A
Anytime AI
Anytime AI is an agentic AI platform sold exclusively to plaintiff law firms, covering a case from intake through settlement. Its centre of gravity is the medical record: it turns thousands of pages into structured chronologies and case overviews that set out injuries, treatments, diagnostic findings, provider actions and the progression of care, flags treatment gaps and deviations from the standard of care, and carries the result through into demand letters that are case-specific, medically grounded and built around itemised damages.
Around that sit agentic review of high-volume discovery and drafting of discovery responses, summarisation of depositions, records and filings, generation of motions, briefs and case memos, and legal research. An agentic chat assistant branded Talk to Teddy answers questions across the matter file and, on the vendor's account, answers only from the documents in that matter, tying every response to a cited source and saying so when the record does not contain an answer.
Five practice areas have pages of their own: personal injury, nursing home litigation, medical malpractice, trucking accidents and traumatic brain injury. The platform is described as a closed, private-cloud environment with AES-256 encryption at rest, TLS 1.2 or later in transit, role-based access controls, audit logging, encrypted backups and a stated policy that customer data is never used to train models. It connects two-way to SmartAdvocate and to Litify, and to SharePoint, Dropbox, OneDrive, Google Drive and Box.
The product was shaped with four named trial attorneys in nursing home, medical malpractice and personal injury work, and the company publishes a Master Software Subscription Agreement and an eighteen-page data privacy and security policy. Anytime AI is based in White Plains, New York.
Capability grades
All 15 axes, graded from public sources on the date shown. 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 models are the product and there is nothing underneath them to sell, which is the A band. Every published capability is model output: turning medical records into structured chronologies and overviews, identifying deviations from the standard of care, drafting demand letters from the resulting record, reviewing discovery, summarising depositions and filings, drafting motions and memos, and answering questions across a matter through the Talk to Teddy assistant.
There is no repository, case management system, calendar or billing layer being licensed underneath. The point is settled by the vendor's own integration strategy rather than inferred: it connects two-way to SmartAdvocate and to Litify, both of them case management systems already in this index, and to SharePoint, Dropbox, OneDrive, Google Drive and Box for storage. It plugs into the systems a firm already runs rather than replacing them, which is what a pure intelligence layer does.
The commercial framing matches, the company describing itself as an all-in-one agentic AI platform for plaintiff lawyers. This grade says nothing about how much is disclosed about those models; on that, see the Model Supply Chain row, which sits far lower. Verified 12 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.
Grounding is real, documented and described down to the technique, short of any published measurement, which is B. The vendor names its method rather than gesturing at it: hallucination is addressed by combining information extraction with retrieval-augmented generation, and the assistant is described as a closed system that answers using only the documents inside a specific matter, with every response tied to a cited source in the case file.
The vendor puts that as a rule, no citation, no answer, and states that answers do not come from a general training corpus but from the medical records, deposition transcripts, discovery responses, photographs and police reports in the matter. Output carries the citations through: chronologies are produced in timeline format with citations, and demand letters are described as citation-ready with integrated treatment summaries.
That is a described retrieval method with linked sources, which is what B asks for. What holds it off A is measurement. No accuracy figure, error rate, test set, sample or third-party evaluation is published anywhere, and no failure mode is named. R15 applies to the authority limbs: the product's citations point to the customer's own case file rather than to legal authority, so citator-style checking does not bite, and the legal research tool's sources are not identified. Verified 12 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.
Autonomy is claimed and oversight is asserted without a mechanism, which is C. The autonomy claims are strong and repeated: agentic review of high-volume discovery, agentic workflows across the case lifecycle, an assistant that in the vendor's words thinks through facts, timelines and legal theories rather than merely summarising, and a platform that lets firms run more matters at once without adding headcount. The oversight side is a posture rather than a control.
The vendor says the platform works alongside a firm's attorneys through the full lifecycle of every case, and its practice-area material describes output as professional documents that attorneys can refine and use immediately, which positions a human as the reviewer. Nothing published turns that into a structure. No mode is described, no threshold is stated at which the system proceeds or stops, no approval step or review surface is named, no escalation route is set out, and nothing distinguishes a task the platform completes from one it proposes.
