Opus 2 vs TrialView: how they compare in 2026

The models drive several capabilities a buyer would miss, on a platform that ran without them for most of its life. Opus 2 Ingenuity is presented as an agentic framework and the feature list is substantial: summarisation and analysis of key documents, extraction of people, organisations, events and legal topics, natural-language querying scoped to a single matter or a selected document set, a prompt builder with reusable saved prompts, drafting and sense-checking of work product, speech recognition producing legal-grade transcripts, and live hearing analysis that compares testimony as it is transcribed against the documents in the bundle to surface inconsistencies. That last one is the most machine-dependent thing in this lane. Underneath it sits eighteen years of case management, shared document workspace, annotation, chronologies, electronic bundles, evidence presentation and hearing infrastructure, none of which needs a model to work, and the vendor's own framing places the AI on top of that rather than in place of it, describing the platform as the source of the matter intelligence the AI runs on. Graded level with Casepoint, Consilio and Nextpoint's peers on the same reasoning: the engine of real capabilities, layered on a workflow system that stands without it.

The models run through the product without being the thing that makes it work. Strip the AI out and a complete trial platform remains: court compliant bundling with OCR, pagination, indexing and cross referencing, evidence presentation synchronised across devices, integrated video conferencing and real time transcription. Those are the capabilities courts actually buy it for and none of them requires a model. What the AI adds is genuine and spans the lifecycle rather than sitting in one module: timeline and chronology construction backed to source documents, pattern and inconsistency detection across witness statements and depositions, natural language question answering over the document set, transcript interrogation, and live identification during a hearing of material that corroborates or undermines a proposition. Same shape as Everlaw and Relativity, both B, where a mature platform hosts the model layer rather than depending on it.

Quality is asserted repeatedly and never measured. The published claims are confident: better data relationships produce more accurate insights, AI features only contribute information that is relevant and helpful to lawyers, and the design prevents errors and erroneous data from having a negative impact on a case. Grounding is claimed in general terms, with query answers described as based on the content inside the selected documents and the data surrounding the matter, and the whole system described as locked down to a single matter. What is absent, checked across the AI page, the home page, the security page and the arbitral institutions page on 31 Aug 2026: no accuracy figure of any kind, no test set, no benchmark, no evaluation, no error rate, and no hallucination disclosure. Nothing states whether a generated answer or summary cites the document it came from, or whether a reader can open the source from the output, which is the verification surface this axis turns on. The gap matters most on the hearing feature, where the system flags inconsistencies between live testimony and the bundle in real time and counsel may act on that within minutes. One limb does not apply and is neither credited nor penalised: a citator or good-law check is out of scope for a platform working on a matter's own evidence.

Grounding is claimed consistently and nothing behind it is published. The vendor states that question answering over the document set returns sourced references, that automated timelines are directly backed by source documentation, and that first draft witness statements carry automatic hyperlinking to cited evidence. Architecturally this is the easier case: the corpus is the closed case file rather than the law, so a citation resolves to a document a user already holds and can open, and the failure mode is misattribution rather than invention of authority. That distinction is credited here. What is absent: no accuracy figure for inconsistency detection, no precision or recall for pattern identification, no measure of timeline extraction fidelity, no hallucination statement, no evaluation, and no account of what the system does when it cannot support an answer from the file. Searched the home page, the AI litigation page, the case preparation page, the hearing services page, the case management page and the product overview on 29 Aug 2026. Inconsistency detection presented during a live hearing is a high consequence output and nothing published tells a barrister how often it is wrong.

The clearest published statement of an autonomy limit in this pull, and it comes with a worked example. The Control principle commits that lawyers must control the process, using their own knowledge, experience and ingenuity, before any data is saved permanently, which is a persistence gate rather than a slogan: AI output is a suggestion until a person commits it. The chronology feature shows what that means in practice, and the vendor explains the reasoning rather than just the behaviour. It states that automatically generating a chronology fills a timeline with irrelevant placeholders a team must then review and remove, and that its approach instead surfaces events the AI identifies for the lawyer to assemble, deciding relevance themselves. The entity workflow has the same shape, presenting extracted people, events and organisations with a summary of each so the user can remove irrelevant ones and eliminate duplicates before anything is added to the case. Against that, checked on the AI page, the home page and the security page on 31 Aug 2026: no threshold at which a feature declines to answer, no confidence or uncertainty signal exposed, no route to report or correct a wrong output, and nothing describing what happens after an error reaches a work product.

The oversight is real but structural rather than stated. Output is consistently framed as suggestive rather than decisive: the AI identifies material that corroborates or undermines propositions, surfaces inconsistencies and proposes timelines, while the advocate marks up, annotates, tags and decides. Witness statement generation is explicitly a first draft. A legally qualified support team is stated to be available 24 hours a day through the dispute process, which is human oversight resourced with people who understand the work and is uncommon enough to name. Held at C because none of it is published as a model: no statement of where the system acts without review, no confidence indication, no escalation behaviour, and no description of what the AI does unattended during a live hearing while attention is on the witness. The naming of a first draft carries the assumption of review without ever stating it.

