CoCounsel Legal vs Lexis+ AI: how they compare in 2026
CoCounsel Legal and Lexis+ AI are the generative layers of the two largest legal research platforms, so for most firms the choice follows the research subscription already in place. CoCounsel Legal, from Thomson Reuters, grounds its work in Westlaw and Practical Law. Its August 2026 release, rebuilt on Anthropic's Claude Agent SDK, shows its plan before acting, then researches, reasons and drafts with KeyCite flags beside each authority. It connects to iManage, NetDocuments, SharePoint and Word, and reaches other AI clients through MCP. Lexis+ AI, from LexisNexis, offers conversational search, drafting, summarization and document upload over case law, statutes and Practical Guidance. It checks every citation in an answer against Shepard's. The terms behind them differ in kind. Lexis+ AI's no training commitment is a warranty in its General Terms, and those terms add an intellectual property indemnity outside the liability cap. CoCounsel Legal's no training commitment sits in a product FAQ, and the product carries no indemnity or liability position of its own.
At a glance
All 15 axes, side by side
The same grid applied to every vendor in the index, graded from public sources. Hover a grade to see what the letter means on that axis.
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 artificial intelligence is the product. CoCounsel Legal exists only as a generative and agentic layer, rebuilt in 2026 on an agent framework so it plans, reasons and executes multi step legal work. Remove the models and there is no product, only the underlying Westlaw and Practical Law content, which the parent sells separately as its own products.
Lexis+ AI is a generative layer over a research platform that is decades older and works without it. Remove the models and Lexis+ remains a working research service, with Shepard's, case law, statutes and Practical Guidance intact. What is lost is conversational search, drafting and summarization. LexisNexis says the AI was built internally with technology partners rather than acquired, and the retrieval platform is described as proprietary, so the model work is real and owned. A mature platform hosts the AI layer rather than depending on it.
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.
Grounding is described rather than asserted. The research agent identifies key legal concepts, selects Westlaw tools including KeyCite, the Key Number System and annotated statutes, runs multiple search cycles and synthesizes a structured result. Output grounds to primary authority the user can open, with inline citations, a Sources tab with direct links and highlighted excerpts, and KeyCite validity flags shown with the authority. A citation level verification feature, Deep Research Verify, is published, and prompting guardrails restrict output to content validated by people rather than the open internet. Measurement is independent. CoCounsel took part in the February 2025 Vals Legal AI Report, a third party benchmark against a measured lawyer baseline, scoring between 73.2 and 89.6 percent across four tasks with a 79.5 percent average, each task more than 10 points above the lawyer baseline. The vendor also publishes CoCounsel Bench, under which attorneys reviewed 50 complex tasks of about six lawyer hours each and judged the product's response stronger than the expert reference answer on nearly 40 percent. The 2025 study measured task accuracy rather than citation validity or hallucination rate, and the vendor did not take part in the later Vals study on citation authoritativeness.
Lexis+ AI is well grounded, but LexisNexis has never withdrawn an absolute claim that independent testing contradicted. The product was marketed on legal citations free of hallucination, and quoted elsewhere as 100 percent free of hallucination. 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. That was 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. Lexis+ AI hallucinated on more than 17 percent of queries, was accurate on 65 percent, and gave incomplete answers on 18 percent. It was still the best performing system tested. Its grounding is real: 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, but has not published that data. Caveats: the 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.
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 published commitment with documented review surfaces, without the full control structure. The vendor states that its AI is built with human oversight at its core and publishes a Fiduciary-Grade AI framework built around transparency and accountability. The product is described as showing its plan before executing: it interprets a request, builds a plan specific to the matter, and returns work product with traceable citations the user can open and check. Not located: the threshold at which an agent stops and returns to a lawyer, and what the vendor commits to when an agent is wrong. The August 2026 shift from structured skills to a fully agentic experience makes the missing threshold more material than it would have been a year ago.
LexisNexis states oversight as a principle and builds it into the product, but sets no limits. 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. In the product, oversight is practical. 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. Nothing is bounded or measured. There is no autonomy threshold, no statement of what the system does without review, no description of what expert review covers or how often, and no escalation behavior. The independent study found incomplete answers on 18 percent of queries, so the system does decline or under answer at a measurable rate, and LexisNexis documents no refusal behavior at all.
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.