The nearest thing to a boundary is the grounding rule graded on the row above, that the assistant answers only from the matter file and says so when the record does not support an answer, which constrains what it will assert rather than who checks it. Verified 12 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.
Named customers in unusual depth, without an outcome carrying a method, which is B. The volume and specificity of attribution is the strongest in this lane: a logo wall of nine firms including Brown Chiari, Lomurro Law Trial Lawyers, Mandell Boisclair & Mandell, Truck Wreck Justice, Ernst Law Group, Romano Law Group, Tosh Law Firm, Clement Law Group and Garcia & Coman, and roughly eighteen testimonials every one of which carries a named attorney, a title and a firm, from Stephen Garcia at Garcia & Coman to Blake Dickson at The Dickson Firm and Carma Henson at Henson Fuerst.
Several describe the work in domain terms a practitioner would recognise, MARs, TARs, ADLs and nursing home flow sheets. One testimonial contains a figure, Ernest Tosh of Tosh Law Firm describing eighty per cent of the time freed on some of his tasks, and one describes a validation exercise, Stephen Garcia running the platform against a past case with results consistent with the firm's own work. Neither carries a method, a period or a dated artifact behind it, so neither reaches the A limb; a vendor-reported figure inside a quotation is a claim rather than a measurement.
Two things are recorded and deliberately not credited: four named attorneys published as product co-designers, two of whom also appear as customers, and eight state trial-lawyer association partnerships, which are sponsorships rather than deployments. Verified 12 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.
Substantive published commitments short of the full picture, and this record reaches the limb that R33 makes decisive and that most of the corpus misses. Privilege and work product are addressed head-on rather than folded into a general confidentiality claim: the home page carries a block headed Attorney-Client Privilege stating that the architecture ensures no third party can access privileged communications or case strategy, and the FAQ answers the question directly, describing end-to-end encryption and firm-held encryption keys and stating that only the firm can access and decrypt its data and that Anytime AI cannot view client information, case strategy or communications.
Training is addressed as a flat policy commitment, that case files, medical records and legal work product are never used to train any AI model. Access is documented: role-based permissions, a need-to-know rule backed by contractual non-disclosure agreements, audit logging, and a closed private environment contrasted with public AI tools. Two of the five limbs are not met. Retention and deletion are not stated on any readable surface, and no position is published on what any third-party model provider may retain, no provider being named.
Recorded rather than resolved: the firm-held key claim sits against a platform that performs server-side agentic analysis of those same documents, and the eighteen-page data policy that would reconcile it could not be opened by this index. Verified 12 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.
The intended audience is unambiguous and the advice line is not addressed, which is where this lands. R15 governs which limbs bite. The audience limb is answered completely: the platform is sold exclusively to plaintiff law firms and attorneys, is not offered to consumers or claimants, and its practice-area material positions output as professional documents that attorneys refine and use, which frames the product as drafting support rather than advice.
The vendor also acknowledges the professional context in general terms, noting on its own blog that bar associations have issued guidance reminding lawyers that the duty of competence applies to AI tools, and warning that a general-purpose chatbot is not safe for substantive legal tasks. What is missing is everything that would make that a position. No statement was located that the vendor is not a law firm, that output does not constitute legal advice, or that a lawyer must exercise independent professional judgement before filing.
Nothing addresses supervision of machine-drafted pleadings, and no jurisdictional limit appears despite research being claimed across all fifty states. The published Master Software Subscription Agreement would ordinarily carry that language; it is dated 5 February 2026, is published ungated as a PDF, and could not be opened by this index, so the row would move on a reading. Verified 12 September 2026.
AI Governance and Bias Disclosure
Published governance over model behaviour: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.
