Counsel Stack vs Omnilex: how they compare in 2026
These two never meet on a shortlist, since Counsel Stack researches United States federal, state and local doctrine and Omnilex is a Swiss law workspace. Side by side they read as near mirror images. Counsel Stack placed first among all participants in the 2025 Vals Legal AI Report legal research study, scoring 81 percent accuracy against a measured lawyer baseline of 71. It publishes almost nothing about how it handles what a lawyer puts into it. There is no privacy policy, security page or customer agreement, and no retention terms or hosting statement. Omnilex is the reverse. It publishes per seat prices with onboarding included, hosting on Microsoft Azure in Zurich, and ISO/IEC 27001:2022 audited across the organization. Its trust center carries a dated penetration test report, and the five frontier models a user can choose are named to the version. It publishes no accuracy measurement at all.
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.
For the research product, the AI is the product. The vendor describes a fleet of specialized research language models built by its attorneys, engineers and legal professionals for particular practices and jurisdictions, and the evaluator independently confirms multiple specialized models built for legal research. Retrieval, synthesis and citation verification are all model driven. The company's legal marketing, lead generation and web design businesses have nothing to do with AI and are separate from the research product the benchmark measured.
The models drive what people use, on top of a large content collection. The site counts 1,057,541 sources, mostly 797,321 Swiss court decisions and 22,181 laws and articles, and that collection would still work as a Swiss law database without any model. The AI adds a reasoning layer over it and one thing that exists only because of the models: the Omnilex Commentary, 222,525 annotations the company writes article by article. Agents built without code from a firm's own playbooks sit on top. The pricing page sells sources, models and agents together as one license.
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 measured independently. In the 2025 Vals Legal AI Report legal research study, Counsel Stack placed first among all participants on all three criteria, ahead of every other legal AI product, a generalist model and the lawyer baseline: 81 percent accuracy against a 71 percent lawyer baseline, 77 percent authoritativeness against 68 percent, 71 percent appropriateness against 60 percent, and a 78 percent weighted score against 69 percent. It answered 196 of 200 questions, with four technical timeouts the evaluator published with the scores. The evaluator set the rubric and weights in advance and Am Law 100 firms supplied the questions, so the figures can be checked on the evaluator's own site. The product also ships structured citation verification, saving proof of verification automatically in the research record. The evaluator found that all systems struggled with multi jurisdiction questions and did worse than a generalist model on fifty state statutory surveys.
Grounding is documented and clearly labeled. Omnilex says every claim is cited and ready for the file, and its home page example shows an annotation to article 336c of the Code of Obligations footnoted to its authorities. Users choose whether research draws on court rulings, commentary or internal data. The commentary Omnilex generates is labeled as AI drafted, so a lawyer can tell machine written analysis from published doctrine. No accuracy figure, evaluation, benchmark or test set is published, and clause 6 of the terms says the provider accepts no liability for the accuracy or legal usability of generated content.
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.
Autonomy is present, and oversight is implied by record keeping rather than described as a control. The product automatically saves research, conversations, sources and proof of citation verification, so a supervising lawyer can reconstruct what was done, and a published customer account describes the tool helping a small team rather than replacing people. Reconstructing work afterward is transparency, not a control point. Searched the vendor site, deep research product page and third party coverage and found nothing on what the system does unaided, when it stops or escalates, what a lawyer must review before output is used, or what the vendor commits to when output is wrong. A traceability feature does not amount to oversight.
Review is a contract duty, and agents run when a user starts them. Clause 6 of the terms says the user remains responsible for reviewing, adapting and using all content, which puts the check on the lawyer in the binding document rather than in marketing. The product description matches: clause 2 speaks of AI assisted suggestions, and the product page describes a system that analyzes documents, laws and judgments, structures arguments and highlights relevant content. Agents are built from a firm's own playbooks and review steps and run on demand, so they start only when a person starts them, and the worked example shows its four steps. The boundary is not published. Nothing says what an agent completes without a human once started, what happens when a step fails or returns nothing, or what separates a suggestion from an action. The firm writes the workflow, so the limits are set by the customer and not described.