Named customers, dated, with figures and a method that can be assessed. The vendor publishes case studies naming Justly Prudent, reporting five times return on investment and a doubling of litigation capacity, and Brinks, on reducing outside counsel spend. It also publishes a commissioned Forrester Total Economic Impact study dated 2026, whose method can be assessed and whose commissioned status is disclosed on the page. Aggregate figures are published with the population stated, including 2,700 corporate customers using the CoCounsel family and a 61 percent reduction in time to draft correspondence.
Lexis+ AI has deployment evidence by segment and an independent measure of its performance. The preview from May 2023 is described as covering global law firms, corporate legal departments, small law firms and US courts. 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. Missing: no customer, firm or court is named anywhere in the material found, no usage figure is published, and LexisNexis publishes no outcome measure of its own. The preview group is described only by segment, which a buyer cannot verify or contact.
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.
Substantive published commitments, with the answer given directly in a published FAQ rather than implied. The vendor states plainly that customer content and prompts are not used to train or improve the product or the underlying models. Third party model partners, including OpenAI and Google, are contractually prohibited from training on customer data, and systemic controls turn off those parties' abuse monitoring so there is no human access to customer data. The vendor also publishes a zero retention API architecture and data residency controls. Attorney client privilege and work product handling are not addressed directly in located material. Segregation by matter is not documented on the vendor's own pages, although the product is built around matters.
LexisNexis makes contract commitments on the points that matter most. 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. Missing: any treatment of legal professional privilege or attorney work product by name, and any separation by matter. That matters because document upload is a main feature, so client material enters the system by design.
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. Where the advice line is not the duty a product raises, the axis is read through the nearest professional duty it does raise: judicial conduct rules and the reviewing duty for products sold only to courts, and the duty to bill for time actually spent for products that draft time entries.
The claim is made and the documentation is not published. The product is sold to licensed professionals and distributed to more than 200 law schools and 120,000 law students. The vendor's framing is that the system works at the level of a senior associate under professional supervision. The product page, the AI hub, the trust center and the legal blog were searched. No published position on the advice line, no treatment of competence and supervision duties, and no statement of jurisdiction limits is published. The senior associate framing is a capability claim rather than a professional responsibility position, and for a product marketed to students it is where a position would matter most.
A product wide disclaimer carries a duty to verify. The LexisNexis General Terms, effective 5 June 2026, which govern Lexis+ AI, 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. The independent study that measured a hallucination rate for this product makes that duty concrete. Nothing addresses a lawyer's own professional duties when relying on the output, and no jurisdiction or ethics guidance is named.
AI Governance and Bias Disclosure
Published governance over model behavior: 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 published governance framework with real substance, short of testing results or a named owner. The vendor publishes a named standard it calls Fiduciary-Grade AI with a buying framework attached, describes a dedicated research organization of more than 1,000 specialists, states a figure for annual responsible AI investment, and answers a direct published question on bias: prompting guardrails restrict output to a clearly defined and limited set of human validated content rather than the open internet. That is a stated mechanism, not a principles page. Not located: a named internal owner of model governance, published pre release testing results, or any disclosure of uneven output across matter types, parties or populations. The bias answer describes a content restriction rather than testing for disparate output.
A named governance framework with stated commitments is tied to the product, with nothing behind it. The RELX Responsible AI Principles are named: transparency, fairness, human oversight and respect for privacy. LexisNexis says 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. This sits on a trust page, not in a press release. Nothing behind the principles is published. There is no model card, no bias or fairness testing method or result, no evaluation output, no accuracy monitoring figures, no named governance body or review schedule, and no ISO 42001. A fairness principle without measurement is a commitment, not evidence. The one substantial accuracy measurement on record came from outside researchers, not from this framework.
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.
Substantive published policy covering most of the ground. SOC 2 Type II and ISO 27001 maintained, ISO 42001 certification referenced for the AI management system, a zero retention API architecture published, data residency controls, 24 hour security monitoring and threat detection, corporate identity management integration, and a corporate trust center carrying the information security framework. Not located: a current named subprocessor list for this product, and a stated incident and breach notification practice. The two named model partners are disclosed in the AI FAQ rather than in a maintained subprocessor list, which is the remaining element of the A bar.
LexisNexis states its handling controls, but not retention or separation. It 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. Not found: any retention position for prompts, uploads or research queries, and any description of tenant separation. For a product whose main features include document upload, those are the disclosures a buyer needs most.
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.
