HAQQ vs Harvey: how they compare in 2026
Scale and region separate these two more than features do. Harvey is an enterprise platform for large law firms, in house departments and professional services networks. Its products include Agents for multistep legal work, Vault for bulk document review, Knowledge for legal, regulatory and tax research, Spaces for shared work, and Command Center for analytics on a firm's AI use. Harvey sells through its sales team. HAQQ, built in Beirut, pairs a legal AI chat with eFirm, practice management for law firms, so research, drafting, matters, time and invoices sit in one system. Its engine detects the jurisdiction and language of a request and keeps a maintained set of statutes for 33 countries. It works natively in Arabic and replaces personal details with stand ins before an outside model writes the answer. Plans start free and are priced per user with credits. A global firm will look at Harvey. A firm in the Gulf, the Levant or North Africa will look at HAQQ, which runs the practice and drafts under local law.
At a glance
All 15 axes, side by side
The same grid applied to every vendor in the index, graded from public sources. Hover a grade to see what the letter means on that axis.
AI Centrality
How much of the product is actually AI. Whether the machine learning is the mechanism the buyer is paying for or a feature layered onto conventional software, and whether the vendor is specific about which is which.
Legal AI Chat is the product HAQQ leads with, and it exists to answer, draft and review with a model. Justinian, the engine underneath, rewrites the prompt, classifies intent, detects language, extracts entities, reads documents with OCR and routes the request to tools and a model, all before generation. It then runs sub agents, memory, deeper reasoning with lookahead and rollback, a citation audit and human approval steps. A third party frontier model writes the final answer, and HAQQ says so plainly. Orchestration, retrieval, memory and safety stay in HAQQ's own layer. eFirm, the practice management product, is sold on its own subscription and would still run matters, time and billing without the AI. With both products, the AI reaches the firm's matters, documents and contacts, and the Legal AI Twin drafts in the firm's own style and templates inside its tenant.
Agents, Vault review tables, Knowledge research, Contract Intelligence review, Intake triage and Horizon Scanning all produce their output with language models. No document management or workflow system sits underneath that a firm would buy on its own. Vault stores matter documents, and its purpose is extraction, comparison and reporting across them.
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.
Answers cite the provision or case behind each line. Justinian retrieves statutes, case law, gazettes and regulatory databases for the detected jurisdiction before it writes, and where sources conflict or the law is unsettled it flags the uncertainty rather than picking a side. HAQQ runs two benchmarks of its own. The task benchmark puts 19 systems through 11 legal task categories and marks each output out of 50 against criteria that include hallucination and source linking. In the July 2026 run Justinian averages 45.8, with Spellbook, LexisNexis +AI and ChatGPT ahead of it on contract drafting, legal research and plain language explanation. A cross domain study draws on public sets such as LegalBench, CUAD and LegalCiteBench. The individual tasks are not published, and nothing describes a check of whether a cited case was later overturned. A claim of 85% fewer hallucinations in HAQQ testing appears without a method beside it.
In the February 2025 Vals Legal AI Report, a third party benchmark set against a lawyer baseline, Harvey Assistant was tested on six tasks. Its results ranged from 65.0 to 94.8 percent, and it beat the lawyer baseline on five. Document question answering reached 94.8 percent, 24.7 points above the baseline, with responses under a minute. Answers on scanned and badly formatted court transcripts reached 77.8 percent, 24.1 points above it. That study measured task accuracy rather than citation validity, and Harvey did not enter the later Vals study of citation authority. BigLaw Bench, Harvey's own benchmark, publishes its task categories and judging criteria on a public repository, with hallucination and source results by model. It defines a hallucination as a factual claim that a source of truth can disprove and tracks reasoning errors apart. Harvey reports that its systems hallucinate less often than the foundation models on legal tasks while giving longer answers. Every agent claim links to the source passage behind it, and the full BigLaw Bench dataset is available on request.
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 is written into the terms. Court submissions, legal advice, regulatory analysis, client communications, settlement analysis, deadline calculation and professional opinions need review by a professional before anyone relies on them. Users stay responsible for verifying citations, dates, calculations and deadlines. In the product, the agent harness carries guardrails, audit trails, lookahead with rollback, and configurable human approval before consequential agent actions. Each answer shows the sources behind it and the credits it used. A contract review returns the analysis first, then the corrected draft and a comparison of what changed. Which agent actions pause for approval by default, and what happens after an agent step goes wrong, are not published.