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.
Testimonials stand in for deployment evidence, though they are specific about the practice. Published accounts describe a solo attorney working with one other attorney, a paralegal and an intern who replaced their practice management software and use the product for research, and a legal content creator using it for Pennsylvania and federal law. These are real descriptions of use, honest about very small firm size. Searched the vendor site, product pages and third party coverage and found no named organization with figures, a date or a method. The independent benchmark measures test performance, not production outcomes.
Ten customers are named by logo: the Swiss Confederation, Bayer, the legal protection insurer Dextra, the pension fund Asga, the disability charity Pro Infirmis, and the firms SwissLegal, Ruoss Voegele, Neuwerk, Domenig and Seeger Frick. Two have case pages with named people on video: Florian Probala, Chief Claims Officer at Dextra, and Dr. iur. Marco S. Marty, Founding Partner at MV Legal Partners. One figure is published: over 6,000 lawyers use the platform. That counts adoption, not outcomes; no figure is tied to a named customer, and nothing is dated.
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.
Searched the vendor site, the deep research product page and third party coverage, and ran targeted searches for a Counsel Stack security page, privacy policy or trust center without reaching one. Nothing is published on how client confidences are handled: no statement on training, no retention or deletion terms, no segregation model, no treatment of attorney client privilege or work product. The product saves research, conversations and sources by design, so material accumulates in it, and no published terms attach to that material.
Omnilex addresses professional secrecy in Swiss terms but says nothing about training. Its home page says models are processed on Swiss infrastructure so attorney client privilege stays intact, and that the platform was built for professional secrecy from day one. The license agreement and data processing agreement are described as drafted to respect professional secrecy and auxiliary person liability, the Swiss doctrine on a lawyer's responsibility for third parties they engage. Clause 7 of the terms covers deletion after the contract ends, subject to statutory retention. Two gaps remain. Nothing describes separation between customers, users or matters, although work is organized into named projects and mandates. And nothing says whether customer content is used for training.
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.
A real position is enforced as a product gate rather than stated as a disclaimer. Access is limited to US licensed attorneys, verified at signup, and the vendor says only US licensed attorneys qualify at this time, so a nonlawyer cannot get the tool. That sets who may hold the product but not what the lawyer remains responsible for: searched the vendor site and product pages and found nothing on competence or supervision duties, the advice line, or jurisdiction limits within the United States, though the product spans federal, state and local doctrine.
The terms take a position on tool versus advice. Clause 2 describes the service as software that supports users on legal questions with AI assisted suggestions. Clause 6 says the provider accepts no liability for the accuracy or legal usability of generated content and that the user is responsible for reviewing, adapting and using all of it. The product covers Swiss law throughout, and the terms are governed by Swiss law with jurisdiction in Zurich. Professional secrecy and auxiliary person liability are named. No statement says output is not legal advice, and nothing covers the professional relationship. No bar or professional body guidance is named, neither the BGFA nor Swiss Bar Association material.
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.
Agreeing to independent measurement and to publication of the results is the governance act here. Counsel Stack was one of four systems evaluated and named in the Vals legal research study, where the evaluator and trade press both reported that major vendors declined and at least one took part but then withheld permission to publish. Allowing publication of results that included four technical failures is a real accountability position, but it is a one time evaluation, not a management system. Searched the vendor site, product pages and third party coverage and found no AI governance framework, AI management certification, named owner of model governance, testing regime before release, or anything on uneven output across matter types, parties or populations.
Omnilex's SafeBase trust center lists AI Security, AI Governance and AI Monitoring as published items, which shows a governance position exists, but the items themselves do not display on the public portal. Nothing addresses bias or uneven output, such as whether results differ across cantons, languages, court levels or party types, which matters in a federal system with three official languages.
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.
Searched the vendor site, the deep research product page and third party coverage, and ran targeted searches for a security page or trust center without reaching one. No retention period, no deletion control, no access control detail, no encryption statement, no hosting disclosure, no subprocessor list and no incident or breach notification practice is published. The gap is sharpened by the product's own design: it automatically saves research, conversations, sources and timekeeping entries, so it accumulates a durable record of a lawyer's work with no published terms governing that record.