No published indemnity scope, liability cap, exclusion, warranty on output or insurance position is published for this product. Commercial terms are reached through a negotiated subscription agreement. The Fiduciary-Grade AI framework treats accountability as a design principle rather than saying who bears the loss when output is wrong. A fiduciary framing sits here without a published liability position.
A published agreement carries a real indemnity, and AI output is left for the user to check. 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. 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. So nothing stands behind a wrong answer or citation, despite earlier marketing of results free of hallucination.
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.
Real and documented integrations, short of implementer level depth. Named connections span document management with iManage, NetDocuments and SharePoint, Microsoft 365, the parent's own HighQ collaboration platform and Legal Tracker, contract lifecycle through Icertis, plus DeepJudge, Supio and Smokeball, and a Word add in carrying a personal clause library. An MCP integration lets a lawyer reach the product from a separate AI client without switching platforms, which is a genuinely forward integration surface. Not located: documentation describing what each integration moves, in which direction, and what an administrator must configure.
No integration was found in the material read, and that material did not cover integration documentation. No integration, connector, API or document management system link was found across the product pages, launch material, trust center material and UK trust explainer checked. That review covered the product's capability claims and its trust pages, not its integration documentation, and a provider of this size very likely publishes some. This is an absence on the pages checked on that date, not an established gap in what LexisNexis discloses.
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.
Deployment is stated clearly with partial residency detail. The vendor publishes data residency controls that keep information within required geographic boundaries, operates a rollout sequenced by country across the United States, Canada, the United Kingdom and Australia, and holds FedRAMP In Process status for the underlying Westlaw and Practical Law products with stated investment toward authorization for this one. Not located: the list of available regions, the tenancy model, what changes between tiers, and where processing happens as distinct from where data is stored.
Where data is held gets no specific answer from LexisNexis. Its material refers to secure cloud infrastructure on enterprise platforms, naming no provider, region or residency commitment, and to strong encryption, which is not a deployment disclosure. No single tenant or dedicated option is described, and nothing says where customer queries or uploaded documents are processed and stored. Residency matters for a product 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 may hold residency detail behind its registration step, which was not completed.
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.
Certification is real and stated, without accessible evidence. SOC 2 Type II and ISO 27001 are stated as maintained, and ISO 42001 is referenced for the AI management system. FedRAMP In Process status is published for Westlaw and Practical Law, with authorization stated as an investment goal for this product, and a corporate trust center publishes the information security management framework. No attestation report, coverage period, audit scope or named auditor is published, and no route to request the reports without a sales conversation is published either. Certifications are stated at the parent company level rather than scoped to this product, which is the distinction a buyer's security team asks about.
LexisNexis runs a live trust center and names the full scope of its SOC 2 examination. The trust center at trust.lexisnexis.com runs on SafeBase, with self serve registration through which authorized users reach compliance reports, a request route rather than a sales gate. An independent auditor performs an annual SOC 2 Type 2 examination of Lexis and Lexis+ against all five Trust Services Principles: security, availability, processing integrity, confidentiality and privacy. A SOC 1 Type 2 report for digital content management and IT services is also published, and its update announced. The annual schedule addresses currency. Two limits. The auditing firm is not named. And one LexisNexis page describes the Lexis+ AI specific examination as scheduled for Q1 2024, with no later confirmation found, so the AI product's current attestation is unclear even though the platform's is not. The trust center also carries a customer notice about a security matter. Its text was not read in full, so nothing is said here about its scope, date or which LexisNexis entity it concerns.
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.
The supply chain is partly disclosed, and specifically. The vendor names Anthropic's Claude Agent SDK as the architecture of the current release, names OpenAI and Google as third party AI partners in a published FAQ, and states the contractual position imposed on them. Not located: which model serves which task, where the models run, a maintained subprocessor list for this product, and any commitment to notify customers before the model supply chain changes. That last gap matters here because the product was rebuilt on a different vendor's agent framework within the past year.
LexisNexis confirms it uses outside model partners but names none. It says the AI was built internally with world leading technology partners, which confirms third parties are involved while identifying none, and describes the retrieval platform as proprietary. No foundation model provider, model family or version is named, and no subprocessor list was found. Nothing says which models process customer uploads as opposed to the public content. Regulated buyers whose own supervisors ask about subprocessors will notice the gap.
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.