Agents show their plan before they run, so a lawyer can adjust the scope and approve the work before Harvey begins. While an agent works, Harvey brings the lawyer back in when a judgment call is needed, logs every step and cites a source for every claim. Intake applies routing rules the legal team defines to send each request to an approved agent, to Contract Intelligence or to a lawyer. Reports and reviews can run on a schedule, and admins govern access through roles and workspace permissions. The point at which an agent stops on its own, and what Harvey commits to when an agent's work is wrong, are not stated.
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.
More than 17,000 legal teams across law firms, in house departments and enterprises use the product by HAQQ's count, and the mobile app reached 7,500 downloads after its July 2026 launch. Named users are few. Zouhdi Yakan, a partner at Yakan Law Firm, recommends eFirm to other firms for practice management, and the president of the Beirut Bar Association is quoted on the home page. In October 2026 HAQQ signed a protocol to become the exclusive technology partner of the New Cairo Bar Association, which has about 22,000 lawyers, with the association's requirements still to be worked out. Partnerships with the Jordanian Arbitrators Association, Amman Arab University and ONEIC in Oman were announced without a published deployment or result. Outcome figures such as 70% less drafting time and a 95% review pass rate come from HAQQ's internal benchmarking. The claim that firms move from one contract a day to ten or fifteen comes from what firms report, with no firm named.
Customer stories name the firm and give figures. Reed Smith, with about 1,800 lawyers in 34 offices, reports 700 Harvey users a day, 1,400 a month and AI certification for every associate. Repsol's legal division started a pilot in January 2024 with 50 lawyers in 12 jurisdictions and reports cutting the time on one process by more than 90%. CMS reports 95% adoption. Other named customers include A&O Shearman, Gleiss Lutz, Cuatrecasas, Macfarlanes, KKR, PwC and Deutsche Telekom. The figures are usage counts and time savings as each customer reports them, with no stated method for how they were measured.
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.
The terms give matter data its own section on confidentiality and privilege. HAQQ acknowledges that matter data may carry attorney client privilege, work product protection and professional secrecy. It limits its own access to support, security, customer instruction and legal compulsion, and states that such access waives no privilege. The agreement bars training or fine tuning any foundation, cross customer or third party model on customer data, prompts or outputs without a separately signed agreement. Each firm is its own tenant, eFirm permissions run record by record, and HAQQ describes ethical walls and cross matter isolation as part of the platform. Before a Legal AI Chat request reaches an outside model provider, HAQQ's servers swap names, emails, phone numbers, addresses, company names and account numbers for stand ins. The request stops if that step fails. The rest of the content still goes to the provider, and the terms do not say what the provider retains.
Neither Harvey nor its subprocessors may train any AI model on customer content or data, and model providers work under zero data retention. A firm can ask for a bespoke model trained on its own data and used only for that firm. Harvey syncs and enforces the firm's existing ethical wall policies and never creates, modifies or deletes walls, so the walls provider stays the system of record. Workspaces are logically separated, access follows roles, and data is deleted within 30 days of termination. If the law or a court compels disclosure, Harvey gives advance written notice where it is legally permitted. Attorney client privilege and work product are not addressed as such.
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.
As the terms put it, HAQQ is a technology provider, not a law firm, giving no legal advice, opinions or representation. Users keep responsibility for professional judgment, client advice, filings, deadlines, conflicts and the supervision of staff and technology, and higher risk output needs a professional's review before use. The product is also sold to people who are not lawyers. The mobile app and the plans for everyone promise a clear answer to any legal question, a reading of a lease or a drafted contract, and HAQQ estimates $25K a year in legal fees avoided per consumer. Its consumer guides carry a line saying they are legal information, not legal advice. Jurisdiction limits are stated: 33 countries have a maintained set of statutes and codes, and more than 200 others get general reasoning only. A guide walks lawyers through ABA Formal Opinion 512 and the duties of competence, confidentiality, candor and supervision.
