P
Prevail
Prevail is a testimony management and court reporting platform from Prevail Legal, Inc. of San Diego, built for depositions, arbitration hearings, internal investigations and other proceedings. Its sessions combine video conferencing, exhibit management and AI generated rough transcripts available at once, with certified transcripts that Prevail says go through human review by credentialed court reporters before delivery.
CheckMate, launched in February 2026, puts live transcription with speaker identification, shared notes and AI analysis on one screen, and lets a legal team query Prevail's native language model or connect its own. AI summaries produced during a deposition link each section to page and line ranges in the rough transcript, and Session Review gathers the record and the analysis after a proceeding ends. Prevail also offers AI transcription and translation in more than 90 languages and sells to federal agencies.
It states SOC 2 Type 2 attestation and ISO 27001:2022 certification and hosts on AWS in the United States. Shook, Hardy & Bacon is named as a customer, and Perkins Coie and Greenberg Traurig appear among its client logos. Pricing is not published.
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.
AI drives the features the vendor leads with, on a court reporting and proceedings platform that would still run without it. The models produce the instant rough transcript of a session, the summaries written while a deposition is under way, and CheckMate's analysis through a connected language model. Transcription and translation of audio and video are AI work too. Around that sits a platform that does not depend on the models: video conferencing built for proceedings, exhibit tools and scheduling, and certified transcripts made by people.
The About page says the platform includes a certified court reporter in every session. A firm buys a proceeding with a human record maker, and the AI makes that record available sooner and searchable. The models are the engine of core capabilities rather than the whole product.
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.
Summaries are grounded to the record in a way a lawyer can follow, and the one accuracy figure has no method behind it. A post of 24 September 2026 on the AI summaries says each section carries a page and line range that opens the rough transcript at that point. It tells users to read the lines themselves before quoting an answer. The commercial page says the AI generally produces a 95% accurate rough transcript, with no test set, sample or definition of accuracy behind the figure.
CheckMate's grounding is not published: the help center describes it querying connected models, and nothing states whether an answer cites the transcript line or document it rests on.
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.
Human review stands between the AI rough and the certified record, and the analysis tools carry a verification instruction with no control structure. Every certified transcript is said to go through a round of human review by credentialed court reporters. A post of 13 August 2026, written after an Indiana appeals judge noted a speaker attribution error, describes the layers. Remote participants get separate audio channels, trained monitors tag speakers live in hybrid and in person proceedings, and scopists check the transcript word for word against the recording before certification.
CheckMate has no equivalent. It queries language models during live testimony and is described as helping teams find relevant documents and lines of inquiry. Nothing states what its output may not be used for, when it declines to answer, or how a lawyer confirms a point before acting on it in the room.
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.
One named firm stands behind a described deployment, with no figures attached. A release of 23 February 2026 launching CheckMate says Shook, Hardy & Bacon has built Prevail AI into its Complex Litigation Strategic Counseling Practice Group for early case assessment, case work up and trial. It quotes Patrick Oot, a partner who co chairs that group, on how the product changes preparation and questioning during testimony.
The home page lists Discount Tire, Perkins Coie and Greenberg Traurig as clients, without stories. No case study with figures is published, so what changed for the Shook practice in time, cost or outcome is not shown.
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.
Confidentiality sits in the agreement, while the training and privilege commitments sit in published posts that the contract does not include. The terms bind each party to treat confidential information as confidential, and do not mention training or models. A post of 1 April 2026 on protecting privilege discusses United States v. Heppner and states the training position, and a later post extends that position to model providers.
Nothing in the contract a firm signs carries those commitments, so they rest on the vendor's published word rather than on terms a firm could enforce.
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.
Supervision of AI work is engaged in public, and the advice line is never drawn. The privilege post names ethics opinions on evaluating AI tools for confidentiality risk and on informed consent. The post on the speaker attribution error quotes the Indiana appeals court's view that anyone using AI to help prepare a transcript is responsible for proofreading it, and the vendor draws a clear line between its AI rough and its certified transcript.
The terms carry no statement that the service or CheckMate's analysis is not legal advice, nothing on who may use the product, and no position on the licensing of digital court reporting by jurisdiction.
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.
Nothing published sets out how the AI is governed. The pages read, which include the security page, the help center and posts on privilege and on AI summaries, name no person or function accountable for model behavior. None describes testing before a release, publishes evaluation results, or sets out principles for how the AI is built and used. The security certifications cover an information security management system, not the conduct of the models.
Uneven output is a live question for this product and goes unaddressed. Transcription and speaker identification can fail unevenly across accents, dialects and the many languages handled, and nothing states how performance is measured across them. Human review of certified transcripts is an oversight step on the record, not a governance framework for the AI.
