LegalMation vs Turbo Law: how they compare in 2026
LegalMation and Turbo Law both sell AI to defense firms and insurance carriers handling high volumes of litigation: LegalMation drafts answers, affirmative defenses and discovery responses from an organization's past pleadings, and Turbo Law builds a line cited fact graph from the matter file and drafts, evaluates and models settlement value from it. Turbo Law sits in the top two bands on ten of fifteen axes and LegalMation on five of fifteen. The gap is almost all data handling. Turbo Law states that customer content never trains shared or public models, that it neither asserts nor waives privilege for the customer, that each firm runs in its own isolated Azure environment, and that it holds SOC 2 Type II. LegalMation's privacy policy expressly excludes the pleadings and discovery customers upload and points to service agreements it does not publish, and no security certification was located. LegalMation's counterweight is who uses it and what it requires: Walmart, Ogletree Deakins and Fisher Phillips are named with titles, and its terms require a licensed attorney in the forum state to review and sign every output. Neither publishes a price or its platform contract.
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 product is the model. Every one of the eight named modules is a generation or extraction capability, and there is no non-AI tier, no core platform underneath and no severable layer: Complaint Response reads a pleading and drafts the answer, Discovery Response and Subpoena Response draft responses, Demand Letter Response handles EEOC and similar claims, Deposition Assistant analyses transcripts, Case Summaries, Matter Profiling and Data Analytics extract and structure. The vendor's own site title is A.I. Agents for High-Volume Litigation and Claims, and a vendor announcement describes the technique stack directly as natural language processing and machine learning including generative AI. The customer onboarding step is itself a model step: work samples and company data are submitted to train the customer's agent so that output reflects the organisation's guidelines and legal approach. Remove inference from this platform and nothing remains that a buyer could use, which is the opposite of the position on two other records built in this pull where a published price tier or a contractual opt-in proved the AI severable. This is the strongest AI Centrality position located so far and it is graded on structure rather than on marketing emphasis. Checked 7 September 2026.
The models are the product, which is the A band. Turbo Law reads the whole matter file with AI to build its Fact Graph, and every module (review, drafting, research, strategy and settlement, and the assistant) generates work product from it: chronologies, contradiction detection, privilege flags, motions, discovery responses, case evaluations and settlement-value models. There is no underlying case management system that works without the models. Verified 18 September 2026.
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 asserted, grounding is claimed at a high level, and nothing is measured or verifiable. The grounding claim is real and specific to this product's design: generated answers and affirmative defences are drawn from the organisation's own historical and best responses, so output is anchored to the customer's precedent bank rather than to a general model's recall, and documents are described as tailored to the jurisdiction in which the action was filed. The accuracy claim comes from the vendor's own announcement of the demand letter capability, which states the automation ensures a higher level of accuracy and compliance. Neither is testable. No accuracy figure, no test set, no evaluation, no error rate and no failure mode is published anywhere, and nothing describes how a drafted allegation response links back to the precedent it came from or how a reviewer verifies it. The published website terms cut against the accuracy claim in plain terms: LegalMation does not warrant the accuracy, completeness or usefulness of the information it provides, states that because it is not a law firm it cannot guarantee the information is current, and warns that because the law varies between jurisdictions the user must ensure any document provided is correct for their specific jurisdiction. A product that drafts pleadings for filing and disclaims currency of the law is the shape this axis exists to surface. Checked 7 September 2026.
Grounding is real and described, and accuracy testing is claimed without figures, which is the B band. The security page says fact-based outputs are line-cited to a source document or the public record and that the platform will not state what it cannot cite, and that every drafting and analysis workflow is measured against vertical-specific accuracy standards before it ships; the product pages say every output is line-cited to the file and case-law search cites the relevant holdings. No accuracy standard, test set or result is published. Verified 18 September 2026.
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 written commitment that a supervising lawyer stands between the model and the filing, with a real review surface, short of the full control structure. The commitment is contractual and unusually strong in its wording: all information and documents provided must be reviewed and ultimately signed by a licensed attorney in the state where the matter is pending, and by using the service the customer agrees they will have a licensed attorney review any output generated. Tying review to the forum state rather than to review in the abstract is a level of specificity no other record in this pull matches. The review surface is described at the product level too, and in two shapes. In one, the user uploads the complaint, verifies key information, and then point-and-clicks the desired answer for each allegation through the interface, which is allegation-by-allegation human selection rather than bulk acceptance. In the other, the system auto-generates answers and affirmative defences from the organisation's historical responses and the user reviews them. What is not published is the control structure around those modes: nothing states which mode applies when, what the system does without a user, whether a confidence or uncertainty signal is surfaced, what happens after a wrongly drafted defence is filed, or whether any log records what the model produced against what the reviewer changed. Checked 7 September 2026.
A categorical review constraint on named high-stakes outputs, which meets the A band under R124(2). The security page says case evaluations, settlement recommendations and motion drafts surface for attorney review before anything goes external, and states the operating model as the platform proposes, the team verifies and partners decide; key conclusions are hyperlinked to source material for that verification. The assistant runs multi-step and scheduled tasks, but its output returns to the team as cited work product. No confidence threshold is published. Flagged for sampling. Verified 18 September 2026.