Output is not legal advice and may contain errors, misstatements or gaps, and the Platform Agreement calls the service a research tool. Harvey sells to lawyers, legal departments and professional services firms, and its agent pages leave the judgment on every task with the lawyer. Access credentials belong to one named user and may not be shared. Nothing addresses competence and supervision duties under professional rules, or the jurisdictions where the product may be used.
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.
Controls are described as aligned with ISO 42001, the AI management standard, without certification, and no AI policy, governance framework or accountable owner is published. Testing is public in outline. The task benchmark gives 19 systems the same prompt for each of 11 legal task categories and marks every output out of 50. The criteria include Sharia, statute, forum, clause construction, risk, hallucination, partner readiness and source linking. A second study runs 25 benchmarks across legal, tax, journalism, general and safety work, including political neutrality, robustness and adversarial testing. The limits are stated alongside the results. HAQQ builds the benchmark, runs it and competes in it, and coverage leans toward corporate and commercial drafting and research, leaving some practice areas covered lightly or not at all. Who marks the outputs is not stated. An open source model router under AGPL 3.0 recommends a model for each legal task from the same results.
ISO/IEC 42001:2023 certification covers Harvey's AI management system, and the Statement of Applicability published with it sets out the controls in scope. Its AIUC-1 certification, an AI security and safety standard, was conducted by Schellman and, in Harvey's account, validates its adversarial testing and AI security program. The trust center adds IRAP attestation, an EU AI Act item and AI Governance, AI Monitoring and AI Overview documents, with an AI Acceptable Use Policy among its policies. An AI Policy dated 26 December 2025 commits to testing the accuracy, reliability and suitability of the AI in the service and to change management for continuing legal compliance. No accountable owner is named, and nothing is published about uneven output across matter types or populations.
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.
Customers set their own retention period. Deletion runs automatically when it ends and produces a verification log, and data can be exported at any time. After a subscription ends there is a 90 day export window, the customer chooses return or deletion, and deletion is the default. Data that cannot yet be deleted is anonymized or made inaccessible until it can be. HAQQ staff have no access to customer matters by default, and support access to a matter needs the customer's approval and is time limited. Book III of the terms serves as the data processing agreement. It lists subprocessor categories (hosting, storage, AI infrastructure, identity, payments, communications, analytics and security tools) and gives customers 15 days to object to a new one. The named subprocessor list is sent on request. Breaches affecting personal data are reported without undue delay, within 72 hours where feasible, and data is encrypted in transit and at rest.
Customers choose what to upload, set their own retention policies and can delete data at any time. Within 30 days of termination, Harvey deletes remaining customer data and content unless told otherwise, and the DPA extends return or deletion to every subprocessor. Access runs on role based controls, logical workspace separation, SAML SSO, audit logs and IP allow listing. The Subprocessor List, updated 21 September 2026, names each provider with its purpose and location, and subscribers receive at least 30 days' notice of an addition with a right to object. Under the Security Addendum, Harvey notifies the customer of a security incident without undue delay and within 48 hours. It then reports the incident's nature, the investigation and the mitigation, and keeps relevant logs for at least a year.
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.
The terms, last updated 29 June 2026 and governed by Delaware law, set out the allocation of loss. Each party's total liability in any 12 month period is capped at the fees paid or payable for the services behind the claim, and indirect and consequential losses are excluded. Fraud, willful misconduct, gross negligence, indemnity obligations, material breach of confidentiality and misuse of the other party's intellectual property sit outside the cap. Against claims that the services infringe third party intellectual property, HAQQ defends customers, excluding customer data and materials, third party integrations, unsupported changes and combinations it did not supply. It warrants that the services are provided in a professional and workmanlike manner and offers the rest as is. For AI output the terms are explicit. HAQQ is not liable for inaccuracies, hallucinations or bias, or for reliance without human review, and customers indemnify HAQQ for claims arising from their own professional services.