Most of the ground is covered, with two gaps. Access controls are stated: SAML single sign on, audit logs, IP allowlisting and data lifecycle management, plus credential management, brute force protection, account hardening, endpoint detection and response and disk encryption in the trust center. Data is encrypted at rest and in transit. Amazon Web Services and Microsoft Azure are named as subprocessors. A dedicated incident response team is stated, and backups, business continuity, disaster recovery and DDoS protection appear as trust center items. The gaps: retention is given only as deletion after the contract ends unless the law requires otherwise, with no period during the term, and there is no deadline for telling customers about an incident, with the trust center saying more detail is available on request.
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, liability cap, carve out, warranty on output or insurance position is published, and no customer terms of service is published as published on the property. Recorded as a pure absence on the surfaces reached. Worth naming the shape of it: this vendor's entire positioning rests on measured accuracy and structured citation verification, and it publishes nothing about who bears the loss if a verified citation is wrong.
Clause 6 of the terms says use is at the user's own risk, the provider accepts no liability for the accuracy or legal usability of generated content, the user is responsible for reviewing and using all content, and liability is limited to gross negligence and intent. That is the widest exclusion Swiss law permits, so nothing is recoverable for ordinary negligence, and no monetary cap is set. Cyber insurance is listed as a trust center document. No indemnity runs either way, and clause 6 disclaims any warranty on output. Clause 5 targets at least 90 percent availability outside maintenance windows and says this is not guaranteed. Clause 9 says use rests on a separate license agreement, which is not published and sits behind the trust center access flow, so its terms are not public.
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.
Integration routes exist and none is documented. API access is offered on request, and a customer account describes the product replacing their practice management software, which suggests matter and time tracking built into the product rather than connected to it. Searched the vendor site, product pages and third party coverage and found no integrations page, no named connector for document management, practice management, court filing or Microsoft Word, and no API documentation available without contacting the vendor.
Integration is claimed but not specified. The product page says a firm can connect its document management and legal tools without media breaks and keep its workflows. None of those systems is named: no document management, practice management or electronic filing product, no connector list and no integrations page. Enterprise plans include custom integrations and APIs, but no API or developer documentation is published. What is concrete is SAML single sign on across plans and company SSO at enterprise level.
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.
Searched the vendor site, deep research product page and third party coverage, and ran targeted searches for infrastructure or security documentation without finding any. Nothing was found on deployment: no tenancy statement, hosting provider, region, residency option or processing location. The product is evidently cloud delivered and reached in a browser, but that is inferred from how it is sold, not stated.
Residency is stated with a named city and cloud, and tenancy is not addressed. Omnilex says it is hosted on Microsoft Azure in Zurich, that data is encrypted at rest and in transit and hosted entirely on its Swiss Azure cloud, and that models run on Swiss infrastructure. The pricing FAQ says data is processed in Switzerland and in selected EU countries that the Swiss Federal Data Protection and Information Commissioner recognizes as adequate. Two gaps remain. Nothing says whether the platform is multitenant or single tenant, or how one firm's data is kept apart from another's. And hosting described as entirely Swiss sits beside processing that extends to EU countries, with no explanation. No on premises or customer cloud option is offered.
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.
Two targeted searches for the vendor's security page, privacy policy, terms, trust center or certifications found no vendor material. A failed search is weaker evidence than a page, so this shows that a buyer doing ordinary diligence would not find security material, not that none exists, and a single link could change it. The vendor publishes an independently measured accuracy score but nothing about how it handles data.
Omnilex runs a SafeBase trust center at trust.omnilex.ai that shows ISO/IEC 27001:2022 and lists documents including a penetration test report dated 5 September 2025, the Omnilex License Agreement with a DPA template, technical documentation, cyber insurance, a data security policy and a network diagram. Omnilex's home page says the certification is audited across the whole organization. The portal splits public and private documents behind a request flow and does not say whether access is instant on an email or NDA click, or only after review. No auditor, certificate number, issue date or validity period is published.