The shape is visible and the number is not. Tier names and feature splits are published and reachable from the main navigation: CoCounsel Essentials, Westlaw Advantage with CoCounsel Essentials, Practical Law Dynamic Tool Set with CoCounsel Essentials, and CoCounsel Legal. Each has its own plans and pricing page, with variants for law firms, corporations and government. Third party analyses report that the pages run a configurator returning a figure for firms of up to ten attorneys and route larger firms to a sales quote. The configurator could not be checked directly because that subdomain blocks automated access, so no vendor published figure is recorded here; it may say more. The unit of charge is per user per month by third party report rather than vendor statement.
Pricing for Lexis+ AI is unpublished at every level. There is no price, range, tier or unit of charge. Nothing says whether Lexis+ AI is licensed separately or bundled into a Lexis+ subscription, the first question a buyer with an existing contract would ask. Every route is a contact or demo request. Independent commentary describes tools in this segment as costly per month without giving a figure, which is description, not disclosure. Checked the product pages, the launch and press material and the trust center material. This is common for large legal research providers, and still a gap a buyer cannot work around.
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.
Who the product serves is documented precisely across many segments. Published navigation separates law firms under ten attorneys from firms over ten, and separately addresses corporations, government (including courts, prosecutors, public defenders, city attorneys and law enforcement), and academic use across more than 200 law schools. Practice areas are listed with their own pages, spanning business and commercial, civil litigation, criminal defense, family, labor and employment, mergers and acquisitions and personal injury. The limits are stated structurally rather than in prose: the published table of plans shows exactly which content each plan reaches, so a buyer can see where coverage stops.
LexisNexis describes its content by type, but does not list its coverage. 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. Shepard's itself implies broad coverage, since a citator only works where the case law is complete. Nothing is listed at the level a researcher checks: no jurisdiction or court list, no date range, and no update lag or refresh frequency. Nothing says which countries the AI features are actually available in, as opposed to where the platform is sold. Largest repository is a comparative claim, not a measurable one.
The 12 legal signals, side by side
Recorded rather than graded. These are the questions a practitioner has to answer before a tool touches a client matter, and the answers are taken from public material only.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
A published FAQ answers the question directly and in the negative, stating that user content and prompts are not used to train or improve the product or the underlying models, and that third party AI partners including OpenAI and Google are contractually prohibited from using customer data to train their models. The vendor also states it has established controls to turn off those third parties' abuse monitoring so there is no human access to customer data.
The commitment as located sits on a public FAQ and trust page. No matching term was located in a published customer agreement as of 29 Aug 2026, and the subscription agreement is not published.
The LexisNexis General Terms and Conditions (effective 5 June 2026), the provider terms for LexisNexis online services, warrant that subscriber inputs, including prompts, and subscriber 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, whether or not an AI feature also draws on web-sourced content. The annual SOC 2 Type 2 examination of Lexis and Lexis+ covers confidentiality and privacy criteria.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
The vendor publishes a direct answer in a product FAQ. Retention of user content and prompts is customized to the customer, set at the organization account level with the support of a client manager, and the vendor deletes in line with the customer's own retention policy. Users can also delete their content and prompts in the product at any time. That is real customer control over the window. No retention period is published, so the window is whatever the customer negotiates, and setting it runs through a client manager rather than an administrator control the customer operates alone.
The separately published zero retention statement covers the model provider API layer rather than the product's own storage.
Not addressed. No retention period is published for prompts, generated answers, conversational search history or uploaded documents, and nothing indicates whether retention is configurable or can be set to zero. Data minimization is named as a principle, which speaks to collection rather than to duration. Checked the product pages, the UK trust explainer, the trust center material and the launch material on 29 Aug 2026. Same registration gate limitation as above applies.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
Named document system integrations come without a stated rule on whether their permissions or ethical walls apply at query time. Integrations with iManage, NetDocuments and SharePoint are named, and corporate identity management integration is stated, but no vendor material was located addressing whether retrieval enforces source system permissions at query time per user, or how ethical walls and matter level segregation are handled.
A third party review refers to granular roles and ethical walls, which is not vendor material and was not treated as evidence.
Not addressed. No permission model, matter level restriction or tenant segregation description was located, and no document management system integration was found that would let retrieval inherit a firm's own permissions at query time. The product accepts uploaded documents and is sold into large firms where walls are a routine requirement, so the absence is material rather than incidental. Checked the product pages, the UK trust explainer, the trust center material and the launch material on 29 Aug 2026.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
Government and law enforcement requests for customer data have no product specific handling or transparency report. No clause addressing government or law enforcement requests for customer data was located for this product, and no transparency report was located. Noted for a future reader: a commitment to inform the customer of third party data access requests, audited under ISO 27701, was located on the trust center for a different Thomson Reuters product line, ONESOURCE Pagero. It was not treated as covering this product and does not support a value here.