A governance position is published without a mechanism, a testing regime or anything a buyer could audit, which is C. What exists is a set of stated commitments rather than a programme: a closed AI architecture contrasted with public tools, a zero data training policy stated repeatedly and unconditionally, encryption and role-based access as design principles, and a compliance wall that names NIST 800-171 and the Cloud Security Alliance alongside the security standards.
Those are real positions and they keep the row off the floor. None of what the higher bands require appears. No AI governance or responsible AI page exists, no AI-specific framework is named and no certification against one is claimed, nobody inside the company is identified as accountable for AI decisions, nothing describes what is evaluated before a model or feature ships, and no testing or red-teaming regime is set out.
Bias is addressed nowhere, and the gap is worth naming on this product rather than passing over: the platform reads medical records to identify deviations from the standard of care and to build damages narratives, and nothing published considers whether that analysis performs evenly across patient populations, care settings or record quality. Verified 12 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.
Protection is documented and the lifecycle is not, which is C. The protection half is specific and published: AES-256 encryption at rest and TLS 1.2 or later in transit, role-based access controls and permissions, a need-to-know rule backed by contractual non-disclosure agreements, audit logging, encrypted backups, continuous monitoring, and a closed private-cloud environment. For a platform holding protected health information that is a real and legible security posture, and the vendor sets it out as such rather than asserting enterprise-grade security and stopping.
The other half of the axis is absent from every readable surface. No retention period is stated for case files, prompts or generated documents; no deletion right or export-on-termination commitment is published; no subprocessor list exists anywhere and no processor is named; and no breach or incident notification commitment was located. The eighteen-page Data Privacy and Security Policy is published as an ungated PDF and is exactly the document that would carry all four, and this index's fetcher would not open it on the date shown.
Under R20 that is a reader limit rather than a finding about the vendor, and the row would move in either direction on a reading. Verified 12 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.
Nothing published on who bears the loss when the system is wrong was located, which is the D band, and the note must be read with the retrieval position stated in full. No warranty, indemnity, liability cap, exclusion, service credit or insurance position appears on any readable surface: not on the home page, not on the security page, not on the practice-area pages, and not in any published FAQ. Nothing addresses what happens if a medical chronology misses a treatment entry, if a demand letter misstates damages, or if a discovery response is wrong, on a product whose outputs go to insurers and into court.
The vendor does publish a Master Software Subscription Agreement, dated 5 February 2026 and hosted as an ungated PDF, and that is the document that would ordinarily carry every one of those terms. This index's fetcher would not open it: the direct fetch was refused and R8 step 4 recovered nothing beyond the link itself, the file not being held in the search index. **So this grade records what is established as located on the date and is not a finding that no such terms exist.** The row would move, potentially several bands, on a reading of that agreement, and it is named here so a future pass knows exactly which document to open first. Verified 12 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.
Real integrations, named and specific to the buyer's stack, short of the depth an implementer could work from. The named connections are the right ones for a plaintiff firm and two of them are checkable against this index: SmartAdvocate, described as a two-way integration, and Litify, both case management systems already indexed here. Storage integration covers the four platforms a firm is likely to hold records in, SharePoint, OneDrive, Google Drive, Dropbox and Box.
The vendor's framing is that a firm connects Anytime AI to the systems it already uses rather than migrating, which is consistent with the platform being an intelligence layer rather than a system of record. What holds it off A is that only one integration has its direction stated and none has its contents described. Nothing published sets out which objects sync, whether documents are pulled or pushed, whether generated work product returns to the case management system, what a firm must configure, or what permissions the connector requires.
No API, developer documentation, connector catalogue or integration page was located, and no authentication method is described. Verified 12 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.
Cloud delivery is stated with a tenancy posture and no region anywhere, which is C. The deployment side is described more than most: the platform is presented as a closed, private environment contrasted explicitly with public AI tools, built on secure private cloud infrastructure with encrypted backups and continuous monitoring, and reached through a hosted application at an enterprise subdomain. That tells a buyer the shape of the delivery, and the closed-environment claim is doing real work in the vendor's own positioning.