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.
The deepest named-customer evidence in the lane, and no method behind any figure. Attribution is unusually good: four testimonials carry full name, title and organisation, being Alan Bryan, Managing Counsel and Head of Commercial and Employment Litigation at Walmart; Ron Chapman, Partner and Executive Committee member at Ogletree Deakins; Evan Shenkman, Chief Knowledge and Innovation Officer at Fisher Phillips; and Michael Sperounes, Senior Vice President for Litigation at American Integrity Insurance Group, who states his company has been a customer since 2021. Logos add Baker Donelson, Nelson Mullins, Wood Smith, CSAA, AAA and Wawa, spanning corporate legal, AmLaw firms and insurance carriers. Figures are published and specific in form: over 100,000 matters processed per year, clients achieving up to 10x return, savings of up to 80 per cent of team time, and a forecast to save clients more than 100 million dollars this year. What holds this below the top band is that no figure survives scrutiny as a measurement. Every one carries an up-to or forecast qualifier, none states a baseline, a sample, a period or a method, and the largest is explicitly a projection rather than a result. A reader can see who uses this and cannot assess what it did for them. Checked 7 September 2026.
Outcome claims without named customers or method, which is the C band. Turbo Law reports more than 1,800 active matters on the platform and says firms using it report a 60 per cent reduction in non-billable hours per matter, 10 per cent fewer write-offs and profitability on alternative fee arrangements; no customer is named on the surfaces read and no method is given for the figures. Verified 18 September 2026.
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.
No published commitment covers the customer's legal documents, and the absence is express rather than inferred, which is the important distinction on this record. The privacy policy, effective 1 July 2024, carves the platform out of its own scope twice. Its introduction states that the policy does not apply to information collected from legal documents provided by customers relating to their use of the LegalMation platform. A dedicated section headed Legal Documents from Customers for Processing through the LegalMation Platform repeats the point for pleadings, discovery requests and responses, subpoenas and demand letters, stating that their processing is governed by the terms of the service agreements with customers and not by the privacy policy. Those service agreements are not published. The website terms, which are published, address confidentiality only in respect of the user's own password and account credentials and say nothing about LegalMation's obligations toward the material a firm uploads. The result is that the documents at issue are pleadings and discovery in live matters, frequently privileged and work product, and a buyer cannot establish from any published surface what confidentiality, segregation, access or handling commitments attach to them. Privilege and work product are named nowhere. The index looked, found the instrument that would normally carry this, and found it declining the subject and pointing to a contract the vendor does not publish. Checked 7 September 2026.
A substantive public position that addresses privilege, with the customer agreement unpublished, which is the B band. The privacy policy says Turbo Law does not use customer content to train or improve foundation models, that any fine-tuning happens only at the customer's direction, that third-party AI providers are contractually barred from training on customer data, that Turbo Law does not assert or waive privilege or work product on the customer's behalf, and that staff access is limited to authorised purposes; the security page adds tenant isolation that its own team cannot cross without a support ticket and client approval. The security page says the no-training commitment is enforced in contracts, but the customer agreement and DPA are not published. Verified 18 September 2026.
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 most complete treatment of this axis located in the corpus, and every limb of the top band is met in the published terms. What the product is and is not: LegalMation states it does not provide legal advice, opinions or recommendations, remedies, defences or strategies, does not apply the law to the facts of a particular situation, and does not engage in the practice of law as it is not a law firm and does not perform services performed by an attorney. Who may use it: the licence is granted only to a user who is a licensed attorney or working on behalf of a licensed attorney, so eligibility is a term of the grant rather than a recommendation. Competence and supervision: all information and documents provided must be reviewed and ultimately signed by a licensed attorney in the state where the matter is pending, and the customer agrees as a condition of use that a licensed attorney will review any output. Jurisdiction limits: a dedicated Geographic Restrictions section states the company is based in California, that its products are intended to assist attorneys practising under their respective jurisdictions' procedures and formatting requirements, that it does not guarantee accuracy for any particular jurisdiction, and that the website is for use only by persons located in the United States, with a further warning elsewhere that because the law varies between jurisdictions the user must ensure any document is correct for theirs. No consumer-facing surface exists; the product is sold only to corporate legal, law firms and insurance staff counsel, so the consumer disclosure limb does not arise. Checked 7 September 2026.
A clear statement that the product is software and not legal services, with the attorney's role stated, which is the B band. The Terms of Use say Turbo Law provides software, not legal services, that nothing creates an attorney-client relationship, and that legal decisions should be made by licensed attorneys exercising independent professional judgement; the security page says the platform is built to make litigators faster, not to practise law. The buyers are law firms and insurance carriers. No jurisdictional limits are stated. Verified 18 September 2026.
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.