Deep, named and current deployment evidence, with no measured outcome attached to any of it. Five customer quotes carry a name, a title and a firm: a contentious legal technology portfolio manager at Linklaters, the chief knowledge and innovation officer at Dorsey & Whitney, the director of practice technology at Vorys, a litigation partner at Pillsbury, and a paralegal at Shepherd and Wedderburn, with two more from barristers' chambers clerks. Around nineteen firm logos appear including Baker Botts, Dentons, Herbert Smith Freehills Kramer, Holland & Knight, Mishcon de Reya, O'Melveny, Troutman Pepper Locke and Wilson Sonsini. The institutional deployments are the strongest element because they are independently checkable: the Singapore International Arbitration Centre's SIAC Gateway is stated to be powered by Opus 2, the Dubai International Arbitration Centre is named as a partner, and the transcript and hearing services are stated to have supported the Manchester Arena Inquiry, the Grenfell Tower Inquiry phase 2 and the Hillsborough Inquiry. Two adoption figures are published, 98 per cent of Am Law 50 firms and 92 per cent of the Financial Times top 25 most innovative firms. What is missing is measurement: those figures describe reach rather than result, no quote carries a number or a date, and no case study with a stated method was located.

The strongest forum level adoption evidence in the litigation category and the third record in this pull with public authority adoption. Named UK courts: the Commercial Court, the Business and Property Courts and the Technology and Construction Court, with the vendor stating wide use across UK and Ireland courts. Named arbitral institutions: the ICC, the IAC and the DDRC, described as a trusted provider for connected hearing services in domestic and international arbitration. A named piece of litigation, the Russian Aviation litigation, plus large investigations, regulatory hearings and statutory inquiries. Independent recognition that is dated and checkable: Digital Evidence Management Software of the Year at the 2025 LegalTech Breakthrough Awards, and recognition at The Lawyer Awards. Held at B rather than A because no law firm or chambers is named, no usage volume is published, and no outcome measure of any kind exists: nothing on hearing time saved, bundle preparation time, or error reduction. The evidence establishes that serious forums use it, not what happens when they do.

One precise commitment on the question that matters most for this product, and nothing else. The Secure principle states that clients can be certain their data will never be shared outside their instance or between two or more cases inside an instance, which is an explicit cross-matter and cross-tenant undertaking rather than a general assurance, and it is reinforced by the query feature being described as locked down to a single matter or selected documents, and by named infrastructure controls including firewalls, encrypted volumes, logical identifiers and ownership tags. The arbitral institutions material acknowledges the confidential nature of arbitration directly. That is the whole of it. No customer agreement of any kind is published, so none of it is contractual: the footer carries a privacy policy, a cookie policy, a modern slavery statement, a supplier code of conduct and a code of business conduct, and no terms of service, master agreement or data processing addendum. In consequence nothing addresses whether client material is used to train or refine models, no retention position exists for documents, prompts or generated output, and privilege and work product are nowhere treated in writing despite the platform holding the working papers of some of the most sensitive disputes and public inquiries in the country. The privacy policy was not opened on 31 Aug 2026 and is the rebuttal route.

Access control is claimed and the hardest confidentiality question on this index is left unanswered. The vendor states user permissions and encryption, secure cloud storage, and the ability to collaborate selectively with colleagues, counsel, witnesses, solicitors, clients or third parties. Those are real controls and the sharing model is described in the right terms. The unanswered question is structural to the product: this platform deliberately brings law firms, barristers, judges and arbitral bodies into one workspace, and in a contested hearing that means opposing parties and the tribunal operating inside the same system. Nothing published describes how one side's privileged working materials, annotations, private mark up and preserved digital copy of proceedings are separated from the other side's or from the tribunal's. No treatment of legal professional privilege or work product was located. That is not a hypothetical concern for a product whose selling point is a shared workspace. Checked the home page, the case management page, the hearing services page, the case preparation page and the product overview on 29 Aug 2026.

Nothing published addresses the advice line, and no document on the property would carry it. The footer legal inventory was read in full on 31 Aug 2026 and consists of a privacy policy, a cookie policy, a modern slavery and human trafficking statement, a supplier code of conduct and a code of business conduct. There is no terms of service, no disclaimer page and no professional responsibility statement. Nothing states that generated output is not legal advice, nothing addresses where responsibility sits when a drafted document or a flagged inconsistency proves wrong, and no bar or regulatory authority is named anywhere, including ABA Formal Opinion 512 and, for a vendor headquartered in London and selling to English firms and chambers, any Solicitors Regulation Authority or Bar Council guidance. What does exist is recorded rather than credited, because it is the inverse of what the band above describes: a supervision posture is published and a disclaimer is not. The vendor states that it is not trying to replace a lawyer's intelligence, that the AI is designed to enhance human ingenuity, and that lawyers must apply their own judgement before anything is saved. That is a statement about how the work is done, not about who answers for it.

