Lex Machina

AI Centrality

How much of the product is actually AI. Whether the machine learning is the mechanism the buyer is paying for or a feature layered onto conventional software, and whether the vendor is specific about which is which.

The machine learning is the mechanism that manufactures the asset, on a product whose value to the buyer is the resulting database. Lex Machina states that proprietary technology and AI-assisted attorney review converts raw legal documents into comprehensive data sets and fills gaps in court records, and that proprietary AI analyses documents including reading signature blocks, updating case data and grouping related entities. Without those models the structured analytics would not exist in usable form. What holds this at B rather than A is that the thing sold is a queryable dataset rather than model output: a buyer pays for coverage and accuracy of the analytics, and the generative layer arrives as Protégé sitting over the top rather than as the product itself.

Citation Accuracy and Hallucination Disclosure

Whether the vendor publishes measured accuracy on citations and assertions, grounds output to primary sources, and says plainly what its system does when it does not know. Legal has a documented public record of fabricated citations reaching filed briefs, so an untested claim of accuracy is not evidence.

Accuracy is claimed in superlative terms and measured nowhere. The API section describes the offering as the industry's most accurate litigation data, and the product is positioned on filling gaps in court records that others miss, but no figure supports either claim. Searched the product page in full on 31 Aug 2026 and located no accuracy rate, no error rate for extraction or entity resolution, no benchmark, no test set and no published evaluation. Nothing addresses what Protégé does when the underlying data does not support an answer to a prompt. What is published instead is provenance and scale: coverage figures carrying an explicit as-of date of April 2025, which tells a reader how current the data is without telling them how right it is.

Autonomy and Oversight Model

What the system decides on its own, what a lawyer must approve, and whether the vendor documents where the review point sits. A tool that drafts under review and a tool that files without one are different products and different risks.

A real review point is published for the data pipeline and nothing is published for the generative layer. The product describes AI-assisted attorney review as the method by which raw documents become data, which places qualified lawyers in the loop between model output and what a customer sees, and the RELX Responsible AI Framework the vendor commits to includes human oversight as a named principle. That is a written commitment with a real review surface. What is absent, checked 31 Aug 2026, is the rest of the control structure: nothing states what the extraction models decide alone, what proportion of output an attorney reviews or against what threshold, what happens when an error is found in published analytics, or what oversight applies to Protégé when a user reaches the data through a prompt.

Operational and Outcome Evidence

Named, dated evidence that the product works in production at real firms or legal departments. Case studies with figures and identified customers count. Unattributed testimonials and launch announcements do not.

Real adoption evidence with a cited basis, and testimonials that thin out under inspection. The adoption claim is unusually well footnoted for this market: trusted by over 90 per cent of the largest firms, with the basis given as the 2025 Law360 Pulse Leaderboard Report and the 2025 AmLaw100 Report and the underlying data dated to April 2025. Against that, only one of the three testimonials is a named practitioner at a named firm, John Johnson, a partner at Fish & Richardson, and his quote carries no figure. The second is attributed only to a Chief IP Litigation Counsel at an unnamed Fortune 500 company. The third is from Miriam Rivera, described as former Deputy General Counsel at Google, and is framed hypothetically, saying what she would do if she were at Google today, which is a statement of opinion rather than evidence of a deployment. No case study with figures for what changed was located.

Privilege and Confidentiality Posture

How client confidences are handled: attorney client privilege and work product treatment, segregation of one client matter from another, whether client data trains any model, and what the vendor commits to in writing rather than in marketing.

A clear and unqualified training commitment, short of the detail around it. The product page states that LexisNexis never uses customer data to train AI models, which is stated flatly with no carve-out for internal models, external models or aggregates, and adds that identifiable information is removed from AI interactions so performance can be improved without compromising privacy. Robust data retention and deletion policies are asserted, privacy by design is claimed at every stage of systems, products and business processes, and advanced encryption is referenced. Three gaps checked 31 Aug 2026: no retention period is stated, no segregation or access model is described, and privilege and work product are not addressed. The commitments also sit on a product marketing page rather than in any agreement, since the LexisNexis terms were not opened. Note the exposure here differs from a document platform, because what a customer puts in is largely queries and research rather than client files.

UPL and Professional Responsibility Posture

Whether the vendor is clear that it supplies a tool rather than legal advice, who its audience is, and how it addresses unauthorized practice of law, competence and supervision duties, and jurisdiction limits. ABA Formal Opinion 512 is the reference point.

