Bench IQ vs Pre/Dicta: how they compare in 2026

The models make up both the product and the dataset. Bench IQ sells a description of how a named federal judge reasons, produced by large language models and AI agents from rulings never issued in writing. Without the models nothing remains to sell: no repository, workflow, document store or search over published opinions, and no dataset, because the dataset is itself extracted by machine from oral rulings and other unwritten indicators rather than a collection that existed beforehand. The privacy policy names large language models as the means of analyzing rulings, and the vendor describes fine tuning commercial models on the dataset it built.

The model is the product. A user enters a case number and receives a forecast; there is no repository, workflow, drafting tool or document store underneath that would work without the model. Only the training data would remain. The vendor says it is neither a self serve analytics tool nor a case law research vendor, and the interface is built around the prediction rather than search. This is a platform that exists to deliver model output.

Nothing is published on accuracy or grounding for a product that makes claims about what a named judge thinks. The site has five pages, with the home page serving as the product page: three FAQ blocks, a social proof claim and a demo request. Checked those, plus the about, press, privacy policy and responsible disclosure pages,, and found no accuracy figure, test set, evaluation, error rate, failure modes or hallucination disclosure. Verification matters especially here: the product's value is reaching the roughly ninety seven percent of rulings with no written opinion, so the source behind an insight is material a reader cannot easily check in a conventional research tool. Nothing says whether an insight links to the ruling it came from, whether a user can inspect that source, or what happens when the record is thin. Citator checks do not apply to a product that describes judicial reasoning rather than validating authority.

Pre/Dicta publishes a measured accuracy figure rather than only asserting accuracy. The claim is 85 percent on motions to dismiss, with the test set described as more than 50,000 motions to dismiss across all 94 US federal district courts, drawn from about two decades of federal data. Its limits are stated against the vendor's commercial interest. The 85 percent covers motions to dismiss only, and for the other nine motion types the vendor says it models outcomes rather than predicting them, a distinction shown on the home page. An interview transcript on Pre/Dicta's own site gives 87 percent, against 85 everywhere else, and nothing reconciles the two. The product forecasts outcomes rather than making cited legal assertions, so grounding to primary authority and citator checks do not apply. The vendor discloses instead that forecasts are produced independently of case facts and precedent. The method behind the figure is not described, so whether it came from a holdout backtest, live tracking or in sample validation is unknown, and no outside validation is published.

No oversight structure is published for a system whose output feeds live litigation strategy. Searched the home page, about page, press page, terms of service, privacy policy and responsible disclosure page: nothing describes what the models decide unattended, no review surface is documented, no confidence or abstention behavior is described, no statement tells a lawyer what weight an insight should carry, and no route exists to report or correct an insight that proves wrong. The vendor is publicly careful about the framing, its chief executive saying the aim is not to predict the outcome of a case but to help lawyers shape it, which locates judgment with the lawyer. That distinction is real and is recorded here, but it was made in an interview rather than in product or policy material, and a framing is not a control.

A real review surface is published and the rest of the control structure is not. The vendor commits that every output is inspectable down to the underlying docket, that every underlying comparable case is visible, inspectable and linked, and that outcomes are verifiable, which gives a lawyer a genuine route from a forecast back to the matters that produced it. The platform also allows a user to re-weight or change the factors behind a prediction, which is a form of human control over model output rather than passive consumption. What is absent, checked 31 Aug 2026: nothing states that a lawyer must review before relying, no threshold or confidence gate is described, and nothing addresses what happens when a forecast proves wrong. On a product whose output feeds settlement valuation and venue strategy, the absence of any published statement about the weight a forecast should carry is the gap.

Adoption counts and anonymous testimonials stand in for evidence. The vendor publishes two figures, both self-reported and neither naming anyone: four of the top five AmLaw firms are said to use the platform, and an earlier announcement put twelve of the hundred largest US firms as pilot customers actively using the product on live cases. Three testimonials appear on the home page attributed only by seniority and firm rank. No customer is named anywhere, no case study exists, and no figure is published for what changed at any organization. One conflation to guard against, because the same names circulate in both roles: Cooley, Fenwick & West, Wilson Sonsini and several Kirkland & Ellis partners are investors in Bench IQ, not disclosed customers. Both bands have some purchase here, since an adoption count is closer to real deployment evidence than a logo strip, and the lower is taken because the count is unverifiable, unattributed and silent on outcome.