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 vendor describes its own model layer without identifying what sits underneath it. Published and independently confirmed: a fleet of multiple specialized research language models built for particular practices and jurisdictions, described by the vendor as assembled over years by its own attorneys, engineers and legal professionals. That tells a buyer the architecture is specialist rather than a thin wrapper. Searched the vendor site, the product pages and third party coverage and located no named foundation model or provider underneath that layer, no statement of where models run, no subprocessor list, and no commitment to notify customers when the supply chain changes.
Five models are named, with versions, as selectable on Omnilex's home page: GPT-5.6, Claude Opus 5, Mistral Large 3, Kimi K2.6 and Grok 4.6, with the user choosing which answers. Omnilex says every frontier model is processed on Swiss infrastructure so attorney client privilege stays intact, and hosting is on Microsoft Azure in Zurich. Omnilex's trust center names Amazon Web Services and Microsoft Azure as subprocessors. No model provider is named as a legal entity or listed as a subprocessor, so it is not clear who contracts for each model or what it may keep, and it is not stated whether the models run inside the Swiss setup or through the providers' own APIs. There is no commitment to notify customers when the model set changes.
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.
Real commercial terms are published at the entry point, but no rate is. A free tier can be reached without a sales conversation: the vendor says signup needs only a name and professional email and no credit card, with free trial access to its chat product. That is a published term and a self serve route. Searched the vendor site, product pages and third party coverage and found no rate for any paid tier, no unit of charge and no tier structure, with API access by contact request.
Rates are published with unit and term: CHF 2,000 per user billed annually for a single user, CHF 1,600 per user billed annually for teams from two users, and a custom agreement for enterprise from twenty users, all excluding VAT. Omnilex says the price scales only with the number of people who use it. Onboarding, setup, training and support are included in every plan, and Omnilex's FAQ says nothing essential is held back for a higher tier. Team adds collaboration and integration of a firm's own knowledge, and Enterprise adds SSO, custom integrations and APIs. A fourteen day free trial needs no card or commitment. Usage limits are published at up to 1,000 pages and 500MB per document. Only the enterprise price is withheld, which the page explains as a scoping conversation about integrations and support.
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.
Coverage is described in substance, and the target is stated honestly. Scope is federal, state and local US doctrine, with specialized models claimed for particular practices and jurisdictions. The vendor positions itself narrowly, for highly technical, complex, boutique, cross jurisdictional and multi doctrine research, which says where it is strong rather than claiming to cover everything. Published customer accounts come from very small firms, and no enterprise scale is claimed. Practice areas are not listed, state level coverage is not stated, and nothing says what the product does not cover; the measured weakness on fifty state surveys comes from the evaluator, not the vendor.
Four buyer groups each have a page described by the work: law firms doing research, drafting and review; in house teams, from contract review to regulatory questions; public administrations applying federal and cantonal sources; and insurers assessing claims against statute and precedent. Named references include the Swiss Confederation and a legal protection insurer. Swiss coverage is shown by the counted federal and cantonal sources. No practice areas are listed; home page examples cover employment, tenancy and accident insurance law, but only as illustrations. No firm size band is given beyond the seat thresholds in pricing, nothing says whether any jurisdiction besides Switzerland is served, and nothing says what is not supported.
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?
Training on customer research has no stated position, a gap that matters because the vendor says it has assembled its own specialized models over a number of years. No located material states whether customer content may be used to train models, either way. Recorded as silent under the rule that a value is never inferred from the absence of a contradiction. The question carries weight here because the vendor states it has assembled its own specialized models over years, so a buyer would want to know what went into them and whether their own research contributes.
Nothing published says whether customer content is used to train or improve models. Clause 7 of the terms covers data protection by reference to the privacy policy and deletion after the contract ends, and clause 8 reserves rights in the software, content, models and generated data to the provider, without mentioning training. The privacy policy covers the website only and points to the data processing agreement in the trust center for the app at app.omnilex.ai.