No governance disclosure was located. There is no AI policy, no responsible AI or ethics statement, no governance framework, no bias or fairness discussion, no model evaluation or testing description, no named internal owner for AI decisions, and no alignment claim to any framework. Nothing addresses how models are selected, assessed before release or monitored in use. The gap sits on a product whose output is filed in court: drafted answers, affirmative defences, discovery responses and subpoena responses, produced at a stated volume of over 100,000 matters per year. Two adjacent things exist and neither is governance. The website terms impose an attorney review-and-signature obligation, which is an allocation of responsibility to the customer and is credited on the professional responsibility axis. And a vendor announcement names the technique stack as natural language processing and machine learning including generative AI, which describes what is used rather than how it is governed. Surfaces read on 7 September 2026: the homepage and full site navigation, the website terms in full, the privacy policy in full, the Complaint Response module page and the vendor's demand letter announcement.
A stated release gate with substance, short of testing results, which is the B band. The security page says every drafting and analysis workflow is measured against vertical-specific accuracy standards before it ships and does not ship if it fails; the Trust Center lists board oversight briefings, annual risk assessments and quarterly critical system reviews. The standards, results and any assessment of uneven performance are not published. Verified 18 September 2026.
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.
Nothing platform-specific is published, and the one document that addresses data expressly removes the platform from its own scope. The privacy policy governs the marketing website. Its data security section is generic, committing to reasonable safeguards with no certification, no controls described and nothing specific to the platform. The third parties it names are website marketing and analytics vendors, being ZoomInfo, Google Analytics and Google and Facebook advertising; those are not platform subprocessors and are not treated here as though they were, because doing so would credit a website advertising stack as though it disclosed who processes a firm's pleadings. For the platform itself the policy states that processing of customer legal documents is governed by the service agreements and not by the policy, and those agreements are not published. One conditional survives the carve-out: where LegalMation is permitted to process such information for its own purposes it will do so under the policy, and the permission itself sits in the unpublished agreement, so a buyer cannot establish what is permitted. Consequently nothing published states how uploaded pleadings and discovery are stored, segregated, encrypted, retained or disposed of, whether any third party processes them, or what happens to the work samples and company data a customer submits to train its agent. Checked 7 September 2026.
Real published practices with the contractual detail withheld, which is the B band. The security page describes tenant isolation, staff access only with a support ticket and client approval, encryption at rest and in transit, enforced two-factor authentication, exportable logs of every access, query and export with one-year retention by default, and HIPAA business associate agreements; the privacy policy says the subprocessor list is available from the account team. Deletion and retention of customer content are left to the unpublished customer agreement, and the Trust Center's answer on incident notification is collapsed and could not be read. The privacy page carries two different policy texts, one covering the platform and one covering the website only. Verified 18 September 2026.
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 affirmative recourse is established, and the agreement that would carry it is not published. The website terms are the only published instrument and they run one way. All warranties are disclaimed in full, expressly including any representation that the website, its content or any services obtained through it will be accurate, reliable or error-free. Liability is excluded for damages of any kind under any legal theory, including direct damages. The customer indemnifies LegalMation for claims arising out of its use of the website, and no indemnity runs the other way. Claims must be brought within one year or are permanently barred, and a class action waiver applies. Two provisions do sit on the customer's side and are recorded so the grade is not read as absolute. A liability cap addressed expressly to customers limits LegalMation's exposure to fees paid in the preceding twelve months, and the licence to use generated draft documents survives termination, so a firm keeps the right to use work product it has already filed. Neither is recourse: a cap limits the vendor's exposure rather than granting the buyer a remedy, and a surviving output licence protects continuity rather than compensating for a defective draft. Nothing anywhere addresses what a firm's remedy is when a generated answer misstates a defence or a discovery response is wrong, and the service agreements that might address it are unpublished. Checked 7 September 2026.
No published position on what the vendor stands behind, which is the D band. The only published terms govern the website, cap liability for the site at $100, and say the platform is governed by separate customer agreements, which are not published. Nothing located addresses warranties, indemnities, caps or remedies for the platform or its AI output. Verified 18 September 2026.
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.
No integration with any practice system was located on any published surface. The site navigation is complete and short, covering the eight platform modules, three case study pages, news, contact, about and careers, and it contains no integrations page, no partner listing and no developer or API documentation. No document management system, practice management system, case management system, billing or e-billing platform, e-discovery platform or court filing system is named anywhere, and no application programming interface is mentioned in the website terms or on any product page. The product is described as browser-based and self-contained: a user uploads a document through the web interface and downloads the responsive documents. Two adjacent facts are named and set aside. The Data Analytics module is described as merging complaint extraction with an organisation's own time, billing or e-billing records and settlement outcomes, which implies data arriving from a billing system without naming one or describing how it gets there. And a third-party webinar describes handing off from an e-filing notice tool to LegalMation, which is a third party's account of a workflow rather than a vendor-published integration. Checked 7 September 2026.
Integration is claimed without named systems, which is the C band. Press material says the platform plugs into existing case files, and a forward-deployed team stands it up inside each firm; no document management, case management, e-billing or carrier system is named on the surfaces read, and PACER and docket monitoring are listed as coming soon. Verified 18 September 2026.
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.