The Platform Agreement, last updated 9 January 2026, sets out what Harvey stands behind. Section 8 indemnifies the customer against third party claims that the service infringes intellectual property. If a claim arises, Harvey may substitute the service, secure a license to continue, or terminate and refund prepaid fees. The indemnity excludes materials Harvey did not provide and output from input the customer knew was likely to infringe. Harvey warrants that the service conforms to its documentation, is delivered in a professional manner and complies with applicable law. It also commits to carry industry standard insurance with insurers rated A- or better by A.M. Best. Total liability is capped at the greater of 12 months of fees or $250,000. Payment, indemnity and claims the law will not let a contract limit, such as gross negligence, sit outside that cap. Data breach and confidentiality claims carry a separate cap of twice 12 months of fees or $500,000. After a security incident, the Security Addendum commits Harvey to an independent forensic investigation at its own cost. The agreement states that output is AI generated and may contain errors.
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.
eFirm connects to Microsoft SharePoint, its only document storage integration today. With SharePoint linked, a Word file opens and edits in place from eFirm and saves back to both systems, keeping version history and tracked changes. Without it, files sit in eFirm's own storage and are downloaded, edited and uploaded again. Email is sent and received inside eFirm and logged to the right client and matter, and eFirm syncs with the Outlook calendar. NetDocuments and World-Check screening appear among eFirm's connections with no description of what moves between them. Legal AI Chat exports drafts to Word and PDF without a watermark on paid plans, and track changes editing with eFirm integration comes with the Pro plan. HAQQ also offers an API and an MCP server so other tools can call the engine, and payments run through its own eWallet.
Word, Outlook, Teams and SharePoint all run Harvey inside them, and Vault syncs documents from iManage, SharePoint and Google Drive. The iManage connection is a direct OAuth integration with an embedded web extension, and NetDocuments access follows the firm's existing permissions. Intake takes requests from email, Jira, Slack, ServiceNow, iManage, Google Drive and SharePoint and sends updates back to the source channel. Command Center syncs employee profiles through SCIM. Aderant and Ansarada integrations are documented in the trust center, and customers can build their own integrations, including through an MCP connector library. LexisNexis content reaches research through Ask LexisNexis.
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.
The same software runs three ways. HAQQ Cloud is operated by HAQQ in a region the customer picks and isolated per firm. A private cloud runs inside the customer's own tenancy with customer held keys. A third option is fully on premises. Hosting regions are the European Union, the United States and the Middle East. Default residency is global; the Business plan adds regional residency in the GCC, the EU or the United States, and Enterprise adds custom residency, on premises deployment and customer managed keys. Contractual residency is written into the order form. Coverage is kept separate from residency, since serving a country does not mean processing there, and cross border transfers may happen for support, security, billing, subprocessors and redundancy under standard contractual clauses or adequacy decisions. Legal AI Chat requests still go, with personal details replaced, to an outside model provider whose processing region is not given.
Microsoft Azure hosts Harvey's cloud environment, with each customer's data logically separated and access set by role. Customers can keep data in region in the EU or Switzerland, the United States or Australia. Customers with localization needs can also have processing in the EU and Switzerland or in Australia, and that commitment extends to Harvey's subprocessors. The Subprocessor List gives each model provider's processing locations. Where Anthropic or Mistral models are not offered in a chosen region, Harvey reaches them through hosting platforms such as AWS Bedrock and Google Vertex. What changes between plans or tiers is not stated.
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.
Controls are designed to align with the SOC 2 Trust Services Criteria, ISO 27001 and ISO 42001, and HAQQ meets GDPR obligations as a processor. It is explicit that this is alignment and not certification: it holds no SOC 2 Type II report and no ISO certificate. The SOC 2, ISO 27001 and ISO 42001 badges on the site are each marked compliant and not certified. Security questionnaires, evidence packs and audit requests go through a contact form, and HAQQ offers to confirm the exact scope in writing where a certificate is a procurement requirement. A responsible disclosure policy and a security.txt file are published. Book III gives customers an audit right at most once in any 12 month period, with the customer paying HAQQ's reasonable costs.
Schellman audits Harvey for SOC 2 Type II and ISO 27001 every year, and the 2026 audit and certification renewal are complete. The trust center also lists ISO 27701, ISO/IEC 42001, AIUC-1, IRAP, TX-RAMP, GDPR, CCPA and an EU AI Act item. It publishes Statements of Applicability for ISO 27001, 27701 and 42001. Schellman, NCC Group and Bishop Fox perform external audits, and third party penetration tests run every year. Reports, the CAIQ v4.0.3, SIG Core and SIG Lite questionnaires and the data flow and network diagrams are released through a self serve access request, with bulk download.