The confidentiality clause of the LexisNexis General Terms, effective 5 June 2026, which govern Lexis+ AI, allows disclosure required by law, regulation or court order only with advance notice to the other party so that it can seek a protective order, and the terms apply that confidentiality standard to prompts and files submitted to the AI services. No transparency report was located.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
The source collection is named consistently and prominently. Output is grounded in Westlaw and Practical Law, proprietary databases the vendor owns and sells separately, maintained by more than 1,200 full time attorney editors. Vendor material is explicit that the system draws on this content rather than scraping the public internet. The rights basis is ownership rather than a stated license, and no update cadence or lag for the underlying content was located as of 29 Aug 2026.
LexisNexis names its sources by type but states no license basis. The grounding content is case law reviewed by Shepard's editors, statutes and Practical Guidance. That names the content types and the editorial review applied to the case law, rather than presenting it as raw text, which is a real provenance statement. Not stated: the license or public domain basis for any content type, a jurisdiction or court list, a date range, or an update lag for the content or the citator.
LexisNexis calls it the largest repository of accurate and exclusive legal content, which implies proprietary licensing without stating terms. Exclusive is a commercial claim, not a provenance disclosure.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
Treatment signals come from KeyCite, the vendor's own named commercial citation service, and appear with the authority. Vendor engineering material states that reports carry inline citations, KeyCite flags and excerpts, and that the research agent uses KeyCite warnings to flag questioned or overruled cases and adjusts its strategy accordingly. Separate vendor material describes the system recognizing red and yellow validity flags and factoring them into its analysis rather than only displaying them.
Lexis+ AI checks its citations against Shepard's Citations, an established commercial citator that LexisNexis owns. The product is stated to check every citation in a generated answer against Shepard's, so treatment checking happens automatically rather than as a separate step the user must remember. A user can also submit a specific citation and be told when it may be wrong. This pairs with the accuracy finding. Automatic Shepard's checking did not prevent an independently measured hallucination rate above 17 percent.
Citator checking shows whether an authority exists and how it has been treated, not whether the point drawn from it is correct.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
Research logs, reasoning steps and guardrails are all described, but not what happens when an answer cannot be grounded. Vendor material describes transparent research logs, step by step reasoning, an audit trail and prompting guardrails that restrict output to a defined set of validated content. None of that states what the product does when it cannot ground an answer. No explicit no answer or abstention path was located, and no confidence or grounding score exposed to the user was located either.
LexisNexis publishes nothing on what Lexis+ AI does when it cannot answer, though an outside study measured it. No statement describes a path for declining, an abstention behavior, a confidence score, or what the system does when its content does not support an answer. The Stanford RegLab and HAI study found incomplete answers, meaning refusals or ungrounded responses, on 18 percent of 202 queries. So some abstention exists, but LexisNexis does not describe it, and a rate measured by others is not a disclosure.
The gap matters for a product marketed as free of hallucination. The honest counterpart to that claim is an account of what happens when the system does not know, and none is published.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
A court record exists in which a predecessor brand of this product is mentioned. In Fletcher v. Experian Information Solutions, No. 25-20086, the Fifth Circuit sanctioned appellate counsel $2,500 on 18 Feb 2026 after a reply brief was found to contain 16 fabricated quotations and 5 further misrepresentations of law or fact. The court's findings address counsel's conduct: failure to verify, and evasive and misleading responses to the show cause order, which the court treated as conduct unbecoming under FRAP 46(c).
The products counsel identified as the possible source were named by the court as Casetext and vLex, described as generative AI products. Casetext is the brand Thomson Reuters acquired in 2023 and no longer sells under its own name, and CoCounsel is the surviving product. The court made no finding about either product's own output. The AI Hallucination Cases database maintained by Damien Charlotin was also searched.
None located, with the search method named. General web searches combining the vendor and product names with court, order, opinion, sanction, disciplinary and fabricated citation terms returned nothing on 29 Aug 2026. No docket database or court record tracker was searched, so the search was weaker than this product's prominence warrants. The Stanford RegLab and HAI study is an academic evaluation measuring a hallucination rate, not a court record.