Residency is absent entirely. No cloud provider is named, no region or data centre is identified, no residency option is offered or refused, and nothing distinguishes where medical records are stored from where model inference happens. That last gap matters on this product more than on most: it processes protected health information at volume, and the vendor markets HIPAA compliance without stating where the processing occurs or who operates the infrastructure.
No on-premises or single-tenant option is described. The eighteen-page data policy is the document that would carry a processing-location statement, and it could not be opened by this index on the date shown. Verified 12 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.
Badges appear on the site with no scope, no date and no report available, which is C. The security page carries an eight-item wall: SOC 2 Type II, HIPAA, GDPR, FedRAMP, FIPS 140-2, NIST 800-171, CSA and PCI. Not one is accompanied by an auditor, an observation period, a certificate number, a scope statement or a route to obtain a report, and there is no trust centre. What lifts this above a pure badge wall is that a substantive artifact exists: an eighteen-page Data Privacy and Security Policy, published ungated as a PDF and referred to by the vendor as such.
It could not be opened by this index and is named here rather than credited. Two problems belong on the record rather than in the grade. FedRAMP is an authorisation status for cloud services sold to United States federal agencies, not a compliance posture a private vendor holds by assertion, and the badge carries nothing to establish which status is claimed. And ISO 27001 appears in a published vendor FAQ, listed among the standards met, while it is absent from the security page's own wall, so the estate does not agree with itself about what is held. Verified 12 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.
The vendor refers to its own models without identifying anything underneath them, which is C. The only supply-chain language located is possessive and unexplained: data is never used to train our models, the platform is a closed private environment unlike public AI tools, and hallucination is managed by combining information extraction with retrieval-augmented generation. Those describe an architecture and a technique.
No model is named, no version is given, no third-party provider is identified, no hosting arrangement for inference is stated, and nothing commits to notifying a customer if any of it changes. There is no subprocessor list on any surface. The closed-environment framing is the interesting part and cuts both ways: it is offered as the reason a firm can trust the platform with privileged material, and it is offered without any statement of whose models sit inside the closed environment, so the buyer is asked to accept the boundary without being told what is behind it.
On a product that sends medical records and case strategy to a model, that is the disclosure a firm's own client would most likely ask about. Verified 12 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, which is the D band. There is no pricing page in the navigation and none was located anywhere on the estate. No plan or tier name, no rate, no band, no minimum, no term, and no statement of whether the platform is licensed per seat, per firm, per matter or per page of records processed. The only commercial route published is a demonstration booking, and the sole contact channel is a sales address and a telephone number.
Under R10's closing discipline a page that only invites a sales conversation is an absence and belongs in this note alone, so no VendorPricing row is written for this record. One document is named rather than credited, because it is the obvious route by which this row would move: a Fulfillment Policy is published as an ungated PDF and linked in the site footer, and a fulfilment policy commonly carries billing cadence, refund and cancellation terms, any of which would lift this axis off the floor and, under R17, make a pricing row owed. This index's fetcher would not open it on the date shown. Verified 12 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.
Coverage is described with real substance and the boundaries are left open, which is B. The segment is stated with unusual precision and consistency: plaintiff firms only, and specifically those handling larger and more complex matters, with the vendor separately stating that it serves plaintiff firms of every size from solo through large. Practice coverage is enumerated across five pages of its own, personal injury, nursing home litigation, medical malpractice, trucking accident litigation and traumatic brain injury, and the depth behind each is evidenced in the customers' own vocabulary rather than asserted, with named attorneys describing medication and treatment administration records, activities of daily living and nursing home flow sheets.
The vendor's association partnerships across eight state trial-lawyer bodies corroborate the segment even though they are not deployment evidence. What is left open holds it off A. No jurisdictional limit is stated anywhere, against a claim that liability and statutes can be researched across all fifty states; nothing states which matter types or claim values the platform is not suitable for; and defence-side work is excluded by positioning without ever being addressed as a boundary. Verified 12 September 2026.