That agreement sits behind an access flow and was not available to read. Statements that models run on Swiss infrastructure and that the license respects professional secrecy do not mention training.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
Retention is a product feature with no period attached. The vendor says the product automatically saves research, conversations, sources, proof of citation verification and time entries, so keeping the record is what the customer buys. Searching the vendor site, product page and third party coverage on 29 Aug 2026 found nothing on how long any of it is kept, whether customers control that, or whether deletion is available.
A product built to keep a durable record of a lawyer's work, with no published terms governing that record, leaves this gap wide open.
Retention is acknowledged in the terms with no period attached. Clause 7 ties deletion to an event, the end of the contract, with no window for how long after and an open exception for statutory retention duties. Nothing says how long prompts, chats, generated drafts or uploaded documents persist during the term, and the customer is offered no retention setting. Omnilex's trust center lists a data erasure item and a data lifecycle management control among its defaults, and its home page names data lifecycle management as a control legal teams expect, but neither gives a figure.
The one specific limit published concerns size, not time: the pricing page caps documents at 1,000 pages and 500MB. The data processing agreement, where retention detail would normally sit, is behind the trust center access flow.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
Segregation between customers, users or matters has no described model, and no document system permissions are inherited. No vendor material addresses segregation of any kind: not between customers, not between users, not between matters. No document management integration was located whose permissions the product could inherit. Noted for context: published customer accounts describe very small firms of two attorneys plus support staff, where formal ethical walls are less commonly operated than at scale, though conflicts obligations apply regardless of firm size.
Nothing describes separation between customers, between users inside a firm, or between matters. The product is organized around matters, with the interface showing named projects such as a share purchase, a due diligence, a litigation matter and an estate, and the pricing page adds team and project collaboration at the Team tier and know how integration, so the structure for separation exists. What is missing is any statement that it works as a boundary: nothing says whether one user in a firm can reach another's project, whether ethical walls can be set up, or how tenants are isolated from each other.
Access controls are published at the account level instead, with SAML single sign on, audit logs and IP allowlisting. The gap sits awkwardly with Omnilex's own focus on professional secrecy, since conflict walls are part of what Swiss professional secrecy asks of a firm.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
Research, conversations and sources stored against a lawyer's matters have no stated notice protection when a government or law enforcement body comes asking for them. No clause committing to notify a customer of a government or law enforcement request for their data was located, and no transparency report was located. The gap matters because the product durably stores research, conversations and sources tied to a lawyer's matters.
Nothing addresses disclosure to authorities or in response to legal process, or notice to the customer if a demand arrives. The terms have no confidentiality article and no clause on required disclosure. The privacy policy lists legal obligations as a basis for processing, but it covers the website rather than the application, and its recipients are website service providers. The license agreement with a DPA template, the document where such a clause would sit, is listed in the trust center behind an access flow.
Omnilex's focus on professional secrecy makes the silence more notable, since a demand for a lawyer's file is exactly the case auxiliary person liability is meant for.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
Coverage is stated by jurisdiction level, without identifying sources. The vendor says the product synthesizes federal, state and local US doctrine using models grounded in US law for particular practices and jurisdictions, and the evaluator confirms broad federal and state coverage on its test set. Searched the vendor site, product page and third party coverage on 29 Aug 2026 and found no named source or publisher for primary law, no license or public domain basis, no completeness statement per jurisdiction, and no update schedule or lag. Local doctrine is claimed without saying which localities, a significant undefined boundary for a research product.
Sources are identified by category and counted: 797,321 court decisions, 222,525 commentaries, 22,181 laws and articles, 13,162 official documents and 2,352 pieces of literature, 1,057,541 in total, described as federal and cantonal material refreshed continuously. The Omnilex Commentary is Omnilex's own, written article by article, labeled as AI drafted and cited to statute and case law. No rights basis is stated: no publisher, database or data supplier is named for any category, and no license or permission is described.