It is covered under citation accuracy, not here, and should not be read as a judicial finding. This describes what this search found, not a clearance.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
A named framework and a buying guide address professional standards in general terms, but no specific ethics opinion is engaged. The vendor publishes a substantial named framework and a buying guide addressed to professional standards in general terms, and markets a responsible AI position to courts and government. No engagement with any named ethics opinion was located, including ABA Formal Opinion 512 and state bar guidance.
No engagement with bar guidance was found in the material read. No named ethics opinion, ABA Formal Opinion 512 or state bar guidance appears across the product pages, launch and press material, the UK trust explainer and the trust center material checked on 29 Aug 2026. LexisNexis publishes a very large body of practitioner content, including professional responsibility material, and that body was not reviewed, so this finding could change. On the pages a buyer evaluating the product would look at, no bar guidance appears.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
Vendor material leads with time compression and quantifies it: a 61 percent reduction in time to draft correspondence, a customer story reporting an hour's task completed in five minutes or less, and a commissioned return on investment study. The product page, the AI hub, the legal blog and the customer stories were searched on 29 Aug 2026. No per matter record of AI assisted work intended for fee purposes was located, and no published guidance on billing, fee or client disclosure treatment. The savings figures are published without anything on the client's side.
Not addressed. Nothing published addresses billing for AI assisted time, and no exportable record was located that a firm could use to show a client what portion of work was machine generated. The product's own commercial terms are not published either, so neither side of the fee question has a public answer. Checked the product pages, the launch and pressroom material and the trust center material on 29 Aug 2026.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
A published FAQ names the third party model providers that may process customer content, OpenAI and Google, and states the contractual prohibition imposed on them, with no agreement needed to read it. A corporate trust center publishes the information security framework. Not located as of 29 Aug 2026: a maintained subprocessor list for this product, and any consent or notification pack a firm could forward to its own client.
LexisNexis offers a working route to its security documents, but nothing a firm can forward without registering. The trust center at trust.lexisnexis.com runs on SafeBase and invites users to register for compliance reports, including the SOC 1 Type 2 report. The annual SOC 2 Type 2 examination of Lexis and Lexis+ is described with its full five principle scope. A firm answering an outside counsel guideline questionnaire has a set place to go and named reports to point to, without a sales call.
Nothing is open: no downloadable summary, published subprocessor list, named model provider or DPA was found outside the registration step. That step was not completed, so the portal's contents were not checked.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Several elements of a disclosure record are available and documented. Vendor material states that each research report includes transparent research logs with verifiable citations and step by step reasoning. Sources are hyperlinked with highlighted excerpts in a dedicated tab, and KeyCite validity flags appear with the authority, which together show what was retrieved and its status. Two elements are missing. The model used is not identified per output, since the vendor says the system works across multiple providers, and no export per document covering model, sources and human verification together was located as of 29 Aug 2026.
Lexis+ AI records its sources well, but nothing else a court might ask for. Citations are linked in every answer and checked against Shepard's, and a user can submit a citation to be checked, so the authorities relied on can be identified and their treatment checked. The other two things a court standing order asks for are missing. Nothing records which model produced an answer, and no record of human verification is captured or exportable.
No export of any kind was found. Given the measured hallucination rate, verification is the part that matters. A court asking whether a person checked the output would find the product keeps no evidence either way.
The questions both sides leave open
Derived from the records above rather than written, so it cannot favor either vendor. Take these into both conversations and ask each side the same question.
- Ethical Walls and Matter Segregation
- Refusal and Uncertainty Behavior
Which one fits
Choose CoCounsel Legal if
- Your firm already runs on Westlaw and Practical Law. CoCounsel Legal is sold in tiers bundled with them and grounds its answers in that content, with KeyCite flags shown beside each authority.
- The assistant should work inside your document system. Named integrations reach iManage, NetDocuments and SharePoint, along with a Word add in, HighQ and an MCP route from other AI clients.
- You want multi step work planned for you. The agentic release shows a matter specific plan before it executes, then returns work product with citations you can open and check.
Choose Lexis+ AI if
- Your firm already researches in Lexis+. Lexis+ AI sits on the same case law, statutes and Practical Guidance, and checks every citation in an answer against Shepard's.
- You want the no training promise in a contract. LexisNexis's General Terms of 5 June 2026 warrant that prompts and uploaded files will not train any large language model.
- You want an intellectual property indemnity that survives the cap. The same terms indemnify subscribers against patent, copyright, trademark and trade secret claims outside the twelve month liability limit.