4 public documents
The public pages on file for Anytime AI, with the recorded signals each one supports and the date it was last read. Open any of them and check the reading against the record.
-
anytimeai.ai4 signals
Client Data in Training, Good Law Verification, Bar Guidance Alignment and 1 more
Read Sep 12, 2026
-
anytimeai.ai/security4 signals
Prompt and Output Retention, Ethical Walls and Matter Segregation, Third Party Request and Subpoena Notice and 1 more
Read Sep 12, 2026
-
Primary Law Corpus Provenance, Refusal and Uncertainty Behaviour
Read Sep 12, 2026
-
Court Disclosure Support
Read Sep 12, 2026
Legal Signals
What each signal meansA signal records what public sources say on the date shown. It is not a grade and it is not a recommendation. Where a signal reads Not addressed, it means the index did not locate the material in public sources on that date, which is a statement about disclosure rather than about the product.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
A public policy or trust page states no training on customer content, with no matching term located in the published agreement.
Public material states plainly that customer content is not used for training, and no matching term could be read in the published agreement. The policy claim is unconditional and repeated across three surfaces: a home-page block headed Zero Data Training; an FAQ answer stating that case files, medical records and legal work product are never used to train any AI models; and a security page stating that data is never used for training, ever.
A further FAQ adds that the vendor does not use or share data for model training and that staff operate under a need-to-know rule backed by contractual non-disclosure agreements. R43(1) was run rather than assumed: the vendor publishes a **Master Software Subscription Agreement dated 5 February 2026** as an ungated PDF, and it was sought. This index's fetcher would not open it, the direct fetch being refused and the file not held in any search index, so under R20 the position is a reader limit rather than a vendor absence.
That matters to the value because R122 makes three outcomes available on a reading. If the agreement permits de-identified or aggregated use with no stated bound, the value is contractual-permitted; if it permits and bounds that use, it is purpose-limited; the value written here holds only while the agreement's data clause remains unread, and a conflict between it and the marketing claim would be the finding.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
No located public material states how long prompts and outputs are retained.
No located public material states how long case files, prompts or generated documents are retained. The security estate is otherwise specific, covering AES-256 encryption at rest, TLS 1.2 or later in transit, role-based access controls, audit logging, encrypted backups and continuous monitoring, and it says nothing at all about lifecycle. Nothing addresses a retention period, a deletion right, export on termination, or whether the assistant's prompts and generated chronologies and demand letters are held separately from the underlying records.
The one reference to backups, that they are encrypted, describes protection rather than duration. Two published documents would ordinarily carry the answer and neither could be opened by this index on the date shown: the eighteen-page Data Privacy and Security Policy dated August 2025, and the Master Software Subscription Agreement dated 5 February 2026, both hosted as ungated PDFs and linked in the site footer. Under R20 that is a limit on this index's reader and not a finding about the vendor.
The question carries more weight here than on most records because the material held is protected health information belonging to the firm's clients rather than to the firm.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
The product maintains its own permission model, documented, requiring the firm to keep it aligned.
The product maintains its own documented permission model that the firm must administer, which is this value. The published elements are consistent across the security estate: role-based access controls and role-based permissions so that only authorised users can view sensitive information, a stated rule that only the firm's own authorised team members can reach its data, a need-to-know policy for vendor personnel backed by contractual non-disclosure agreements, and audit logging over access.
Around that sits an architectural claim that the environment is closed and private and that encryption keys are held by the firm, so that the vendor cannot read case strategy or client information. The alignment burden sits with the buyer, which is what places this value rather than the tier above: nothing enforces an external document management system's access model, and no default configuration is published. What is not addressed is the wall inside the firm.
Nothing published describes segregation between matters, screening of a conflicted user from a specific case file, or how the assistant's reach is bounded when a firm runs adverse or related matters, on a platform whose conflict-sensitive material is medical records belonging to individual claimants.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
No located term or policy addresses third party requests for customer data.
