. The AI is built into the VitalLaw research platform, and VitalLaw offers a clause builder from a partner as an extra, but no connection to document management, Microsoft Word or Outlook, matter management or any other system a firm already uses is named. Work produced must be moved out by hand.

LexisNexis material refers to secure cloud infrastructure on enterprise platforms, naming no provider, region or residency commitment, and to strong encryption. No single tenant or dedicated option is described, and nothing says where customer queries or uploaded documents are processed and stored. The product is sold in the US, UK, Canada, Ireland and Australia under different data protection laws. The Canadian trust material raises PIPEDA and Quebec Law 25 as customer duties without stating LexisNexis's own residency position. The SafeBase trust center gives registered users further material.

Delivery is inside the vendor's platform, and neither tenancy nor region is stated. The vendor says responses are generated inside the VitalLaw platform and that the product can be deployed at once across departments because it is part of the existing platform. Nothing states whether customers share infrastructure, where data is stored or processed, or whether a region can be chosen.

The trust center at trust.lexisnexis.com runs on SafeBase, with self serve registration through which authorized users reach compliance reports, a request flow that does not run through sales. Its May 2026 update announced SOC 2 Type 2, SOC 3 and HIPAA/HITECH reports for Lexis, Lexis+ with Protégé and CaseMap+, the SOC 2 Type 2 focused on security, availability and confidentiality and offered for secure download, and the SOC 3 open to read. An earlier LexisNexis description puts the annual SOC 2 Type 2 examination of Lexis and Lexis+ against all five Trust Services Principles. A SOC 1 Type 2 report for digital content management and IT services was updated in January 2026. The compliance list adds ISO/IEC 27001, ISO 22301, IRAP, NIST CSF, HIPAA and UK Cyber Essentials, certified on 26 October 2026 by Securious for LexisNexis UK's office and production AWS networks. The auditing firm for the SOC reports is not named. A March 2026 notice on the trust center reports that an unauthorized party reached a limited number of servers holding mostly legacy data from before 2020, such as customer names, user IDs, business contacts, survey responses and support tickets, and no client or matter information; LexisNexis engaged a forensic firm, reported the matter to law enforcement and informed affected customers.

No independent security attestation is published for VitalLaw. Other Wolters Kluwer products, such as TeamMate, CCH Integrator and CCH Tagetik, publish certifications on their own pages, but no statement extends any of them to VitalLaw. The product page refers only to security frameworks without naming a standard, and a targeted search for a VitalLaw SOC 2 or ISO 27001 statement found nothing.

Protégé runs in two configurations. Protégé Legal AI blends several large language models from leading providers, which are not named, and grounds every response in LexisNexis sources with Shepard's citations. Protégé General AI gives access to models from OpenAI, Google and Anthropic, each with its own reasoning strengths and writing style, and LexisNexis says new configurations are assessed and added as they come. The retrieval platform is proprietary. Which models make up the legal configuration, which versions are offered, where the models run and which of them process customer uploads are not stated, and no notice of model changes is offered. No subprocessor list is published outside the trust center's registration step.

The retrieval design is described, and the models underneath are not identified. The vendor contrasts its product with generic AI platforms, saying it retrieves from its own platform content, not open web data or opaque training sets. No model, model family, provider or hosting arrangement for inference is named on any surface read, nothing says what a model provider may retain, and no commitment to notify customers of a model change is published.

No price, range, tier or unit of charge is published for Lexis+ with Protégé, and nothing says whether it is licensed separately or bundled into a Lexis+ subscription. LexisNexis prices each organization by its size, the capabilities it needs and the content it licenses, and quotes on request through a form or by phone. A 2 day free trial is offered. Third party listings report about $128 to $494 per user per month, and LexisNexis has not confirmed them.

No pricing information is published at any level, and the unit of charge is not stated. Every route on the product page, the law firm page, the April 2026 article and the launch release leads to a demo request. Trade press at launch reported the AI as an optional upgrade for existing VitalLaw subscribers; the vendor's current material does not say so.

Grounding content is named as case law reviewed by Shepard's editors, statutes and Practical Guidance, described as the largest repository of legal content. The product is available in the US and in other LexisNexis markets, including the UK, Canada, Ireland and Australia. No jurisdiction or court list, date range, update lag or refresh frequency is listed, and nothing says which countries the AI features are available in, as distinct from where the platform is sold.

The product page addresses law firms and corporate legal departments separately, and the vendor describes coverage of more than 26 practice areas grounded in its treatises, reporters and editorial analysis, naming tax, securities and privacy among them, with support for comparing regulations across jurisdictions. The full list of practice areas is not given, no jurisdiction list is published, and nothing says which practice areas or matter types the AI handles less well.