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.
Security controls and breach notice are published in working detail. The security page says customer data is encrypted at rest in AWS S3 with keys managed through AWS KMS. All access to customer data is logged and attributable to individual people, new employees get access only after a background check, and restorable backups are kept. The data security post adds that transcript editors reach only the jobs assigned to them, and that media cannot be downloaded or shared outside the platform. The terms commit to reporting a security breach within 72 hours, with the relevant details.
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.
An infringement indemnity and a fee cap are in the agreement, and nothing allocates the risk of AI output. The terms of service took effect on 16 October 2023, under California law with AAA arbitration in San Diego. Prevail defends and indemnifies the subscriber against claims that the service infringes a copyright, a United States patent or a trademark, including reasonable attorneys' fees. Liability is capped at what the subscriber paid in the six months before the claim arose.
All warranties are disclaimed, including any that the service will be secure or error free, and the subscriber indemnifies Prevail for its own breaches and negligent acts. Nothing says who bears the loss when an AI rough transcript, a summary or a CheckMate answer is wrong. The terms predate CheckMate and do not mention transcripts or court reporting.
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.
Model and repository connections are described with some detail of what moves, and none reaches a firm's document or case management system. A secure API streams live transcripts and notes to the organization's chosen language model, and the help center documents personal connections and organization connections set by an administrator. On 24 February 2026 Prevail and Consilio announced that Prevail's capabilities would be delivered through Consilio's Aurora platform, letting clients query Consilio hosted document repositories from within Prevail.
The announcement said availability would follow, and no later status is published. The government page names integration with RelativityOne and with Teams and Zoom, without describing what syncs. Session Review lets a team download documents singly or as one package. No document management, case management or other litigation support connection is published, and no Relativity implementation documentation was found.
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.
Storage and transcription work are both placed in the United States, and tenancy is not stated. The security page names Prevail's AWS region in the United States, and the data security post says customer data stays within Prevail's cloud environment in United States data centers. The government page goes further on processing: all hosting, editing, transcribing and data transfer take place within the continental United States, with no offshoring of data, labor or services.
The CheckMate release says customer data stays within each organization's environment, and firms can send analysis to their own language model instead of Prevail's. Nothing states whether customers share infrastructure or are separated at the tenant level, whether a region outside the United States is available, or where a connected third party model processes the transcript it receives.
Security Certifications and Trust Center
Independent attestation a buyer can pull without a sales call: SOC 2, ISO 27001, penetration test summaries, a trust center with current reports and named scope rather than a badge image.
Two standards are claimed with the auditor named, and no report, date or scope document can be reached. The home page and the data security post state SOC 2 Type 2 attestation and ISO 27001:2022 certification, both by Prescient Security. The security page says the SOC 2 controls cover security, confidentiality, processing integrity, privacy and availability, and that HackerOne tests the infrastructure several times a year.
The government page says Prevail holds authorizations to operate from multiple large federal agencies, none named. No trust center is published, the reports are not offered for download or on request on any page read, and no audit period or certificate date is published.
Model Supply Chain Disclosure
Which models sit underneath, whose they are, where they run, and whether the vendor commits to telling customers when that changes. A legal buyer inherits every dependency it cannot see.
The architecture is described and the providers are not. A firm can use the vendor's native model or connect its own legal AI model, with Consilio's Aurora given as an example, and the help center says the native solution is open to all members with no setup. Nothing published names the model or provider behind the native solution, the AI rough transcripts or the summaries. Nor does anything say where those models run, or commit to telling customers when they change. Only a firm's own connected model is known to it; what sits under transcription and the native model is not.
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 appears anywhere on the site. The pages read, among them the commercial page, the terms and the CheckMate release, show no price, plan, rate per page or hour, session fee or unit of charge. The terms refer to the amount a subscriber has paid under its subscription, which indicates a subscription without saying what it covers or costs. Nothing shows whether charging is by page, hour, seat or session, or whether CheckMate, translation and the summaries are priced apart from the court reporting itself. Every route to a number runs through a demo request or a scheduling contact.
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.
Customers and proceedings are described with substance, and the limits are not. The commercial and government pages address law firms, corporate legal departments and federal agencies. Depositions, arbitration hearings and internal investigations are the named proceedings, and the Consilio announcement adds compliance reviews. Sessions run remote, hybrid and in person, transcription and translation cover more than 90 languages, and the government offering adds hearings, interviews and redaction of personal information.
The blog addresses arbitration in several posts in 2026. Nothing states which jurisdictions the certified transcripts are prepared for or the court reporters are qualified in, which practice areas the analysis is built around, or whether firms outside the United States are served.