Nothing published on the advice line was located. Searched the product page in full on 31 Aug 2026, including the data security, privacy and governance section and the footer. No statement that the analytics do not constitute legal advice, no professional responsibility or ethics page, no bar or ethics guidance named including ABA Formal Opinion 512, and no jurisdiction limits. The product predicts outcomes and forecasts the chances of success of a motion, which is squarely the kind of output a lawyer must exercise independent judgement over, and the marketing invites exactly that use without addressing it. Exposure is lower than for a consumer-facing tool because the buyer is a lawyer, but the axis asks what the vendor has published and the answer is nothing. The LexisNexis general terms were not opened, so this is rebuttable.

AI Governance and Bias Disclosure

Published governance over model behaviour: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.

A named governance framework that addresses bias explicitly, short of a named owner and any published result. The vendor states that LexisNexis follows the RELX Responsible AI Framework and enumerates its commitments: that its AI protects privacy, is transparent and explainable, avoids unfair bias, includes human oversight, and is designed for real-world impact. Naming bias as a governed dimension puts this ahead of most of the pull, where bias goes unmentioned. What is missing is everything downstream of the principle: no individual or role inside the vendor is identified as accountable for model behaviour, no pre-release testing regime is described, no evaluation method is published, and no finding about uneven output across case types, courts, parties or populations has been disclosed. On a product that profiles named judges and named counsel, published bias testing would matter more here than almost anywhere else on this index.

AI Safety and Data Stewardship

Retention, deletion, access control, and what happens to prompts and documents after they are processed. Whether the vendor states its subprocessors and its incident practice, or leaves the buyer to assume.

Principles are published where controls should be. The product page offers privacy by design applied at every stage, protection of customer data wherever stored, sent or accessed, advanced encryption, enterprise-grade cloud providers, continuous monitoring, regular third-party audits and robust data retention and deletion policies. Every one of those is an assurance rather than a specification. Searched the page in full on 31 Aug 2026 and located no retention period, no named subprocessor, no incident or breach notification practice, no encryption standard and no access control model. The one concrete statement is that identifiable information is removed from AI interactions. A LexisNexis Trust Center is linked openly at trust.lexisnexis.com and is credited here as a genuine access route rather than scored against the vendor; it was not opened, so this grade is rebuttable on its contents.

AI Liability and Recourse

What the vendor stands behind contractually when its output is wrong. Indemnities, caps, carve outs, insurance, and whether any of it is published or only reachable through a negotiated agreement.

Written 31 Aug 2026 as an R7 amendment; left unwritten in the original build because the governing agreement had not been opened and section 8 puts this axis's evidence solely there. Surfaces read 31 Aug 2026: lexmachina.com/terms/, which resolves to the LexisNexis product page rather than to any Lex Machina agreement, and the LexisNexis General Terms and Conditions effective 5 June 2026. The scope connector is the GTC's own opening line, which states that the terms govern use of any LexisNexis product or service linking to them, together with the absence of any separate Lex Machina agreement in market. What is published is real and specific. Section 7.1 gives an indemnity against third-party claims of patent, trademark, service mark, copyright or trade secret infringement, with five named conditions and three remedies, the last being termination with a pro-rata refund, and 7.2 makes it the sole and exclusive remedy. Section 6.3 caps aggregate liability at the lesser of actual direct damages or fees paid in the preceding twelve months, which is more restrictive than the plain twelve-month cap common in this corpus, and states that those damages are in lieu of all other remedies. Section 6.4 excludes consequential damages with four named carve-outs. What keeps it off A is that the indemnity reaches infringement and nothing reaches output: section 6.1 disclaims liability for loss resulting in any way from errors or omissions in the Online Services or Materials, section 5.2 provides them as is with all other warranties disclaimed, and section 4.3(a) states that AI systems may not be accurate or error-free and that users are responsible for verifying. No insurance position is named. On a litigation analytics product sold for case assessment, the error-and-omission disclaimer is the clause a buyer should read first.

Practice Systems Integration Depth

How deeply the product reaches into the systems legal work already lives in: document management such as iManage and NetDocuments, Word and Outlook, contract lifecycle management, matter management, e-billing, and court filing systems.