Not located. The product generates first draft witness statements from recorded interviews and produces analysis presented in live proceedings, and the buyer set includes judges and arbitral institutions as well as advocates. Nothing published addresses the professional responsibility framing of any of that: no statement on the advocate's duty to verify AI identified inconsistencies before putting them to a witness, no positioning on machine drafted witness evidence, and no engagement with Bar Council, Law Society, Law Society of Ireland or judicial guidance on AI in proceedings. Checked the home page, the AI litigation page, the case preparation page, the hearing services page and the product overview on 29 Aug 2026. The gap is notable because the vendor supplies both advocate and tribunal, which is exactly the configuration professional guidance addresses.

Five published principles that are better written than most and still not a governance framework. Empower, Enhance, Control, Secure and Trust each get a paragraph, and unusually the paragraphs explain design decisions and rejected alternatives rather than restating the principle: the vendor describes why it does not auto-generate chronologies, why it preserves existing legal workflows rather than replacing them, and what it will not let the AI write into a case without a lawyer's action. Two of the five carry real commitments, on human control before persistence and on data never crossing an instance or a case boundary. What is absent is the apparatus. Searched the AI page, the home page, the security page and the arbitral institutions page on 31 Aug 2026 and located no accountable owner for model behaviour, no described pre-release testing or evaluation regime for the AI specifically, no published result of any kind, no model documentation, and no external standard or certification covering AI as distinct from information security. STRIDE threat modelling and CREST-accredited penetration testing are published but cover software security rather than model behaviour. Nothing anywhere addresses bias or uneven output, which has a specific edge here: the hearing feature analyses live testimony, and how it performs across accents, languages and interpreted evidence in international arbitration is exactly the question, unasked.

Nothing published. No AI policy, no evaluation methodology, no bias or fairness assessment, no accuracy monitoring, no drift statement, no model card, no named governance body and no external standard such as ISO 42001. The specific risk left unaddressed: a system that flags inconsistencies in witness statements and identifies material undermining propositions is making adversarial judgements about credibility, and any systematic tendency in how it does that would operate inside live proceedings without anyone being able to inspect it. Checked the home page, the AI litigation page, the case preparation page, the case management page and the product overview on 29 Aug 2026.

Strong on how the system is built and silent on what happens to the data inside it. What is published is specific and technical: AES-256 encryption at rest, TLS 1.2 in transit with internal component traffic also over TLS and email encrypted end to end, logical data segregation using firewalls, encrypted volumes, logical identifiers and ownership tags, client-defined password policy, multiple multi-factor authentication options, single sign-on linking to the client's own provider, and full client control over creating, modifying and removing users. The development side is better documented than most vendors manage, with STRIDE threat modelling and an assigned risk owner for each identified threat, regular CREST-accredited external penetration testing, automated and manual code assessment covering third-party dependencies, a continuous integration pipeline, and a formal change management protocol with documented risk and impact assessment. Against that, the data lifecycle is absent. Searched the security page, the AI page and the footer inventory on 31 Aug 2026 and located no retention period for anything, no deletion or return commitment, no data export provision, no subprocessor list, no named cloud provider, and no incident response or breach notification practice at all. Four of the five elements this axis asks for are missing. The privacy policy was not opened and is the rebuttal route.

No AI specific stewardship position located. General platform controls are stated, being secure cloud storage, user permissions and encryption, and those are graded on the Security axis and not double counted here. What is missing is everything about the model layer: nothing states whether case documents, transcripts, witness interview recordings or hearing material are used to train or improve any model, how long AI generated output and queries persist, whether processing is segregated by matter, or what happens to a case file after proceedings conclude. This product ingests recorded witness interviews and complete litigation files, which is among the most sensitive content on this index, and the training and retention position is not published. Checked the home page, the case management page, the case preparation page, the hearing services page and the product overview on 29 Aug 2026.

Nothing is published on who bears the loss when the system is wrong, because no customer agreement is published. The footer legal inventory, read in full on 31 Aug 2026, is a privacy policy, a cookie policy, a modern slavery and human trafficking statement, a supplier code of conduct and a code of business conduct in PDF. There is no terms of service, no master services agreement, no subscription terms, no service level agreement and no data processing addendum. Nothing therefore states a liability cap, an indemnity in either direction, a warranty on the software or on generated output, an insurance position, a limitation period, a governing law or a forum. The security page refers in passing to Opus 2 Services and Agreements meeting global legal and regulatory requirements, and names six regimes they are said to satisfy, which confirms agreements exist without publishing any of them. A buyer cannot read the allocation of risk in advance at any level. This is the second vendor in this category to publish no customer agreement at all, and the pattern in both cases is the same: an enterprise vendor whose customers negotiate individually and never require a public one.