No located public material addresses what happens when a third party demands customer data. Nothing on any readable surface refers to subpoenas, warrants, court orders, law enforcement or government requests, and no commitment or reservation about notifying the customer appears. No transparency report exists. The two documents that would ordinarily carry the position are published as ungated PDFs and could not be opened by this index on the date shown: the eighteen-page Data Privacy and Security Policy and the Master Software Subscription Agreement dated 5 February 2026.
Under R20 that is a limit on this index's reader rather than a finding about the vendor, and this row would move on a reading. The question has particular force on this record. The custodian holds protected health information belonging to injured claimants and their treating providers, gathered for litigation, and a demand served on the platform rather than on the firm would reach material the firm's clients would expect the firm alone to control.
The vendor's architectural claim, that only the firm can decrypt its data, would be the natural place to address it and does not.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
Sources are identified without stating the licence or rights basis.
The working corpus is identified precisely and nothing behind it is, which is this value. On the customer's side the vendor is unusually explicit: the assistant does not draw on a general training corpus but on the documents inside the matter, itemised as medical records, deposition transcripts, discovery responses, photographs and police reports, and the platform is described as a closed system answering only from those files.
That is a clear statement of where an answer comes from and it is the reason the accuracy claims on this record hold up as well as they do. What is not identified is everything else. The platform also offers legal research, described as finding relevant case law, statutes and legal arguments and elsewhere as researching liability and statutes across all fifty states, and no source, database, publisher or licensing arrangement is named for any of it, nor any coverage table or update cadence.
Nothing states what the underlying models were trained on, and no model or provider is named anywhere. So a buyer is told what the product reads from its own file and nothing about what it knows before it opens one.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
No located public material addresses whether authority is checked for subsequent history.
No located public material addresses whether authority is checked for subsequent history. The product does publish a legal research capability, described as finding relevant case law, statutes and legal arguments with AI that understands the context of a specific matter, and elsewhere as researching liability and statutes across all fifty states in minutes. Nothing published says where those authorities come from or whether their current status is checked.
No citator relationship is named, no treatment taxonomy is described, and nothing states that the platform reports whether a decision has been reversed, vacated, overruled or superseded. The gap sits oddly against the rest of the record, because this vendor has thought carefully about grounding and states as a rule that every response ties back to a cited source in the case file. That rule is about the customer's own documents; it does not reach the currency of external authority, and the vendor's own comparison material notes that a competing plaintiff platform builds treatment status into its research while making no equivalent claim for itself. Product pages, the security estate and the published FAQs were read on the date shown.
Refusal and Uncertainty Behaviour
What does the product do when the answer is not in the corpus?
The vendor describes refusal or abstention behaviour in public materials.
The vendor describes abstention behaviour in public material, which is this value, and does so in operational terms rather than as a slogan. The worked example is specific: asked what an orthopaedist said about a client's range of motion at a six-month follow-up, the assistant locates the note in the relevant record and reports what is there, and if the note does not exist it says so instead of inventing something that fits.
The vendor states the design rule behind it as no citation, no answer, with every response tied to a cited source in the case file, and describes the system as closed, answering only from documents inside a specific matter. Together those describe what the product does at the edge of its grounding, which is what this signal asks and what most records in this corpus leave unstated. What is not present, and is the distance to the top value, is any demonstration.
No evaluation, benchmark, error rate or published example set shows the behaviour holding, no confidence or grounding score is exposed to the reviewer, and nothing describes what a user sees when an answer is declined.
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 addressing fabricated or hallucinated legal citations produced by this product has been located as of the date shown. This is a statement about the public record on that one subject, not a finding about the product, and this signal is not a litigation history.
Searched on 12 September 2026, on the product name with a plaintiff-litigation qualifier and on the corporate name, against published trackers of decisions on AI-generated fabricated citations including coverage of the Damien Charlotin AI Hallucination Cases database and two independent sanctions trackers, for any court order, opinion or disciplinary record addressing fabricated or hallucinated legal citations produced by this product.