One named customer of real weight, with no figures. John B. Quinn, Executive Chairman and Founding Partner of Quinn Emanuel, is quoted saying all attorneys at the firm now have access to the platform. That is a named firm wide deployment at a major litigation firm, not a logo or anonymous quote. The vendor also won ALM Legal Week's Innovator of the Year award. No figure shows what changed at Quinn Emanuel or anywhere else, and there is no case study with a stated method and no customer count. No other named customer is published on the pages read. The 85 percent accuracy figure measures the product, not its deployment.

The published policy is a website and app privacy notice, not a customer data commitment. Effective 20 February 2025, it covers site visitors and app users and handles that well: personal data limited to name, job title, employer, work email and work phone, which the vendor notes is already public through firm websites, state regulators and PACER; no sale or rental of personal information; thirteen named subprocessors; storage and processing in the United States only; and a named security officer, the cofounder and chief technology officer. The platform takes in judicial records, not client documents, so there is no client matter material to separate, no ethical wall to inherit and no privilege question about ingested content. What does apply is unanswered: a litigator researching a named judge on a named issue reveals the matter they are working on, and nothing says whether that query history is kept, analyzed, used to improve models or disclosed. No customer confidentiality commitment of any kind was found.

The privacy policy carries a dedicated section on case reports that addresses the actual exposure this product creates. It recognizes that case numbers and reports may be sensitive, and commits that Pre/Dicta does not share, sell or disclose case numbers, case report contents or access history to third parties, that encryption and strict access controls prevent unauthorized viewing, and that case access logs are used only for security monitoring and troubleshooting rather than marketing or profiling. It also states that case numbers are not stored against a personal profile unless the user creates an account. That is a considered answer to the real risk here, which is that the matters a litigator looks up reveal the firm's docket and strategy. Three gaps: no statement on whether anything a user enters trains any model, no retention period, and privilege and work product are not addressed. Note the underlying exposure is lower than for a document platform, since the input is a case number rather than client files.

Nothing is published on the advice line, and no page exists that would usually carry it. The site has login, about us, careers and press in the navigation, and privacy policy, responsible disclosure, security and terms of service in the footer. There is no disclaimer or professional responsibility page, and none of the pages read says an insight is not legal advice, that no attorney client relationship arises, or that the lawyer remains responsible for the argument built on it. No bar or ethics guidance is named, including ABA Formal Opinion 512, and no jurisdiction or practice limit is stated. The product exists to shape how counsel argues to a specific judge in a live matter. The terms of service could only be read in fragments through search, so a disclaimer could sit in the unread part.

Nothing is published on the advice line, and there is nowhere it could sit. There is no disclaimer, ethics or professional responsibility page, no bar or ethics guidance named, including ABA Formal Opinion 512, and no statement that a forecast is not legal advice or no substitute for professional judgment. There are no terms of service either: the footer link labeled Terms of Conditions leads back to the home page, and no terms are published anywhere. The exposure is high because the published use cases are settlement valuation, venue selection, whether to move to dismiss and whether to challenge an expert, all decisions a client relies on counsel to make.

No governance position is published for a system that models the reasoning of named federal judges. Searched the home, about, press, privacy policy, responsible disclosure and terms pages: no responsible AI statement, accountable owner for model behavior, testing regime before release, or evaluation of uneven output, and nothing on what it means to describe how identifiable federal judges think. The method is withheld, which the vendor attributes to pending patents, so a buyer can examine neither the governance nor the technique. Bench IQ reconstructs judicial reasoning by large language model from material that was never published and cannot be independently checked, which makes the missing governance more significant.

No governance position is published for a model whose central design choice raises exactly these questions. Pre/Dicta's forecasts rest on judicial biography, and the vendor names the variables: education, appointment history, political affiliation, net worth and location, with judges compared against cohorts of similar background. Its marketing has promoted dismissal rates for Republican appointees, and a biographical intelligence feature lets users isolate a judge's political affiliation to see how it bears on outcomes in similar cases. No responsible AI statement, accountable owner, testing regime before release or fairness evaluation is published. Nothing discusses what it means to model identifiable federal judges by politics and personal wealth. The vendor may have defensible answers, but none is published, and on this product the question is central.