Integrations are named and all of them are inside the parent's own portfolio. The page states that Lexis+ with Protégé, CourtLink, Law360 and CaseMap+ AI have integrations with Lex Machina and complement it, and Protégé itself is embedded in the product. An API is offered and described as a route for customers to build their own solutions using the underlying litigation data, linked to a community article. What was not located on 31 Aug 2026 is any integration into the systems litigation work otherwise lives in: no document management, no matter management, no billing, no filing system and nothing outside the LexisNexis family. Nor is there documentation an implementer could use, since neither the API article nor any developer index was opened, and nothing on the page describes what any integration moves or in which direction.

Deployment Model and Data Residency

Where the software runs and where the data sits. Multi tenant cloud, single tenant, private deployment, on premises, and whether region of residence is a published option or an enterprise conversation.

Nothing published on where the software runs or where customer data sits. Searched the product page in full on 31 Aug 2026, including the data security, privacy and governance section. The only statement touching infrastructure is that LexisNexis partners with trusted, enterprise-grade cloud providers, which names no provider, no region, no jurisdiction and no tenancy model. No multi-tenant or single-tenant statement, no residency option, no processing location, and nothing distinguishing where data is stored from where it is processed. The LexisNexis Trust Center is the reachable route to any of this and was not opened, so the grade is rebuttable on its contents; nothing on the product surface a buyer reads first answers the question.

Security Certifications and Trust Center

Independent attestation a buyer can pull without a sales call: SOC 2, ISO 27001, penetration test summaries, a trust center with current reports and named scope rather than a badge image.

A trust centre exists and the product page names no standard at all. The security section states that LexisNexis systems meet the highest standards for information security, that the company undergoes regular third-party audits to maintain industry-leading certifications, and links to a trust centre at trust.lexisnexis.com. No certification is identified by name anywhere on the page, so a reader cannot tell whether that means SOC 2, ISO 27001 or something else, and no auditor, coverage period, scope or report route appears. The trust centre is openly linked and is credited as a genuine access route under the retrieval rules; it was not opened on 31 Aug 2026 and is a SafeBase style portal, so whether it fulfils on registration or requires a sales conversation was not established. This sits at C because assurance language without a named standard is weaker than a badge with no scope.

Model Supply Chain Disclosure

Which models sit underneath, whose they are, where they run, and whether the vendor commits to telling customers when that changes. A legal buyer inherits every dependency it cannot see.

Nothing published about the model supply chain a customer inherits. Searched the product page in full on 31 Aug 2026. Protégé is named as the AI assistant delivering generative analytics and the extraction technology is described as proprietary, but no model provider is identified, no model or version is named, no architecture is described, no processing location is given for the model layer, no subprocessor list exists, and nothing commits the vendor to notifying customers when any of it changes. Referring to proprietary technology and a branded assistant without saying what sits underneath is the same shape recorded against Paxton AI in pull 1 and graded the same way here.

Commercial Transparency

Whether a buyer can learn what this costs without entering a sales process: published rates, the unit being charged, what sits behind an enterprise tier, and what implementation adds.

No pricing information is published at any level, including the unit of charge. Searched the product page in full on 31 Aug 2026, including its on-page navigation covering features, audience, resources, testimonials and related products. There is no pricing page, no tier structure, no per-seat or per-practice-area unit, no volume banding and no statement of what implementation adds. Every call to action is a demo request, and the demo form is the only route offered. Nothing published would let a prospective buyer form any view of cost before entering a sales process, which is notable on a product sold to segments as different in scale as AmLaw firms, solo practitioners, law schools and state government.

Firm and Practice Coverage

Who the product is actually built for. AmLaw, midlaw, small firm and solo, in house departments, government and courts, and which practice areas are supported rather than merely claimed.

The most complete coverage statement read in this pull, on both who it serves and what it covers. The audience is enumerated across eight segments rather than gestured at: large law firms, small law firms, corporate legal departments, insurers, courts and judges, state and local government, federal government, and law schools. That includes the government and court use that this axis asks about and that almost nothing else in the pull addresses. Coverage is stated with equal precision and, unusually, with its limits: all 94 federal district courts, the 13 courts of appeal, the PTAB and specialty venues, but state coverage described as enhanced state courts rather than complete, with a separate pool of 18 million state cases supporting party analytics only; appeals analytics bounded to federal civil cases filed since 2012; motion metrics bounded to more than 40 motion types; and the whole coverage claim carrying an explicit as-of date of April 2025. Practice areas are evidenced through published litigation reports covering patent, class action and administrative law rather than merely claimed.