Neither limb is stated, which is what the band requires. The service is plainly hosted software, delivered through a browser with the application at a subdomain and no installation required, but that establishes the delivery method rather than either limb this axis grades. No hosting provider is named anywhere. No region, data centre location or residency option is published, and no commitment restricts where uploaded pleadings and discovery material sit. Tenancy is equally absent: nothing states whether the service is single-tenant, pooled or configurable, and nothing describes separation between customers, which matters more than usual on a platform that ingests one organisation's precedent bank to generate documents for that organisation. The nearest geographic statement is a restriction on users rather than a commitment about data: the website terms state the company is based in California and that the website is provided for use only by persons located in the United States, with access from outside undertaken at the user's own initiative. That tells a buyer who may use the service, not where their matter documents rest. No self-hosted or private deployment option was located. Checked 7 September 2026.
The host is named without regions or options, which is the C band. The security page says the application and data run in dedicated Microsoft Azure environments with tenant isolation; the Trust Center's answer on where data is stored is collapsed and could not be read, and no region, residency option or processing location for AI calls is published. Verified 18 September 2026.
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.
No certification and no described controls were located. The site navigation contains no security page, no trust centre and no compliance page, and the full page inventory was established across several reads, so this is a finding about what the vendor publishes rather than a limit on the researcher. The only security material anywhere is the data security section of the privacy policy, which commits in general terms to reasonable safeguards and stops there: no certification is claimed, no control is described, no encryption standard, access control, testing regime or incident process is named, and no auditor, report or portal exists. No SOC 2, ISO 27001 or equivalent attestation appears. Nothing unsupported is badged either, so there is no misrepresentation to record; the position is simply that a buyer evaluating a platform that ingests live pleadings and an organisation's entire precedent bank has no published security information to assess. The absence is more consequential here than the letter alone conveys, because the privacy policy that carries that single generic sentence also expressly excludes the platform from its scope, so even the generic commitment does not clearly reach the matter documents. Checked 7 September 2026.
A named standard with a readable trust centre, short of dated reports, which is the B band. The Trust Center at trust.turbolaw.ai lists SOC 2 Type II and HIPAA as compliant, names Prescient Security, sets out control families and more than 150 documents available on request, and the security page says SOC 2 documentation is available under NDA. The security page describes controls as aligned to the Trust Services Criteria and continuously monitored; no audit period, report date or auditor opinion is published. Verified 18 September 2026.
Model Supply Chain Disclosure
Which models sit underneath, whose they are, where they run, and whether the vendor commits to telling customers when that changes. A legal buyer inherits every dependency it cannot see.
Nothing about the models is disclosed. No model is named, no version or family is given, no provider is identified, and no statement says whether the models are built in house, adapted from open weights, or licensed and run on the vendor's infrastructure. No subprocessor list exists, no data processing agreement is published, and no trust centre or portal offers one. The closest published material is a vendor announcement describing the technique stack as advanced natural language processing and machine learning techniques including generative AI, which names a category of method rather than a supply chain: it tells a reader what kind of system is at work and nothing about whose. The website terms acknowledge service providers only in passing, in a hold-harmless clause referring to LegalMation's affiliates, licensees and service providers, naming none. The gap is material on this record because the platform ingests an organisation's historical work product to train a customer-specific agent, so the question of which party's infrastructure that training runs on is precisely what a buyer would need answered, and no published surface addresses it. Checked 7 September 2026.
The model architecture is described without naming providers, which is the B band. The security page says Turbo Law routes each workflow to the model best suited to it and changes models as the state of the art changes, with no single-vendor dependency, on dedicated Azure infrastructure; the privacy policy says third-party LLM providers are contractually barred from training on customer data. No model or provider is named, and no notice of model changes is committed. Verified 18 September 2026.
Commercial Transparency
Whether a buyer can learn what this costs without entering a sales process: published rates, the unit being charged, what sits behind an enterprise tier, and what implementation adds.
No pricing information is published at any level, including the unit of charge. The full page inventory was established and the navigation carries the eight platform modules, three case studies, news, contact, about and careers; there is no pricing page, and every commercial route on the site resolves to Request a call or Contact us. Nothing states whether the product is charged per matter, per document, per seat, per module or by subscription, and no figure, band, tier or term appears anywhere. The website terms confirm that pricing exists and is set elsewhere, reserving the right to modify prices and undertaking to give notice before a customer incurs charges under a new pricing scheme, with any change applying only going forward. That is a change-notification commitment rather than a disclosure of price, and it points to separately executed agreements for payment of invoices which are not published. The one commercially framed figure the vendor does publish is a return claim rather than a price: clients achieving up to ten times return. A buyer can learn what the product is claimed to be worth and nothing about what it costs. Checked 7 September 2026.
Nothing about pricing is published, which is the D band. The site has no pricing page and routes every buyer to a demo request; neither the unit, the structure nor any figure is stated. Verified 18 September 2026.
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.
Three buyer segments are addressed distinctly and each has its own case study page, which is more structure than most records carry. The homepage names them and characterises each: corporate legal, described as in-house legal departments of all sizes; law firms, described as from AmLaw firms to litigation boutiques; and insurance carriers, described as ideal for staff counsel operations. That third segment is the distinguishing one, since insurance staff counsel is a buyer few vendors in this corpus address directly. Practice depth runs to the matter types the modules cover: complaints and petitions, written discovery, subpoenas, demand letters including EEOC complaints, and depositions, all on the defence side of high-volume litigation and claims. Jurisdictional handling is a real coverage claim rather than a boast: documents are described as tailored to the jurisdiction in which the action was filed, and the terms frame the product as assisting attorneys practising under their respective jurisdictions' procedures and formatting requirements. What holds this at B is the narrowness beyond that. The product is defence-side only, United States only by express restriction, and organised by document type rather than by practice area, so no substantive area beyond employment is addressed and no coverage statement describes which matter types are supported in which states. Checked 7 September 2026.