9 public documents
The public pages on file for Prevail, 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.
-
Client Data in Training, Bar Guidance Alignment
Read Oct 2, 2026
-
prevail.ai/commercial2 signals
Primary Law Corpus Provenance, Good Law Verification
Read Oct 2, 2026
-
prevail.ai/security2 signals
Prompt and Output Retention, Outside Counsel Guideline Readiness
Read Oct 2, 2026
-
Ethical Walls and Matter Segregation
Read Oct 2, 2026
-
Court Disclosure Support
Read Oct 2, 2026
-
Refusal and Uncertainty Behavior
Read Oct 2, 2026
-
prevail.ai/about-us1 signal
Billing and Fee Posture
Read Oct 2, 2026
-
prevail.ai/terms-of-service1 signal
Third Party Request and Subpoena Notice
Read Oct 2, 2026
-
Fabricated Citation Record
Read Oct 2, 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.
Client data is never used for training, according to published posts, and the agreement does not address training. The terms never mention training, models or AI, and grant Prevail no improvement right over customer content. The post of 1 April 2026 states that client data is never used to train models, retained in shared datasets or exposed across users. A post of 17 June 2026 says contractual protections stop language model providers from training on client content.
The CheckMate release qualifies the position: customer data is never used to train AI models without explicit permission.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
A specific retention period is published and the customer cannot change it.
Customer data is kept for at least seven years unless otherwise specified, according to the security page, which does not say who may specify otherwise or how. The terms give a customer ninety days after termination to retrieve its content and say nothing about deletion after that. No separate period is published for CheckMate queries and answers or for AI summaries, and the vendor says language model providers are contractually barred from retaining client content.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
Segregation is asserted in public materials with no published detail on how it is enforced.
User defined permissions are asserted to govern who sees what. Session Review keeps private files visible only to the person who uploaded them, with notes shared to organizations or to members taking part in a session, and CheckMate searches the files in the user's own briefcase when it queries a model. Nothing published describes walls between matters or teams within a firm, how organization administrators restrict access to a case, or how a connected model is kept from reading across matters.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
Terms commit to notice where lawfully permitted. No transparency report located.
Prompt written notice before a compelled disclosure is committed, to the extent the law permits. The terms reserve Prevail's right to provide confidential information to third parties as required and permitted by law, such as under a subpoena or court order, with that notice going to the subscriber first. No transparency report or count of requests is published.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
No located public material identifies the corpus behind the product’s answers.
The product works on the record of a customer's own proceedings and the files a team uploads, and does not describe retrieval of case law or statutes; no outside legal corpus is described. Checked the home, commercial and government pages and the CheckMate release on 2 October 2026.
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.
The product does not cite legal authority, so no check of subsequent history applies, and none is described. Checked the home and commercial pages, the help center and the CheckMate release on 2 October 2026.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
No located public material addresses what the product does when it cannot ground an answer.
No located material describes what CheckMate or the AI summaries do when the transcript or the uploaded files do not contain an answer. Checked the help center articles on CheckMate, the CheckMate release and Prevail's posts on AI summaries and Session Review on 2 October 2026.
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.
No court order, opinion or disciplinary record naming Prevail as the source of fabricated authority was located as of 2 October 2026. The AI Hallucination Cases database maintained by Damien Charlotin was searched for Prevail; its one result, an Israeli decision of 11 January 2026, does not name Prevail or its products.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Public materials engage with at least one named ethics opinion.
Three named ethics opinions are engaged, applied to confidentiality and consent rather than to features. The vendor's post of 1 April 2026 on protecting privilege names ABA Formal Opinion 512 and, on the duty to evaluate AI tools for confidentiality risks before use, North Carolina's 2024 Formal Ethics Opinion 1 and Texas Opinion 705. No surface maps the product to guidance by jurisdiction.
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.
Savings are claimed and billing is not addressed. The About page claims up to 40% savings on litigation costs, with no baseline, customer or method attached, and the product is used by law firms on litigation for their clients. Nothing located addresses how AI rough transcripts, summaries or CheckMate analysis are billed or disclosed to a client, or how the claimed savings would show on a bill.
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.
Neither a subprocessor list nor a model provider list is published, and no disclosure material for a client's AI terms was located. Checked the home, security, government and terms pages, the privacy policy (last updated 10 August 2021) and the help center on 2 October 2026.
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.
The record shows which transcript a person has checked, and nothing shows how the analysis was made. Because the AI version is labeled a rough and the certified version passes human review, a lawyer can tell which one has been verified. Nothing records which model produced a summary or a CheckMate answer, no query log or export for an AI use disclosure is described, and no template for a court's standing order was located.