Published policy covers most of the ground, and the subprocessor list is specific. The privacy notice effective 20 February 2025 names thirteen subprocessors, including Amazon Web Services, Heroku, Segment, Mixpanel, FullStory, Intercom, HubSpot, Notion, Slack, Sentry and Datadog, without any access request. It also states storage and processing in the United States only; retention for the life of the relationship plus a further period for the vendor's own operations and archiving; deletion by email request; data processing agreements and model clauses with vendors where feasible and appropriate; only essential cookies in the application; and a named security officer, the cofounder and chief technology officer. A responsible disclosure program lists what is in scope, gives a reporting address, offers a discretionary reward, promises not to pursue researchers acting in good faith, and asks them to avoid privacy violations and disruption. The retention period after the relationship ends is not quantified, no access control model is described, and no breach notification practice was found.

A general privacy policy covers the ground in principle without specifics. It cites industry standard technical and organizational measures, including encryption, secure servers and limited employee access. It says case report access logs are purged automatically after a limited retention period unless the law requires longer, and handles transfers abroad through Standard Contractual Clauses. No security page is published: no encryption standard, access control model, named subprocessor, breach notification practice or stated retention period. The policy was also published with unreplaced template placeholders where the company name and privacy contact address should be.

What is published disclaims the risk and gives the buyer nothing. The terms provide the services as is, with all faults and as available, and disclaim all warranties, including merchantability, fitness for purpose and noninfringement. The customer's duty to indemnify Bench IQ and the limits on Bench IQ's liability survive termination, and Bench IQ may end the service without notice for any reason. No indemnity runs to the customer, no warranty covers any insight, and no insurance position appears. No liability cap figure was found, so whether one exists is unknown. The text found reads as an evaluation agreement, framing permitted use as testing the services and giving feedback, which raises whether paying customers sign a separate commercial agreement. The terms could only be read in fragments through search, and a full reading could change the picture.

Nothing is published, because there is no agreement to publish it in. The footer of every page carries a link labeled Terms of Conditions, and it resolves to the home page rather than to any document. Searched for terms of service, terms and conditions and any subscription agreement across the site and the wider web and located none. The consequence is total: no indemnity, no liability cap, no warranty or disclaimer of warranty on output, no insurance position, no allocation of risk of any kind between vendor and customer. This is a product that forecasts whether a motion will be granted and is marketed for settlement valuation and reserve-setting by insurers, sold without published terms addressing what happens when a forecast is wrong. Recorded as a vendor-side absence rather than a retrieval failure: the link exists and points to the wrong place.

No integration into any practice system exists, and this is an absence rather than a gap in the reading. The site inventory was established directly from the navigation and footer: login, about us, careers, press, privacy policy, responsible disclosure, security and terms of service. There is no integrations page, no developer or API documentation, and no product page on which an integration might be described, because the home page is the product page. Nothing names a document management system, matter management, e-billing, court filing, Word or Outlook, and no export or data feed is described. The product is reached through a browser at its own application and stands alone.

No integration into any practice system is published. No API is offered or referenced, no document or matter management system is named, no billing or filing system connection appears, and there is no integrations page. What the vendor publishes in this area are workflow conveniences rather than integrations: the platform is cloud-based and reachable from any location, reports can be shared under the firm's own letterhead, matter numbers can be assigned to reports for cost tracking, and no technical training is required to implement. Useful to a buyer, but none of it moves data between Pre/Dicta and the systems a litigation team already runs.

Residency is stated plainly. The privacy notice says personal information is stored and processed in the United States only, so a buyer knows where data sits and where it is handled, and the subprocessor list names Amazon Web Services and Heroku as hosts. Not published: a tenancy model, whether customer environments are separated, region options, an enterprise or private deployment option, or where inference runs as distinct from where records are stored. A five page site from a three year old company states its residency clearly.

Cloud delivery is implied and neither the tenancy model nor the region is stated. The only infrastructure statements located are that the platform is cloud-based and accessible from any location, and that data may be transferred to and processed in other countries where the vendor's servers or partners operate, with Standard Contractual Clauses and other lawful safeguards applied to those transfers. That acknowledges cross-border processing while naming no country, no region, no hosting provider and no customer-selectable option. Nothing describes single or multi-tenancy, and nothing distinguishes where data is stored from where it is processed. No residency commitment of any kind is offered.