Segments and matter types are set out with substance, short of stated limits on the AI, which is the B band. Turbo Law sells to defence law firms and insurance carriers and names medical malpractice, mass tort, toxic tort, transportation, mergers and acquisitions, business litigation, employment, construction and real estate, and says it builds one litigation vertical at a time; drafting follows the jurisdiction's local rules and case-law search covers federal and state courts. Where the models are weaker is not stated. Verified 18 September 2026.
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?
The governing instrument is not published, and the vendor's own privacy policy says so expressly rather than leaving it to inference. The policy, effective 1 July 2024, states in its introduction that it does not apply to information collected from legal documents provided by customers relating to their use of the LegalMation platform, and a dedicated section repeats the point for pleadings, discovery requests and responses, subpoenas and demand letters, stating that their processing is governed by the terms of the service agreements with customers and not by the policy.
Those service agreements are not published anywhere on the site. One conditional survives the carve-out and is worth stating precisely because it is the only bridge back: where LegalMation is permitted to process such information for its own purposes it will do so under the policy, and the permission itself lives in the unpublished agreement, so a buyer cannot establish whether any such permission exists or what it covers.
The question matters more on this record than on most. The platform's design involves the customer submitting work samples and company data to train its agent, and generated documents are drawn from the organization's own historical and best responses, so customer content is unambiguously used to shape model behavior for that customer. What no published surface establishes is whether anything derived from it reaches models used for anyone else. This is recorded as an express referral to an unpublished contract rather than as vendor silence.
The privacy policy and security page say customer content is never used to train shared, public or foundation models, by Turbo Law or any model vendor, and that any fine-tuning on customer content happens only at the customer's direction. The security page says this is enforced in contracts, but the customer agreement is not published, so no matching term was located.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
Nothing published addresses retention of the documents submitted, the drafts generated, or anything in between, and the instrument that would ordinarily carry it declines the subject. The privacy policy expressly excludes customer legal documents from its scope and routes them to unpublished service agreements, so its retention treatment does not reach the pleadings, discovery, subpoenas and demand letters that are the platform's entire input and output.
The website terms address the website rather than the platform and set no period for anything. Nothing states how long an uploaded complaint persists, whether a generated draft answer is retained after download, what becomes of the work samples and company data a customer submits to train its agent, or whether any of it survives termination of the subscription. The one durable statement in the published estate runs the other way and concerns rights rather than retention: the license to use generated draft documents survives termination, so a firm keeps the right to use output it has already downloaded.
That protects the customer's continuity; it says nothing about what LegalMation keeps. Surfaces read on 7 September 2026: the privacy policy in full, the website terms in full, the homepage and full navigation, and the Complaint Response module page.
Retention is mentioned without a period for customer content or AI inputs and outputs. The one-year default sits under the audit-log heading on the security page, and the privacy policy says retention and deletion of customer content are governed by the unpublished customer agreement.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
Nothing published describes segregation of any kind. No surface addresses how one customer's material is separated from another's, how matters are partitioned within a customer's account, whether users can be scoped to particular matters, or whether a conflicted fee earner can be excluded. No access control model, permission structure or administrative capability is described anywhere, and the privacy policy's exclusion of customer legal documents means even its general handling language does not reach the matter material.
The only confidentiality obligation in the published estate runs the other way, requiring the user to keep their own username and password confidential and not to share account access. The absence is worth stating against the product's design rather than in the abstract: the platform ingests an organization's precedent bank and generates documents from it, so separation between customers is the property a buyer would most want described, and separation within a customer matters for a defense firm acting for multiple insurers on related claims.
Surfaces read on 7 September 2026: the website terms in full, the privacy policy in full, the homepage and full navigation, and the Complaint Response module page.
Tenant isolation between customers is claimed; walls within a customer are not yet available. The security page lists firm, practice-group and matter-level ethical walls under a coming-soon heading, and one version of the text says they are on the way. Roadmap items are not graded.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
Disclosure is addressed in the published terms, no notice accompanies it, and the customer is asked to waive claims arising from it. The website terms reserve to LegalMation the right to take appropriate legal action including referral to law enforcement, and to fully cooperate with any law enforcement authorities or court order requesting or directing it to disclose the identity or other information of anyone posting materials on or through the website or engaging in unlawful activity through it.
The same clause has the user agree to waive and hold LegalMation and its affiliates, licensees and service providers harmless from any claims resulting from action taken during or as a consequence of investigations by those parties or by law enforcement. There is no undertaking to notify, no commitment to resist or narrow a demand, no minimization term and no record-keeping obligation, and no transparency report is published.
One scope limit belongs on the record and is not a mitigation so much as a widening of the gap: that clause is framed around the website and materials posted through it, while the customer's pleadings and discovery are excluded from the published estate altogether and routed to unpublished service agreements. So for the matter documents a firm most needs protected, the published position is not a weak commitment but no commitment at all.