No published position located. Nothing was found on liability for AI output, warranty, indemnity, service levels or remedy. Two failure modes make the absence concrete rather than formal: an AI identified inconsistency that is wrong and gets put to a witness in cross examination, and a bundle compliance or pagination failure at hearing, since court compliant bundling is the operational spine of the product and a defective bundle has immediate consequences in front of a judge. No service level commitment was located for a platform used live in court. Checked the home page, the product overview, the case management page and the site navigation on 29 Aug 2026.

Integration is referenced by category and never by name. The AI page states that core documents flow seamlessly from eDiscovery and document management systems into a single workspace, which tells a buyer the direction of travel and identifies neither system. The security page describes linking to a client's own single sign-on provider to centralise account control, which is a real and useful enterprise integration but an identity one rather than a practice systems one. Beyond those, checked the AI page, the home page, the security page, the arbitral institutions page and the full navigation on 31 Aug 2026: no document management system is named, with nothing located for iManage or NetDocuments, and nothing for matter management, e-billing, court filing, Word or Outlook. No application programming interface, developer documentation or integration index was located. There is a real physical integration the record should not miss, since it is unusual: the vendor operates hearing infrastructure and has partnered with the International Dispute Resolution Centre in London to build a dedicated suite, which is an integration into a venue rather than a system. The platform page was not opened and is the most likely home for named connectors.

Integration is real inside the hearing and absent outside it. Documented: an integrated video conferencing platform with breakout rooms, real time transcription feeds viewable in platform with AI transcript analysis over them, speech to text for witness interviews, and evidence presentation synchronised across multiple devices in the hearing room. Bringing a live transcript feed and evidence display into one synchronised environment is genuine integration work and is the core of what a hearing provider must do. What does not exist in public material: no named document management system, no practice management or case management integration, no ediscovery platform connector for matters where review happened elsewhere, and no API or export documentation beyond court compliant bundle output. A firm cannot tell how material reaches the platform from the systems where it already sits. Checked the hearing services page, the case management page, the case preparation page and the product overview on 29 Aug 2026.

Residency is offered explicitly, and where processing happens is not addressed. The security page states that the vendor hosts the client's solution and data in the geographical location that best fits their needs, which is a residency option stated as a customer choice rather than a fixed footprint, and it is corroborated by the shape of the business: three regional operations with separate EMEA, North America and Asia Pacific sites, and a published list of the data protection regimes the services are said to meet across those markets, naming GDPR, UK GDPR, Canadian PIPEDA, the California Consumer Privacy Act, the Singapore Personal Data Protection Act and the Australian Privacy Act. Tenancy is described in substance if not in name, with data said to be self-contained through firewalls, encrypted volumes, logical identifiers and ownership tags, which is a logical rather than physical separation model. What is not published: no list of the available regions or data centre locations, no named cloud provider or statement of whether the infrastructure is the vendor's own, and no distinction between where data is stored and where processing or model inference happens, which matters for a customer choosing a jurisdiction precisely to keep material inside it.

Cloud storage is stated and nothing else is. The vendor states that data is securely stored on cloud based servers accessible anywhere, which names no provider, no region and no residency commitment, and describes no single tenant, dedicated or on premise option. Residency is a live question for this product specifically: it serves UK and Irish courts and international arbitral institutions including the ICC, so material from proceedings in multiple jurisdictions sits in one platform under undisclosed geography, and a tribunal or a party may have residency requirements of its own. Checked the case management page, the home page, the hearing services page and the product overview on 29 Aug 2026. Correction candidate: the site links a dedicated security features page which was not read in this pass and may carry residency detail.

Real, current, named certification with nothing attached to let a buyer verify it. The security page, last modified 19 Aug 2026, names ISO/IEC 27001:2022 and Cyber Essentials Plus, and adds a genuine third-party testing element in regular CREST-accredited penetration testing by external groups, alongside STRIDE threat modelling with assigned risk owners and automated and manual code assessment. Naming the 2022 revision of ISO 27001 rather than the superseded 2013 version is worth crediting, since the transition deadline has passed. What is missing is everything around it: no certification body or auditor is named, no certificate number or scope statement is given, no validity period or issue date appears, no route is published for obtaining the certificate or any report, there is no trust portal, and no SOC 2 of any type is claimed, which a US buyer will look for first. One inconsistency belongs on the record: the arbitral institutions page states ISO 27001:2013 while the security page states 2022, so the same site publishes two different versions of the same certification and one of them is out of date.