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.
A third party frontier model writes each final answer, and Justinian is shown routing across models from OpenAI, Anthropic and Google. The specific models in production are not named; the models in HAQQ's benchmarks are competitors, not its stack. Model choice is a configuration setting, so a change of model is not a rebuild. AI infrastructure providers are one of the subprocessor categories in Book III, which gives customers 15 days to object to a new subprocessor. The named list is sent only on request, and no notice covers a change of model within a provider. Under the terms, customer data, prompts and outputs may not train any third party model.
The Subprocessor List names ten AI providers with their processing locations, among them Microsoft, OpenAI, Google Cloud, Amazon Web Services, Anthropic and Mistral. ElevenLabs supplies voice services, DeepL machine translation, and Baseten and Fireworks.ai further model hosting. Claude and Gemini models run through Amazon Bedrock and Google Vertex AI, as Harvey's subprocessor FAQ explains. Model providers work under zero data retention, with no human review and no training on customer data. Under the DPA, customers who subscribe to the list receive notice at least 30 days before a new subprocessor is added and may object. Which model serves which feature is not stated.
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.
Prices are published in USD, and the self serve plans start without a sales call. Legal AI Chat is priced per user with a monthly allowance of credits. Free comes with 20 credits as a sign up gift, and Starter is $30 a month with 100 credits. Pro is $100 with 360, and Business is $300 with 1,080 plus regional residency and the firm wide Legal AI Twin. Top up packs run from $1 to $100 and do not expire. Larger packs and higher tiers buy more credits for the same money. eFirm is priced per seat with a two seat minimum, Boutique at $25 and Purple at $60 per user a month, with free trials of three to seven days. Annual billing gives two months free, students get 60 credits a month for $15, and Enterprise is quoted. The credits a request uses depend on the prompt, attachments and research depth, so its cost is not known before it runs.
Fees are set in each customer's order form and invoiced under the Platform Agreement. If an order form is terminated early under the agreement, Harvey refunds a pro rata share of prepaid unused fees. Undisputed late amounts carry a finance charge at the Federal Funds Rate plus 1.5% a year. No rate, plan, tier or unit of charge appears on Harvey's site, and buying starts with a contact sales form.
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.
Buyers range from lawyers and law firms to in house legal teams, founders, students and individuals. Enterprise is aimed at law firms, corporate legal teams and regulated businesses in banking, telecom and government. The prompt library spans 391 prompts across 23 practice areas, and use case pages cover contract drafting and review, litigation support, due diligence, and banking and finance. Jurisdiction depth is stated. Thirty three countries across the Middle East, Europe, Latin America and Asia have a maintained set of named statutes and codes, while more than 200 get general reasoning and drafting only. The interface runs in seven languages, with native right to left Arabic. eFirm shapes matter types and fields for litigation, arbitration and government procedures. Practice areas the product does not support are not listed, and HAQQ's own benchmark leans toward corporate and commercial work.
Large law firms, mid sized firms, in house legal departments and professional services firms are the stated buyers. Industry pages cover healthcare and life sciences, technology, energy and utilities, and media, entertainment and telecom. More than 200,000 lawyers at over 3,500 organizations in more than 85 countries use Harvey, including 80% of Am Law 100 firms and five of the Fortune 10. Solutions pages address litigation, transactional work and innovation teams, and research spans legal, regulatory and tax questions across jurisdictions. Firm sizes or practice areas the product is not built for are not stated.
The 12 legal signals, side by side
Recorded rather than graded. These are the questions a practitioner has to answer before a tool touches a client matter, and the answers are taken from public material only.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
The terms bar HAQQ from using customer data, matter data, prompts, inputs, outputs, user content, client data or support data to train, fine tune or improve any model. The ban covers foundation, cross customer and third party models, and lifts only with a separately signed agreement. Service improvement relies on anonymized, aggregated telemetry with no customer content. The Legal AI Twin in eFirm adapts to a firm's own knowledge inside that firm's tenant and feeds no other model.