The privacy policy says information may be disclosed to comply with law or protect rights, with no commitment to notify the customer first. The customer agreement, which may address this, is not published.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
Two questions arise and the applicable one is unanswered. There is no external legal corpus behind this product: generated answers and affirmative defences are drawn from the customer's own historical and best responses, so the reference material is the organization's precedent bank rather than a licensed case law or forms library, and the sourcing and licensing limb this signal was written for does not bite. The limb that does apply is the provenance of the models, and nothing addresses it.
No published surface states what any underlying model was trained on, whether it was built in house or adapted, or on what rights it rests. A vendor announcement describes the technique stack as natural language processing and machine learning including generative AI, which names a method category rather than a provenance. The customer-precedent design raises a further provenance question that is also unanswered: whether material submitted by one organization to train its agent informs anything used for another.
That question is governed by the unpublished service agreements and is recorded on the training signal. Surfaces read on 7 September 2026: the privacy policy, the website terms, the homepage, the Complaint Response page and the vendor's demand letter announcement.
Searched the product overview, security page, Trust Center, privacy policy and newsroom on 18 September 2026. Case-law search covers federal and state courts and settlement modeling uses jurisdictional outcomes, but the sources and licenses behind them are not named.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
No citator ships and no good-law claim is made, and on this product the vendor goes further and disclaims currency expressly. The website terms state that because LegalMation is not a law firm it cannot guarantee that all the information it provides is current, and that because the law varies in different jurisdictions it is important to ensure that any information or document provided is correct for the user's specific jurisdiction.
That is a published statement that the vendor does not verify whether the legal content of its output remains good law, which is more informative than silence and is the reason the position is recorded here rather than passed over. The product class makes the point live rather than academic: this platform drafts answers, affirmative defences, discovery responses and subpoena responses tailored to the jurisdiction in which an action was filed, so the output asserts legal positions for filing.
The gap is bridged, if at all, by the contractual requirement that a licensed attorney in the forum state review and sign the output, which is graded on the professional responsibility axis and is an allocation of the verification duty rather than a verification capability.
Searched the same surfaces on 18 September 2026. Case-law search returns citations to holdings, but no check of whether cited authority remains good law is described.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
Nothing published describes how the system behaves when it is unsure. No confidence score, likelihood indicator or reliability signal is described as surfaced to a reviewer, nothing states when the system declines to draft a response to an allegation or flags one as weak, and nothing addresses a pleading it cannot parse or an allegation for which the organization's precedent bank holds no comparable response. Two features come nearest and neither answers the question.
The point-and-click interface presents chosen answers for each allegation, which is a selection surface applying uniformly rather than a behavior that varies with the model's own uncertainty. And the vendor's website terms acknowledge fallibility in general terms, disclaiming any warranty of accuracy and stating that the information may not be current, which admits that error occurs without describing what the system does when it detects the possibility.
On a product that drafts every response to every allegation in a complaint, the absence of any published abstention or flagging behavior means a reviewer receives a complete draft with nothing marking where the model was least certain. Surfaces read on 7 September 2026: the Complaint Response page, the homepage and full navigation, the website terms in full and the privacy policy in full.
The vendor states a behavior rule: fact-based outputs are line-cited to the file or the public record, and a claim that cannot be cited is not made. No confidence signal shown to the user is described.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
No matter naming LegalMation or LTLW, Inc. was located in the hallucination case tracking maintained by Damien Charlotin or in the sanctions reporting drawn from it, searched on 7 September 2026 on both the trading name and the second corporate name appearing in the vendor's published terms. The tracked corpus is large and the reporting reviewed names the tools involved where they are known, including the rare instances tied to purpose-built legal products rather than general chatbots.
LegalMation appears in none of it. The product class is worth stating plainly rather than treating the result as reassurance: this platform drafts answers, affirmative defences, discovery responses and subpoena responses that are filed in court, at a stated volume of over 100,000 matters per year, so it sits closer to the exposure this signal tracks than most records in the corpus. What separates it from the sanctioned pattern is that the output is drawn from the customer's own precedent rather than generated citations to authority, and that the terms require attorney review and signature in the forum state. Recorded as none located rather than as a positive finding about the vendor.
Searched the AI Hallucination Cases database maintained by Damien Charlotin and trade press reporting on 18 September 2026 for court records addressing fabricated or hallucinated content in output from Turbo Law. None located. This signal does not record litigation history of any other kind.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Professional obligation is engaged in real terms and no source of guidance is named. What the published terms do is unusually concrete for this corpus: they restrict the license to licensed attorneys or those working on their behalf, require that all output be reviewed and ultimately signed by a licensed attorney in the state where the matter is pending, state that LegalMation does not engage in the practice of law and is not a law firm, and warn that the user must ensure any document is correct for their specific jurisdiction.
That framework tracks the substance of a lawyer's supervision and competence duties closely, which is why this sits above the floor. What it does not do is point to any authority. No bar association, state bar opinion, model rule, court standing order or regulator publication on the use of artificial intelligence is cited, referenced or mapped, and nothing is broken down by jurisdiction despite the product being sold across United States practice and generating documents tailored to the forum.