Controls are claimed, a dedicated security page exists, and no certification was located. Published claims: secure cloud based servers, user permissions, encryption, and a secure workspace as the repeated framing across the site. The vendor maintains a distinct security features page linked from product pages, which is more than a passing mention and indicates deliberate attention to the topic. Searches on 29 Aug 2026 combining the vendor name with ISO 27001, SOC 2, Cyber Essentials and GDPR certification terms returned no certification, attestation, auditor or trust portal for this vendor. Held at C rather than D on an explicit and important limitation: the security features page itself was not read in this pass. Grading D would assert an absence on the one surface most likely to contradict it, which is the error pattern this pull has already corrected twenty five times. Flagged as a correction candidate in both directions, since reading that page could establish certifications that raise this grade or confirm their absence and lower it.

The vendor names its AI and never says what runs it. Opus 2 Ingenuity is described as an agentic AI framework running on the matter intelligence generated in the platform, and the capability set is set out at length across summarisation, entity extraction, natural-language query, prompt building, drafting and real-time speech recognition. Nothing identifies what sits underneath any of it. Searched the AI page, the home page, the security page and the arbitral institutions page on 31 Aug 2026: no model is named, no provider is named, no distinction is drawn between first-party and third-party models, no version appears, no inference location is stated as distinct from the general offer to host in a chosen geography, and no commitment exists to notify customers when a model or its configuration changes. The one adjacent commitment is a boundary rather than an identification: client data is stated never to be shared outside the client's instance, which tells a buyer where data does not go without telling them what processes it inside. For a platform whose AI reads live testimony in a courtroom, the absence of a named model is the disclosure a tribunal would most likely ask about.

Nothing located. No foundation model provider is named, no distinction is drawn between proprietary and third party models, no subprocessor list exists, and the transcription and speech to text capabilities are not attributed to any provider despite almost certainly involving one. The vendor refers to the latest AI capabilities and intelligent technology without identifying anything. For a platform used by courts and arbitral institutions, where a tribunal may need to know which parties processed evidence in a proceeding, the identity of the model and transcription providers is a question with procedural weight and it cannot be answered from public sources. Checked the home page, the AI litigation page, the case preparation page, the hearing services page and the product overview on 29 Aug 2026.

No pricing information is published at any level, including the unit of charge. The full primary navigation covering who we help, platform, AI, solutions, resources and company, and the complete footer across all three regional sites, were read on 31 Aug 2026: there is no pricing or plans page in either, and no pricing entry anywhere in the site structure. No rate, no unit, no tier or package names, no volume band, no term length and no statement of what implementation, hearing services or transcription add, which matters here because the offering mixes software subscription with delivered services including court reporters, transcript production and hearing room setup, each of which would ordinarily be priced differently. Every call to action across the home page, the AI page and the solutions pages is to watch a demo or contact the company. The only cost-adjacent claim located is a general assertion that connected data and workflows drive more efficient AI usage so a customer spends less.

No pricing at any level. No price, no range, no tier structure, and no unit of charge, which matters more than usual here because the plausible units are all different in kind: per matter, per hearing day, per user, per gigabyte of evidence, or per bundle. A litigation team cannot estimate the cost of a three week trial from anything published. Every route is a demo request. Independent software directories list pricing as available on request and carry no figure either. Checked the home page, the product overview and the site navigation on 29 Aug 2026. Common for hearing services providers and still an absence a buyer cannot work around.

Coverage is documented along two axes with real substance, and no boundary is stated. By role, which is an unusual and useful cut: partners and lead counsel, legal innovation teams, associates, and paralegals and litigation support, each with a dedicated page and its own stated benefits. By solution: Cases, Hearings, Transcripts, Arbitral Institutions, and Chambers sold as LEX, which together identify the buyer types as law firms, barristers' chambers and arbitral institutions. Practice focus is clear from the material rather than asserted, covering commercial disputes, international arbitration, public inquiries and inquests, with named inquiry work and named arbitral institution deployments, and the chambers product reaching criminal clerking. Geographic reach is real, with three regional operations and a published list of six data protection regimes across those markets. What is absent is any limit: the vendor states it modernises the litigation and arbitration lifecycle in any jurisdiction worldwide, and nothing identifies a practice area, matter type, jurisdiction or scale that the platform does not serve. Checked the navigation, the home page, the AI page and the arbitral institutions page on 31 Aug 2026.

Forum coverage is stated with unusual specificity and the user types are enumerated honestly. Proceeding types named: litigation, arbitration, mediation, adjudication, conciliation, tribunals, depositions, large investigations, regulatory hearings and statutory inquiries. Named forums: the Commercial Court, the Business and Property Courts and the Technology and Construction Court in the UK, with stated wide use across UK and Ireland courts, plus the ICC, IAC and DDRC arbitral institutions and international arbitration generally. User types named: paralegals, partners, counsel, barristers, solicitors, arbitrators, judges and arbitral bodies, which is a genuinely wider set than any other record in this category addresses. Held at B rather than A because coverage is described rather than enumerated: no list of which courts have the platform available, no statement of jurisdictional limits outside the UK, Ireland and international arbitration, and no indication of whether the AI features are available in every forum or only in some.