That matters most for the 2,352 pieces of literature and whatever underlies the commentary category, since Swiss legal literature is commercially published and, unlike federal and cantonal case law, is not free to redistribute by default.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
A verification mechanism is published and what it verifies is not defined. The vendor puts structured citation verification at the center of its positioning and states that proof of citation verification is saved automatically with each research record, and the independent evaluator scored it highest of all participants on authoritativeness at 77 percent, defined as citing relevant primary sources that are valid and support the statements made.
Searched the vendor site, the product page and third party coverage on 29 Aug 2026 and located nothing stating whether verification checks that authority exists and supports the proposition, or whether it also checks subsequent history and treatment. Those are different questions and no commercial citator license was located. The good law question specifically is unanswered.
Nothing says whether a decision the product returns is still good law: no treatment flags, no subsequent history, and nothing on whether a judgment has been confirmed, distinguished or departed from. The nearest claims are about the freshness of the collection, not the standing of any one authority. Sources are described as refreshed continuously, and the product page offers to keep users informed about the latest changes in legislation and jurisprudence at federal, state and cantonal levels.
That covers what has been added, not what has been overtaken. It matters here because the product retrieves case law and its own commentary as the basis for cited answers, and Swiss practice depends heavily on whether the Federal Supreme Court has since moved.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
Technical timeouts are the only recorded non answers, and no principled abstention path is described. No published material describes what the product does when it cannot ground an answer, and no explicit no answer path or confidence signal exposed to the user was located. The independent evaluator recorded four cases where the product returned nothing due to technical timeouts, which is a failure mode rather than principled abstention, and unlike one other participant in the same study no instances were recorded of it explaining that sources did not support an answer.
Nothing describes what the platform does when it cannot ground an answer, and no confidence, relevance or grounding indicator is described. The published answer is that the user checks: every claim is said to be cited, the commentary says every statement is citable, and clause 6 of the terms makes the user responsible for reviewing and adapting all content. Users can limit which source types are consulted, which shapes retrieval but says nothing about what happens when retrieval comes back thin.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
No court order, opinion or disciplinary record naming this product has been located as of 29 Aug 2026. Instrument searched: the AI Hallucination Cases database maintained by Damien Charlotin, which tracks court decisions worldwide involving AI generated hallucinated content and records the AI tool implicated where it is known. Also checked published 2026 sanctions summaries and secondary sanctions trackers. The entries located name filers, and in some rows other products, rather than this one.
This is a statement about the public record on the date shown and not a clearance. The exposure is direct: a US litigation research product produces authority bound for filings, and this vendor's independently measured authoritativeness score was the highest in its study.
The AI Hallucination Cases database maintained by Damien Charlotin, and reporting based on it, records no court order, opinion or disciplinary record naming Omnilex or Omnilex AG. The database is weighted heavily to United States filings, its European coverage is thinner, and Switzerland is not among the jurisdictions commonly reported in it, so a Swiss product is less likely to appear even if an incident occurred. Swiss judgments are also published in three languages, which limits how reliably an English language search reaches them.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Access is limited to verified US licensed attorneys, but no ethics opinion or professional guidance is engaged. No engagement with any ethics opinion or professional guidance was found, including ABA Formal Opinion 512 and state bar guidance. The vendor limits access to US licensed attorneys and verifies licensure at signup, a real position on who may use the tool, covered under professional responsibility, but that is not engagement with published guidance.
Omnilex refers to professional duties directly but names no authority. It says the platform was built for professional secrecy from day one, that models are processed on Swiss infrastructure so attorney client privilege stays intact, and that its license agreement and data processing agreement respect professional secrecy and auxiliary person liability. Neither the Federal Act on the Free Movement of Lawyers nor the criminal provision on professional secrecy is named, and no Swiss Bar Association or cantonal bar guidance is cited. Nothing says what a lawyer must do to meet those duties when using the product.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
The product keeps a record of AI assisted work. It automatically saves time entries alongside research, conversations, sources and proof of citation verification, a per matter record the firm can use, so a firm asked what was done and how long it took has something to produce. The vendor publishes no guidance on treating that record for fees or client disclosure: searched the vendor site, product page and third party coverage on 29 Aug 2026 and found nothing on billing, no position on whether AI assisted time is billable, and no client disclosure material. The record exists; the guidance does not.