A real third party trust center exists, hosted on Drata at a public address and linked from every page, so the vendor has set up an independent compliance portal and points to it prominently. The site itself names no standard, auditor, scope or coverage period on any page read. The portal could not be read that day, so nothing about its contents is assumed, and it may say more. On the pages read, all that is stated is that a trust center exists.

A badge with nothing behind it. The footer of every page carries an AICPA SOC for Service Organizations mark, in the non-CPA form licensed to service organizations, which asserts a SOC examination without stating which one. No accompanying text names SOC 1 or SOC 2, no Type I or Type II is specified, no auditor is identified, no coverage period or report date is given, no scope is described, and no route to request a report exists. There is no trust center or security page anywhere on the property. A mark of this kind is meaningful only alongside the report it refers to, and the report is neither published nor offered.

The vendor names the type of model but not whose. The privacy notice names large language models as the means of analyzing rulings, and the company has described fine tuning commercial large language models on its own dataset, so third party foundation models sit underneath. No provider, model or version is named, no inference location is given apart from US storage, and no commitment to notify customers of model or training data changes was found. The same privacy notice names thirteen subprocessors for hosting, analytics, session recording, support, marketing and monitoring, and none is a model provider. A vendor that lists Sentry and Datadog by name but omits the systems that produce its core output has drawn a line somewhere, and nothing explains where or why.

The vendor describes its own model in detail and says nothing about anything it may depend on. It describes machine learning models trained on about two decades of federal litigation data, around 15 to 20 million cases, with 50 to 100 data points per case, dozens of variables for parties and firms, and biographical profiles of hundreds of federal judges. No third party model provider is identified. The architecture as described looks like proprietary machine learning rather than a licensed foundation model, though nothing confirms that either way. No model version or release identifier is given, so a buyer cannot tell which model produced a forecast or when it changed. There is no commitment to notify customers of changes to the model or its training data, and no subprocessor or infrastructure provider is named.

No pricing information is published at any level, including the unit of charge, and no pricing page exists to carry it. The site inventory was established directly rather than assumed: the navigation is login, about us, careers and press, and the footer adds privacy policy, responsible disclosure, security and terms of service. There is no pricing page, no plan or tier names, no feature split, no per-seat, per-judge or per-report unit, no term length and no statement of what implementation adds. The single call to action on the home page is a demo request. The only adjacent statement located is a 2024 interview in which the chief executive referred to flexible pricing models and said the product is designed for firms of all sizes, which names neither a figure nor a unit and does not appear on any vendor page.

No pricing information is published at any level, including the unit of charge. Searched the home page, platform page, privacy policy, the full primary navigation covering platform, about, research center, news, support and contact, and the footer. There is no pricing page, no tier structure, no per-seat, per-case or per-report unit, no volume banding and no statement of what implementation adds. Every call to action is a demo booking or a case consultation request. The one adjacent statement is that the platform requires no technical training to implement, which speaks to effort rather than cost. Nothing published would let a buyer form any view of price before entering a sales process.

The product page claims total coverage, and the real boundary appears only elsewhere. The home page describes covering all of a judge's rulings rather than the roughly three percent with written opinions, which reads as complete coverage, and names attorneys as the audience and large litigation firms as buyers. The real limits appear only in funding announcements: coverage is the US federal judiciary, starting with commercial bankruptcy, and state courts, where the vendor notes there are far more judges, are an expansion target along with other common law jurisdictions. A buyer reading only the site would not learn that state court judges are missing. No practice area list, court by court coverage, list of judges or historical range is published.

The buyer segments are carefully reasoned, and coverage is only half described. Four buyers are named with the question each brings: litigators setting strategy and timing motions, general counsel choosing outside counsel, private equity valuing law firm acquisitions on portfolio performance, and insurers setting reserves against how similar matters resolve. Each has its own page. Federal coverage is concrete: all 94 district courts, the appellate level, and ten named motion types from motions to dismiss through Daubert challenges, temporary restraining orders and motions to remand. State coverage is described only as select state courts, with no list, no court detail and no depth statement. The January 2023 acquisition of Gavelytics brought state trial court assets, and their current coverage is not published.