No located statement addresses whether client material is used to train or refine any model, either way. The nearest thing is a boundary rather than a training position: the Secure principle on the AI page commits that client data will never be shared outside their instance or between two or more cases inside an instance, which describes where data does not travel without saying what is done with it inside. Nothing states whether documents, transcripts, prompts or generated output feed model training, tuning or improvement, and no opt-out or opt-in mechanism is described. The absence is structural rather than an omission from one page: no customer agreement of any kind is published, so there is no instrument in which such a commitment could sit. Searched the AI page, the home page, the security page, the arbitral institutions page and the full footer legal inventory on 31 Aug 2026; the privacy policy was not opened.

Silent. The quoted phrase is the extent of the published data handling position and it describes access control and encryption, not model training. No statement in either direction was located on whether case documents, transcripts, witness interview recordings or hearing material are used to train or improve any model. The content at stake is unusually sensitive even for this index: complete litigation files, recorded witness interviews and live proceedings. Checked the home page, the case management page, the case preparation page, the hearing services page, the AI litigation page and the product overview on 29 Aug 2026. Recorded as silent, not as a negative commitment. Correction candidate: a dedicated security features page exists on the site and was not read in this pass.

Nothing published addresses retention of anything. Searched the AI page, the home page, the security page, the arbitral institutions page and the complete footer legal inventory on 31 Aug 2026 and located no retention period, no deletion commitment, no return or export provision at the end of a matter, and no customer-configurable or zero-retention setting, either for the documents and transcripts held in a case or for the prompts and generated output the AI features produce. The security page addresses encryption, access control and development practice in technical detail and stops before the data lifecycle. There is no customer agreement published in which a retention term could appear. The question has particular weight for this vendor because its hearing and transcript work covers public inquiries and confidential international arbitrations, where the period for which a record persists is often itself a matter of order. The privacy policy was not opened and is the rebuttal route.

Not addressed. No retention period is published for AI queries over the document set, generated timelines, transcript analysis or draft witness statements, and nothing states what happens to a case workspace once proceedings conclude. Retention is a live question in this product's own terms: the vendor states that a user can preserve their own digital copy of proceedings, which is retention as a feature, and says nothing about the vendor side duration. Checked the case management page, the hearing services page, the case preparation page and the product overview on 29 Aug 2026.

An explicit cross-matter commitment, published as one of five AI principles rather than buried in a security annex. The Secure principle states that with every AI enhancement clients can be certain their data will never be shared outside their instance or between two or more cases inside an instance, which addresses both tenant separation and the wall between two matters run by the same firm on the same system. It is reinforced at the feature level, where natural-language querying is described as locked down and focused only on a single matter or a selected set of case documents, and at the infrastructure level, where the security page names the mechanisms as firewalls, encrypted volumes, logical identifiers and ownership tags, describing the result as self-contained. Client-side control is also stated, with full customer authority over creating, modifying and removing users and their permissions. What is not published: no description of how the boundary is enforced within the AI specifically as opposed to the storage layer, and no contractual form for any of it, since no customer agreement exists.

Claimed and not documented, and this is the most consequential instance of that value in the pull. The vendor states user permissions and secure sharing with colleagues, counsel, witnesses, solicitors, clients or any third party, so a permission model exists and is asserted. Nothing documents how it works. The product's defining feature is that law firms, barristers, judges and arbitral bodies share one workspace, which in a contested hearing means opposing parties and the tribunal inside the same platform, and no published material describes how one party's privileged annotations, private mark up and preserved copy of proceedings are walled from the other side or from the bench. No document management system exists to inherit permissions from, so a second model is the only available shape and its specification is not published. Checked the case management page, the hearing services page, the case preparation page and the product overview on 29 Aug 2026.

Nothing located addresses third-party demands for customer data. Searched the AI page, the home page, the security page, the arbitral institutions page and the complete footer legal inventory on 31 Aug 2026. No statement covers subpoenas, court orders, warrants, regulatory demands or government requests: nothing commits to notifying the affected client, nothing addresses resisting or narrowing a demand or seeking a protective order, nothing describes conduct where notice is legally prohibited, and no transparency report or disclosure statistics exist. No customer agreement is published in which such a clause could sit. The question is sharper for this vendor than for most, because it hosts the evidence bundles and hearing records of confidential international arbitrations and of statutory public inquiries, where a demand for material would raise questions the parties would expect to be told about. The privacy policy was not opened and is the rebuttal route.

Not addressed. No government or law enforcement request clause, no commitment to notify a customer before producing their data, and no transparency report were located. The question has an unusual edge here because the vendor holds material from live proceedings for courts and arbitral institutions as well as for parties, so a request could arrive from a body that is also a customer. Checked the home page, the case management page, the product overview and the site navigation on 29 Aug 2026.