Nothing addresses how AI assisted work is recorded, billed or disclosed to a client, and no record of it per matter is described. Omnilex makes no quantified savings claim; its language is about being faster, more precise and dependable, with no time or cost figure. A third party report credits users with a 93 percent research time saving, a figure Omnilex does not publish. Work is organized into named mandates with audit logs, but neither is described as something that can be reported against a bill.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
A firm bound by a client AI clause would find no subprocessor list, model provider statement, certification or trust center to answer it with. No subprocessor list, no statement naming which model providers see client content, no published data processing agreement, no certification and no client facing consent or notification material was located. A firm bound by a client AI clause could not assemble a response from this vendor's published material, notwithstanding that the vendor's own models being proprietary would simplify the answer if it were stated.
A subprocessor list is published and reachable without an agreement: the legal section of Omnilex's trust center names Amazon Web Services and Microsoft Azure, and the same portal lists a data processing agreement, cyber insurance and a license agreement with a DPA template. The pricing FAQ says a data processing agreement is part of every license. The two subprocessors are infrastructure only. Five models are named on the home page as selectable, and none of their providers appears in the list or anywhere else as a contracting party.
The license and DPA sit behind the trust center access flow, so what a firm can hand a client without asking is the subprocessor list and the certification, not the documents.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
A strong disclosure trail exists, short of identifying the model. The product automatically saves research, conversations, sources and proof of citation verification, so a lawyer facing a court's AI disclosure order has a kept record of what was asked, retrieved and cited and that citations were checked, created in normal use rather than assembled afterward. No export packaging that record as a single document was found, and no model is named anywhere, so the model used could not be stated. Litigation research output is among the most likely to attract such an order.
Nothing addresses disclosing AI use to a court or authority, and no exportable record for each document of the model used, sources retrieved and human check is described. The product leaves a trail for the lawyer's own use: every claim is cited, the commentary is labeled as AI drafted, and audit logging is a standard control. None of these is presented as a record of model use, and none shows which of the five selectable models produced a given passage.
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
- Third Party Request and Subpoena Notice
- Good Law Verification
- Refusal and Uncertainty Behavior
Which one fits
Choose Counsel Stack if
- You want research measured by somebody other than the vendor. Counsel Stack placed first among all participants in the 2025 Vals Legal AI Report legal research study. It scored 81 percent accuracy against a measured lawyer baseline of 71 percent, 77 percent authoritativeness against 68, and 71 percent appropriateness against 60. The evaluator set the rubric in advance, using questions contributed by Am Law 100 firms, and published Counsel Stack's four technical timeouts alongside the scores.
- Your citations have to be checkable afterward. Counsel Stack ships structured citation verification and automatically saves research, conversations, sources, proof of citation verification and timekeeping entries. A supervising lawyer can reconstruct what was searched, what was relied on and what was verified.
- You want the tool restricted to lawyers. Access is limited to United States licensed attorneys, verified at signup, a professional boundary implemented as a product gate rather than asserted in terms. The entry point is genuinely self serve, requiring a name and professional email with no credit card.
Choose Omnilex if
- You practice Swiss law and the corpus has to be Swiss. Omnilex counts its sources openly at 1,057,541 across federal and cantonal material: 797,321 court decisions, 222,525 commentaries, 22,181 laws and articles, 13,162 official documents and 2,352 pieces of literature. It writes the Omnilex Commentary itself, article by article, labeling it as AI drafted and citing it back to statute and case law. A lawyer can tell machine written analysis from published doctrine.
- Professional secrecy is the whole question. Omnilex hosts on Microsoft Azure in Zurich and states that models run on Swiss infrastructure. It processes in Switzerland and in EU countries the Swiss data protection authority recognizes as adequate, and it holds ISO/IEC 27001:2022, audited across the organization. Its trust center carries a penetration test report dated 5 September 2025, a license agreement with a data processing template and a cyber insurance document.