A firm looking for help aligning its own AI use to its regulator's expectations finds a well-drafted allocation of duty rather than a route to the guidance that defines it.
Searched the product overview, security page, Trust Center, Terms of Use, privacy policy and newsroom on 18 September 2026. No ethics opinion, court AI order or rule, or bar guidance is referred to.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
Savings are the vendor's central published claim and the fee consequence is nowhere addressed. The claims are prominent and quantified in form: reduce your team's time by 80 percent so there is no need to hire when volume surges, clients achieving up to ten times return, over 100,000 matters processed per year, and a forecast to save clients more than 100 million dollars this year. The stated purpose is to shift resources, refocusing lawyers and paralegals on higher-value work and controlling cost as volume rises.
Nothing addresses what happens to a client bill when drafting an answer takes minutes rather than hours, no per matter record of AI-assisted work is described, and no guidance on fee or disclosure treatment is offered. Two qualifications belong on the record. The buyer base is unusual for this signal: corporate legal departments and insurance staff counsel are largely absorbing their own cost rather than billing a client, so for two of the three named segments the compression this signal was written for does not translate directly into a client bill.
And the Data Analytics module is described as merging complaint extraction with an organization's own time, billing or e-billing records and settlement outcomes, so billing data flows into the platform without anything published about AI-assisted work flowing back out to a bill.
Billing economics are marketed without disclosure guidance. The vendor says firms report 60 percent fewer non-billable hours per matter, 10 percent fewer write-offs and profitability on alternative fee arrangements, and generates carrier-ready status reports; nothing addresses how firms should bill or disclose AI-assisted work to clients or carriers.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
A firm cannot answer a client's AI clause from anything published. No subprocessor list exists on any surface, no data processing agreement is published, and there is no trust center or portal from which either could be requested. No model provider is named, and no statement identifies whether any third party processes matter content at all. The privacy policy names third parties, and they must not be mistaken for the answer: ZoomInfo, Google Analytics and Google and Facebook advertising are marketing and analytics vendors serving the website, and the same policy expressly excludes the platform and the customer legal documents processed through it from its own scope.
Naming a website advertising stack discloses nothing about who handles a firm's pleadings, and it is not credited here. The instrument that would carry the answer is named and withheld: the policy routes processing of customer pleadings, discovery, subpoenas and demand letters to the service agreements with customers, which are not published. No forwardable client-facing material of any kind was located. A firm asked by a client which systems and providers touch its litigation documents could report only that the vendor's own privacy policy declines to say and points to a contract the vendor does not publish.
Disclosure material is available on request. The privacy policy says business customers can get the subprocessor list from their account representative, the security page says Turbo Law answers security questionnaires as part of deployment and signs BAAs, and the Trust Center offers more than 150 documents on request. No list is published.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Nothing published supports disclosing AI involvement, and this is the record in the pull where that absence is most consequential. The platform's entire output is court-facing: answers to complaints, affirmative defences, responses to written discovery, subpoena responses and demand letter responses, produced at a stated volume of over 100,000 matters per year and described as tailored to the jurisdiction in which the action was filed.
Nothing states whether a generated document carries any marker of its origin, whether a record distinguishes text the model drafted from text the reviewer wrote, whether any log of the drafting session is retained or exportable, or what a firm should say if a court asks how a filing was prepared. Vendor material runs the other way on distinguishability, describing the system as able to match the style and content of an organization's documents so that output resembles that generated by the firm's own attorneys.
The one relevant published commitment is the contractual requirement that a licensed attorney in the forum state review and sign any output, which places responsibility for the filing squarely on the lawyer and is graded on the professional responsibility axis; it is an allocation of duty rather than a record a lawyer could produce. Surfaces read on 7 September 2026: the website terms in full, the privacy policy in full, the homepage, the Complaint Response page and the vendor's demand letter announcement.
Parts of a record exist. Every access, query and export is logged and exportable, and outputs carry line citations to their sources; no record of which model produced an output or who verified it is described.
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.
- AI Liability and Recourse
- Commercial Transparency
- Primary Law Corpus Provenance
- Good Law Verification
Which one fits
Choose LegalMation if
- You want drafts built from your own past pleadings. LegalMation reads a complaint and drafts each allegation response, the affirmative defenses and the full answer from your organization's historical and best responses, tailored to the jurisdiction where the action was filed, with a user choosing or reviewing each answer.
- You want the vendor's terms to make the supervision rule explicit. LegalMation licenses the platform only to licensed attorneys or those working for them, requires every output to be reviewed and signed by a licensed attorney in the state where the matter is pending, and states that it does not practice law.
- You want evidence from organizations at your scale. LegalMation names Walmart, Ogletree Deakins, Fisher Phillips and American Integrity Insurance Group with attributed senior staff, and covers complaints, discovery, subpoenas, EEOC demand letters and depositions across more than 100,000 matters a year.
Choose Turbo Law if
- You need a firm statement on training and privilege before loading claim files. Turbo Law's privacy policy says customer content does not train foundation models, fine tuning happens only at the customer's direction, third party model providers are contractually barred from training, and Turbo Law neither asserts nor waives privilege for the customer.