The question does not arise in its usual form for this product. Opus 2 operates on the documents, evidence bundles and transcripts of a customer's own matter, so there is no vendor-assembled corpus of primary law, no third-party content licence and no upstream data supplier to identify. Recorded as not addressed because that is the honest value, with the reason stated so it does not read as a gap. One feature does sit close to the question and is recorded here: the vendor describes its AI as running on proprietary matter intelligence generated as legal teams collaborate in the platform over eighteen years, which raises rather than answers a provenance question about whether any of that accumulated intelligence informs models used for other customers. The Secure principle's commitment that data never leaves a client's instance cuts against it, and nothing states the position directly. Searched the AI page, home page, security page and footer inventory on 31 Aug 2026.

Not addressed, and largely inapplicable in the usual sense. This product has no primary law corpus: the AI operates over the parties' own evidence, bundles, transcripts and witness statements rather than over published authority, so there is no external legal source to name, licence or date. The signal is recorded rather than skipped because the underlying question still has an answer here, which is what corpus the models were built or tuned on, and that is not published. Nothing states whether models are general purpose, trained on legal material, or tuned on prior customer matters, the last being the possibility that would matter most to a litigant. Checked the AI litigation page, the case preparation page, the case management page and the product overview on 29 Aug 2026.

Not applicable to this product class, and neither credited nor penalised. The platform manages, analyses and presents the evidence and transcripts of a specific dispute; it does not retrieve primary law or assert propositions of law whose continued validity would need checking, and no citator or treatment signal is claimed anywhere. Searched the AI page, the home page, the security page and the arbitral institutions page on 31 Aug 2026. The accuracy question that does apply is whether a generated summary, extracted entity or flagged inconsistency faithfully reflects the underlying document or testimony, and it is recorded on the Citation Accuracy axis, where no measurement of any kind was located.

Not addressed, and inapplicable on the facts of the product. TrialView analyses the evidence in a case rather than researching the law, so there is no legal authority in its output for a citator to check and no treatment signal would have anything to operate on. Recorded rather than omitted so the absence is visible as a scope fact rather than read as a gap: a buyer comparing this record against a legal research product should understand that this row is empty because the product does not do that work, not because the vendor declined to document it. Checked the AI litigation page, the case preparation page and the product overview on 29 Aug 2026.

Nothing published describes what the system does when it is unsure. Searched the AI page, the home page, the security page and the arbitral institutions page on 31 Aug 2026 and located no abstention path, no confidence or relevance score surfaced to the user, no threshold at which a feature declines to produce output, and no statement of behaviour where the documents do not support the question asked. The vendor's answer to uncertainty is architectural rather than signalled, and is recorded because it is the substitute: AI output is presented for a person to accept or discard, with extracted entities shown for the user to remove irrelevant ones and eliminate duplicates, and with a stated rule that a lawyer must apply their own judgement before anything is saved permanently. That places a human at every commit point without telling them which outputs the system was least sure of.

Not addressed. No explicit no answer path, abstention behaviour or confidence signal is documented for question answering over the document set, for inconsistency detection or for timeline construction. The consequence is specific to live proceedings: during a hearing the platform surfaces material said to corroborate or undermine a proposition in real time, and nothing published tells an advocate whether a weak match is presented with the same confidence as a strong one, or whether the system will decline to surface anything at all when the evidence does not support the proposition being tested. Checked the AI litigation page, the hearing services page, the case preparation page and the product overview on 29 Aug 2026.

No court order, opinion, tribunal ruling or disciplinary record naming this product has been located as of 31 Aug 2026. Instrument searched: the AI Hallucination Cases database maintained by Damien Charlotin, which tracks decisions worldwide where a court addressed hallucinated AI content and records the tool implicated where known, alongside several independent 2026 sanctions trackers and a UK-specific tracker, searched on the company and product name. This is a statement about the public record on the date shown rather than a clearance, and one limit of the instrument is worth noting for this vendor in particular: the trackers cover court decisions, and a substantial part of Opus 2's work is confidential international arbitration, where a comparable failure would not ordinarily produce a published decision. A UK commentator maintaining one of these trackers notes anecdotally that hallucinated material is appearing in arbitration proceedings without being reported or sanctioned.

None located, with the instrument named. General web searches combining the vendor name with court, judgment, order, sanction and fabricated evidence terms returned nothing on 29 Aug 2026. No named docket database or court record tracker was searched in the UK, Ireland or elsewhere. The exposure shape differs from a research product: this system does not generate citations to legal authority, so the classic fabricated case failure mode does not apply, and the analogous risk would be a mis-cited or misattributed document reference within a bundle. Recorded as a statement about what this search found, not as a clearance.

No professional conduct authority is named anywhere. Searched the AI page, the home page, the security page, the arbitral institutions page and the complete footer legal inventory on 31 Aug 2026 and located nothing citing ABA Formal Opinion 512, any state bar guidance, or, for a company registered in England and selling to English firms and barristers' chambers, any Solicitors Regulation Authority or Bar Council guidance on the use of artificial intelligence. Nothing maps a platform feature to a professional obligation a supervising lawyer could evidence. The vendor does publish five AI principles that engage with lawyers' concerns about AI in general terms and commit to human control before data is saved, which is adjacent to the competence and supervision duties those authorities describe without citing any of them. Nothing addresses arbitral institution rules or practice notes on AI use either, despite the vendor supplying case management systems to arbitral institutions themselves.