None located. This is a statement about the public record on that one subject as of the date shown, and under R119 this signal records fabricated citations and nothing else, so it is not a litigation history. A note for a future reader: the vendor's name is close to ordinary English and a bare search returns the general sanctions literature rather than anything about this product, so every query needs a plaintiff or corporate qualifier to be meaningful.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Public materials refer to professional responsibility in general terms without naming guidance.
Public materials refer to professional responsibility in general terms without naming any guidance, which is this value. Two references were located and both are unattributed. A home-page block headed Compliance Ready states that the platform lets a firm meet bar ethics on confidentiality and third-party data sharing and use AI without exposure. And a published blog page notes that federal and state courts have sanctioned attorneys for filing briefs containing AI-fabricated citations, and that bar associations have issued guidance reminding lawyers that the duty of competence applies to AI tools.
Both engage the right subject matter, confidentiality under the duty to protect client information and competence in the use of technology, and neither names a single opinion, rule or issuing body. ABA Formal Opinion 512 does not appear anywhere on the estate, no state bar guidance is cited, and no rule of professional conduct is identified by number. Nothing maps any capability of the platform onto any specific duty.
The absence is worth recording rather than passing over, because the vendor plainly knows the terrain: it argues its case in the language of bar ethics and stops short of citing any.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
Public materials claim time savings without addressing billing or disclosure, and the product sits inside a fee relationship between a lawyer and a client where those savings would change the bill.
Time savings are claimed throughout and nothing addresses billing or disclosure, on a product sold inside a lawyer-client fee relationship. The savings claims are the commercial case: a named attorney describing eighty per cent of the time freed on some of his tasks, projects that took days or weeks handled in hours, firms taking on more cases without adding headcount, and medical record review and drafting compressed from hours to minutes.
Nothing published addresses what any of that does to what a client pays, whether AI-assisted work is identified anywhere, or whether a client is told the tool was used. One feature of this buyer complicates the reading and is recorded rather than resolved. Plaintiff firms commonly work on contingency, so time saved falls to the firm rather than reducing an hourly bill; but medical record review and chronology preparation are also routinely charged to the client as case costs deducted from recovery, which is the disbursement-shaped question already parked from the Verbit build.
Either way there is a fee relationship and a saving, and nothing published connects them. No cost pass-through statement was located, and no pricing is published at all.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
No located public material supports a client side disclosure obligation.
No located public material supports a client-side disclosure obligation, and the three artifacts this signal looks for are absent from every readable surface. There is no subprocessor list anywhere. No model provider is named, so a firm asked which third party processes its client's medical records could not answer from anything published; the vendor's closed-environment framing describes a boundary without naming what sits inside it.
No forwardable client-facing disclosure pack, consent template or notification material was located, and no data processing addendum appears. The certifications wall names eight standards without a scope statement, a date or a route to a report, so it cannot serve the purpose either. The value is not on-request, because nothing indicates such material sits behind a sales conversation: no security contact, trust centre or request mechanism is published.
Two documents would change the position and could not be opened by this index on the date shown, the eighteen-page Data Privacy and Security Policy and the Master Software Subscription Agreement; under R20 that limit is recorded rather than held against the vendor, and this row would move on a reading of either.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Some elements of the record are available, short of a document level export.
Several elements of a record exist and no disclosure-shaped export is described, which is this value. The elements are real and better than most. Every response is stated to tie back to a specific source in the case file, chronologies are produced in timeline format with citations, and the vendor states its rule as no citation, no answer, so the provenance of an assertion is visible at the point of use. The output is portable: timelines are described as exportable for use in memos and trial exhibits and as exportable for sharing or trial use, and medical chronologies are positioned as the foundation for expert reports, deposition preparation and motions.
So a firm can hand on a sourced chronology. What is missing is the disclosure dimension. Nothing identifies which model produced a passage, and no model is named anywhere in any event; nothing distinguishes machine-generated text from attorney-edited text in an exportable form; no verification record captures who checked what; and no disclosure template, certification form or court-facing guidance was located, despite the vendor writing publicly about courts sanctioning attorneys over AI-generated filings.