- You want the price and the implementation cost in the same place. Omnilex publishes CHF 2,000 per user billed annually for a single user and CHF 1,600 per user for teams from two, excluding VAT. Onboarding, set up, training and support are stated as included in every plan. There is a fourteen day free trial requiring no card and published usage limits, and the user chooses which of five named frontier models answers a given question.
In summary
Counsel Stack
Counsel Stack is a deep legal research platform for United States law firms. The vendor describes it as a fleet of specialized research language models assembled for particular practices and jurisdictions. It offers structured citation verification and automatically saves research, conversations, sources and proof of verification. Access is restricted to United States licensed attorneys, verified at signup. According to the AI Legal Index, it placed first among all participants in the 2025 Vals Legal AI Report study. It publishes no privacy policy, security page or customer agreement, and no retention terms or hosting statement.
Omnilex
Omnilex is an AI workspace for Swiss law, bringing statutes, court decisions, commentary and a firm's own know how into one research and drafting environment. Its corpus is counted openly at more than a million sources. It writes its own article by article commentary, labeled as AI drafted, and lets a firm build no code agents from its own playbooks. The AI Legal Index records published per seat prices, with onboarding and support stated as included in every plan. Hosting is on Microsoft Azure in Zurich with ISO/IEC 27001:2022 certification, and five frontier models are named to the version. No accuracy measurement or AI governance framework is published.
Questions buyers ask
Are Counsel Stack and Omnilex alternatives to each other?
No. Counsel Stack researches United States federal, state and local doctrine and restricts access to US licensed attorneys. Omnilex is a Swiss law workspace, hosted in Zurich and sold in Switzerland. They will never appear on the same shortlist. The comparison is worth reading because the two fail in opposite places: one is measured and undocumented, the other documented and unmeasured. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
Which one has been measured?
Counsel Stack. It submitted to an independent benchmark and placed first among participants, with the evaluator publishing the rubric in advance and the failure cases alongside the scores. Omnilex publishes no accuracy figure, evaluation, benchmark or test set. Clause 6 of its terms states that it accepts no liability for the accuracy or legal usability of generated content. That is candor about the limitation rather than evidence about the rate. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
What can you learn about data handling?
On Omnilex, a good deal: Azure hosting in Zurich, processing in Switzerland and adequate EU countries with the supervisory authority named, and ISO 27001:2022 across the organization. It also lists SAML single sign on, audit logs, IP allow listing, named infrastructure subprocessors and deletion after the contract ends. Counsel Stack publishes nothing on data handling: no statement on training, no retention period or deletion route, no encryption detail, no hosting disclosure and no subprocessor list. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
What does each cost?
Omnilex publishes CHF 2,000 per user annually for a single user and CHF 1,600 per user for teams, excluding VAT, with onboarding and support included and only the enterprise tier withheld. Counsel Stack publishes a free tier reachable without a sales conversation and no rate for any paid tier, no unit of charge and no tier structure, with API access directed to a contact request. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
What do Counsel Stack and Omnilex both leave unpublished?
Neither publishes an AI governance framework with a named accountable owner or a pre release testing regime, and neither publishes anything on uneven output, which on Omnilex means nothing states whether performance holds across cantons, languages or court levels. Neither documents an integration an implementer could use, with Omnilex claiming document management connections without naming a system and Counsel Stack offering API access on request without documentation. And neither states what its product does not cover. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
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.
The two fail in opposite places. Counsel Stack publishes no privacy policy, security page, trust center or customer agreement, no retention or deletion terms, and no hosting statement or liability position. Its product automatically saves research, conversations and sources, so a durable record of a lawyer's work accumulates with nothing published governing it. On Omnilex, clause 6 of the terms limits liability to gross negligence and intent. That is the maximum exclusion Swiss law permits, and it leaves nothing recoverable for ordinary negligence. The availability target is 90 percent and stated as not guaranteed, and the license agreement that actually governs a customer is not published. The Vals study also recorded that every system it tested underperformed a generalist model on fifty state statutory surveys. Neither vendor reviewed this page.