Not addressed. No named guidance from the Bar Council, the Law Society, the Law Society of Ireland, the Bar of Ireland or any judicial practice direction on AI use in proceedings was located, and no ethics material of any kind appears on the surfaces a buyer would check. The absence is more visible than usual because the vendor supplies courts and arbitral institutions directly, and judicial guidance on AI in litigation is exactly the framework its own customers operate under. Checked the home page, the AI litigation page, the product overview and the site navigation on 29 Aug 2026.

Efficiency and cost claims without any treatment of how the work is billed. The published claims run to reduced review time, greater productivity, faster delivery of value to clients, and a cost-efficiency benefit framed as spending less because connected data and workflows drive more efficient AI usage. One customer quote goes further into revenue, with a chief knowledge and innovation officer describing the AI functionality winning a client on the phone and calling it a revenue-generating opportunity for the firm. Searched the AI page, the home page, the security page and the footer inventory on 31 Aug 2026 and located no per matter record of AI-assisted work intended for fee purposes and no guidance on billing, fee or client disclosure treatment. One adjacent resource is recorded because it is closer than most: the vendor publishes material on measuring legal technology return on investment through utilisation, time spent on tasks and time spent on matters, which is a measurement framework for the firm rather than guidance on what a client is told.

Savings claims only. The vendor states that the platform reduces administrative workload, cuts manual document preparation time, and allows teams to spend more time on case strategy and less on manual tasks, saving resources. Those describe the firm's own time and no figure is attached to any of them. Nothing appears on the client's side of the equation: no position on billing for AI assisted preparation, no time or audit record a firm could produce to a client, and no guidance on how machine assisted bundling or analysis should be recorded on a bill. Checked the case management page, the home page, the case preparation page and the product overview on 29 Aug 2026.

None of the artifacts a client's AI clause asks for is published or offered. Searched the security page, the AI page and the complete footer legal inventory on 31 Aug 2026: there is no subprocessor or model provider list, no data processing addendum, no security questionnaire response, no trust portal, no AI disclosure pack, and no stated route for requesting any of them. What a firm can point a client to is real but partial and is recorded so the value is not read as nothing published: ISO/IEC 27001:2022 and Cyber Essentials Plus are named on the security page, regular CREST-accredited external penetration testing is described, six data protection regimes are listed as met including GDPR, UK GDPR and the California Consumer Privacy Act, and the AI principles page carries a specific commitment that data never leaves a client's instance or crosses between cases. A firm asked which third parties process its matter data, or whether the vendor trains on it, would find no answer and no one to ask short of a bespoke negotiation.

Not addressed. No subprocessor list, no named model or transcription provider, no data processing agreement, no trust centre and no self serve documentation request route were located, so a firm has nothing it could forward to its own client. General claims of encryption and user permissions are not a forwardable disclosure. Checked the case management page, the home page, the product overview and the site navigation on 29 Aug 2026. Correction candidate: the dedicated security features page was not read in this pass and is the surface most likely to hold anything relevant.

Nothing published would help a lawyer disclose or certify AI use, and this vendor raises the question more directly than any other in the pull. Its hearing product runs AI inside a live proceeding: as testimony is transcribed in real time, the system compares the evidence being given against the documents in the hearing bundle and generates insights such as inconsistencies, which counsel may act on within the same session, and teams can additionally run their own prompts against live evidence. Searched the AI page, the home page, the security page and the arbitral institutions page on 31 Aug 2026 and located nothing on whether that use is or should be disclosed to the tribunal or the other side, no model identification or versioning, no exportable record of which AI outputs were generated or acted on, no log of what a user accepted or discarded despite the stated rule that a lawyer commits every AI-derived datum, and no template or guidance for a court's standing order on AI use. The vendor also supplies case management systems to arbitral institutions, which places it on both sides of the question.

Partial record, and unusually well positioned on the source trail because the whole product is built around evidencing where things came from. Court compliant bundles carry automatic pagination, indexing, bookmarking, hyperlinked indexes and dynamic cross referencing, with built in compliance checks. Timelines are stated to be directly backed by source documentation, draft witness statements carry automatic hyperlinking to cited evidence, question answering returns sourced references, and a user can preserve their own annotated digital copy of proceedings. A party can show what a proposition rests on and where in the bundle it sits. What is missing is the AI specific limb a standing order on AI use would ask for: nothing records which outputs were machine generated, no model is named or logged against an output, and no human verification record is captured showing that an advocate checked an AI identified inconsistency before deploying it. The product evidences the documents thoroughly and evidences its own involvement not at all.