- You need each client's matters isolated and every access logged. Turbo Law runs in dedicated Microsoft Azure environments with tenant isolation its own staff cannot cross without a support ticket and client approval, logs every access, query and export in exportable form, and signs HIPAA business associate agreements.
- You want the high stakes outputs held for an attorney. Turbo Law states that case evaluations, settlement recommendations and motion drafts go to attorney review before anything leaves the firm, that each fact based output is line cited or not made, and that workflows must meet accuracy standards before they ship.
In summary
LegalMation
LegalMation, of Los Angeles, launched in 2018, automates high volume litigation and claims defense for corporate legal departments, law firms and insurance staff counsel, through eight modules covering complaint response, discovery and subpoena responses, EEOC demand letters, depositions, case summaries, matter profiling and analytics. It drafts from an organization's own historical responses, tailored to the jurisdiction of filing. The AI Legal Index grades it in the top two bands on five of fifteen capability axes, with A grades on AI centrality and professional responsibility posture: its terms require a licensed attorney in the forum state to review and sign all output. It names Walmart and Ogletree Deakins among users. As of 7 September 2026 the index located no platform agreement, no security certification and no named model provider.
Turbo Law
Turbo Law, of San Mateo, California, is an AI litigation platform for defense firms and insurance carriers in medical malpractice, mass and toxic tort, transportation and business disputes. It builds a line cited fact graph from the matter file and produces chronologies, contradiction reports, privilege flags, motions, discovery responses, case evaluations and settlement value models. The AI Legal Index grades it in the top two bands on ten of fifteen capability axes, with A grades on AI centrality and on autonomy and oversight, since high stakes outputs go to attorney review before leaving the firm. It states SOC 2 Type II and dedicated Azure environments, and that customer content does not train foundation models. As of 18 September 2026 the index located no named customer, platform agreement or price.
Questions buyers ask
LegalMation vs Turbo Law: which is better for insurance defense?
On published evidence Turbo Law sits in the top two bands on ten of fifteen AI Legal Index capability axes and LegalMation on five of fifteen, mostly because Turbo Law publishes its data handling, security and oversight commitments. LegalMation publishes stronger customer evidence and the clearest attorney review rule, and drafts from a firm's own past responses. Carriers whose security review comes first have more to read from Turbo Law.
Does LegalMation publish how it handles uploaded pleadings?
No. LegalMation's privacy policy, effective 1 July 2024, states that it does not apply to legal documents customers provide through the platform, including pleadings, discovery, subpoenas and demand letters, and that their processing is governed by service agreements with customers, which are not published. No security page or certification was located. Customers submit work samples to train their own agent, and what happens to that material is not stated. Graded by AI Legal Index against 15 capability axes and 12 legal signals, including privilege handling and citation accuracy, from each vendor's own published materials, verified September 25, 2026. No vendor pays for placement.
Does Turbo Law train AI on client files?
Turbo Law's privacy policy says it does not use customer content to train or improve foundation models, that any fine tuning happens only at the customer's direction, and that third party model providers are contractually barred from training on customer data. The security page says the commitment is enforced in contracts, though the customer agreement itself is not published. LegalMation's published documents do not reach the question. Graded by AI Legal Index against 15 capability axes and 12 legal signals, including privilege handling and citation accuracy, from each vendor's own published materials, verified September 25, 2026. No vendor pays for placement.
Which one keeps an attorney in the loop?
Both, in different ways. LegalMation's terms require a licensed attorney in the state where the matter is pending to review and sign every output, and its interface has users choose or review each allegation response. Turbo Law states that case evaluations, settlement recommendations and motion drafts surface for attorney review before anything goes out, summed up as the platform proposes, the team verifies and partners decide. Graded by AI Legal Index against 15 capability axes and 12 legal signals, including privilege handling and citation accuracy, from each vendor's own published materials, verified September 25, 2026. No vendor pays for placement.
What do LegalMation and Turbo Law both leave unpublished?
The platform contract and the price. Neither publishes the customer agreement that governs its platform, so neither states a warranty, indemnity or remedy for a wrong draft, and neither publishes a price or unit of charge. Neither names the AI models it uses or the case, document or claims systems it connects to. Neither names bar guidance on AI, or says how saved drafting time should appear on a bill to a client or carrier. Graded by AI Legal Index against 15 capability axes and 12 legal signals, including privilege handling and citation accuracy, from each vendor's own published materials, verified September 25, 2026. No vendor pays for placement.
Three readings to weigh. Neither vendor publishes the customer agreement that governs its platform: LegalMation's privacy policy expressly excludes customer legal documents and points to unpublished service agreements, and Turbo Law's published terms cover its website only, so neither states what it stands behind when an AI draft is wrong. Turbo Law's commitments on training and privilege are published policy, which it says are enforced in contracts that cannot be read. Turbo Law's figures for fewer non billable hours and write offs carry no customer names or method. LegalMation was verified on 7 September 2026 and Turbo Law on 18 September 2026. Neither vendor reviewed this page.
Neither vendor paid for inclusion, placement or a grade, and neither reviewed this page before it published. Everything above comes from public material on the dates shown. How the index grades.