In summary
CoCounsel Legal
CoCounsel Legal is Thomson Reuters' flagship legal AI product, grounded in Westlaw and Practical Law content and descended from Casetext, which Thomson Reuters bought in 2023. Its August 2026 release was rebuilt as an agentic experience on Anthropic's Claude Agent SDK, so a lawyer describes a matter and the system plans, researches, reasons and drafts with citations. It is sold in tiers bundled with Westlaw and Practical Law and reaches more than 200 law schools. In the AI Legal Index's reading, its research agent uses KeyCite warnings to flag questioned or overruled cases and changes its search strategy in response. Integrations span iManage, NetDocuments, SharePoint, Microsoft 365 and a Word add in, and retention of prompts and content is set by each customer.
Lexis+ AI
Lexis+ AI is LexisNexis's generative AI product for legal research, drafting and summarization, built on the Lexis+ research platform and generally available in the United States since October 2023. It offers conversational search, drafting, summarization and document upload, grounded in case law reviewed by Shepard's editors, statutes and Practical Guidance. The AI Legal Index finds that every citation in a Lexis+ AI answer is checked against Shepard's automatically, and a user can submit any citation to be checked. LexisNexis's General Terms of 5 June 2026 warrant that prompts and files will not train any large language model, and an annual SOC 2 Type 2 examination covers all five trust principles. Its price and its model provider go unnamed.
Questions buyers ask
CoCounsel vs Lexis AI: which is better for a litigation firm?
The one built on the research platform the firm already uses usually fits best. CoCounsel Legal brings Westlaw, Practical Law and KeyCite, plus an agentic mode that plans multi step work, and integrates with iManage and NetDocuments. Lexis+ AI brings Lexis case law, statutes and Practical Guidance with every citation checked against Shepard's, and supports uploading documents for summary and drafting. From the AI Legal Index, based on each vendor's own published materials as of October 17, 2026. No vendor pays for placement.
Does CoCounsel or Lexis+ AI train AI on client data?
Neither, by its own commitment. Lexis+ AI's General Terms warrant that prompts and files will not train any large language model and treat them as confidential. CoCounsel Legal keeps customer content and prompts out of training under a product FAQ, and its partners OpenAI and Google are contractually barred from training on them. The first is a contract term, the second a policy. From the AI Legal Index, based on each vendor's own published materials as of October 17, 2026. No vendor pays for placement.
Which AI models do CoCounsel and Lexis+ AI use?
CoCounsel Legal's current release runs on Anthropic's Claude Agent SDK, with OpenAI and Google named as AI partners that are contractually barred from training on customer data. Lexis+ AI was built internally with technology partners on a proprietary retrieval platform, and its model provider and model family go unnamed. From the AI Legal Index, based on each vendor's own published materials as of October 17, 2026. No vendor pays for placement.
Which one integrates with iManage or NetDocuments?
CoCounsel Legal integrates with iManage, NetDocuments and SharePoint. Microsoft 365, HighQ, Legal Tracker, Icertis and a Word add in sit beside them, plus an MCP route from other AI clients. Lexis+ AI offers no named document management connector, so documents reach it by upload. From the AI Legal Index, based on each vendor's own published materials as of October 17, 2026. No vendor pays for placement.
What gaps do CoCounsel and Lexis+ AI share?
Neither product has a stated fallback for when it cannot ground an answer, such as declining or flagging low confidence. Neither offers walls between matters at query time that a firm can rely on, which matters for firms with conflicts screens. Neither ties its controls to a named ethics opinion such as ABA Formal Opinion 512. The AI Legal Index finds both strong on linking every answer back to authorities a lawyer can open.
Neither vendor paid for inclusion, placement or a grade, and neither reviewed this page before it published. Everything on it comes from public material on the dates shown. How the index grades.
Both products are sold mainly to firms already licensing the parent's research platform, so a side by side switch is rare and the bundled price usually decides. Neither AI product has a price of its own. A commitment in a product FAQ, as on the CoCounsel side, is a policy rather than a contract term, which matters to a firm's general counsel. CoCounsel Legal works mainly as a general legal assistant and Lexis+ AI mainly as a research tool, and each also covers the other's ground. Thomson Reuters owns CoCounsel Legal, which descends from Casetext, and LexisNexis, part of RELX, owns Lexis+ AI. Neither vendor reviewed this page.