Under the Platform Agreement, Harvey will not train any AI model on customer content or customer data, and its subprocessors will not either. The term covers both uploaded documents, which Harvey calls customer data, and the prompts and answers it calls customer content. A firm that asks for it can have a bespoke model trained on its own data and used only for that firm. Harvey may use usage data, such as feature and session statistics, to develop and improve the service, and usage data excludes customer data and content.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
Customers set the retention period, and purging runs automatically when it lapses, with a verification log. eFirm adds retention schedules and legal holds per document class. After a subscription ends the customer has 90 days to export, then data is deleted from production systems, subject to backups and legal holds. No zero retention setting is described for Legal AI Chat, and what the outside model provider keeps is not stated.
Customers decide what to upload, how long it is kept and whether it can be shared inside the organization, and they can delete data at any time. No default retention period is published. Model providers work under zero data retention, and subprocessors other than cloud storage providers may not retain or log content for human review. After termination, Harvey deletes remaining customer data and content within 30 days unless the customer instructs otherwise.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
eFirm keeps its own permission model. Read, add, update and delete rights are set per feature and scoped to everyone, no one, the records assigned to a user or the parent record, and every action is logged against a name. Each firm is a separate tenant, and HAQQ describes cross matter isolation and ethical walls across the platform. The firm configures these permissions in eFirm during onboarding; access rules are not inherited from a document management system.
NetDocuments permissions carry through, so a user sees only the cabinets, matters, folders and files they can already open, and Harvey does not change those permissions. The iManage connection is a direct OAuth integration that respects iManage permissions and ethical walls. Harvey also syncs the firm's own ethical walls provider and blocks walled users from reaching or sharing restricted content.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
The terms let HAQQ access matter data where law, a court order, a regulator, a sanctions obligation, a security investigation or legal process requires it. They contain no commitment to tell the customer about such a request or to resist it, and HAQQ publishes no transparency report. Book III does require HAQQ to warn the customer when an instruction would breach data protection law.
If law or a court order compels Harvey to disclose a customer's confidential information, the Platform Agreement commits it to advance written notice where legally permitted and to help seek confidential treatment. The DPA makes the same commitment for law enforcement requests for personal data. Where the law forbids notice, Harvey seeks a waiver and tells the customer once the prohibition ends. Its Law Enforcement Requests policy sends agencies to the customer first, challenges requests its lawyers consider unlawful and answers foreign requests only through channels such as a Mutual Legal Assistance Treaty. No transparency report 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?
The coverage list, last reviewed 26 August 2026, names the statutes and codes maintained for 33 countries. Examples are the French Code civil and Code du travail, Germany's BGB and HGB, and Lebanon's Labor Code, Civil Code and Rental Law. Justinian also draws on case law, gazettes, regulatory databases and the USPTO, EUIPO and WIPO registers, and can look things up on the web. Where the case law comes from, the license behind each source and how often the sources are updated are not stated.
Knowledge connects questions to legal databases, curated public sources and Harvey's data partners, more than 500 legal sources across jurisdictions. The Subprocessor List names RELX, doing business as LexisNexis, for Ask LexisNexis, and Parallel Web Systems and you.com for web browsing and knowledge sources. Harvey's AI Policy commits it to obtaining the rights and licenses its data sets require. No source list by jurisdiction, license basis per source or update schedule is published.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
Each claim in an answer carries a citation the user can open, and Justinian runs a citation audit and flags conflicting or ambiguous authority. Nothing describes a check of whether a cited case has been overturned or a provision amended since. The terms leave verification of citations and authorities to the user.
Knowledge answers from more than 500 legal sources and draws LexisNexis content through Ask LexisNexis. Harvey does not describe a treatment signal or a later history check on the authorities it returns.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
Justinian works inside legal domains and declines to speculate outside them. Where sources conflict or the law is ambiguous, it flags the uncertainty instead of choosing an answer, and legal claims are generated against retrieved sources so an unsupported claim is visible. HAQQ's benchmark marks hallucination as one criterion, but no published result isolates how often the engine declines to answer. Its engine page tells lawyers to test any legal AI on unsettled law, where a system built for professional work should say the ground is uncertain.
BigLaw Bench publishes hallucination rates and explains how Harvey defines and detects hallucinated claims, and agents cite a source for every claim. No setting or behavior is described for declining to answer when Harvey cannot ground a response.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
The AI Hallucination Cases database maintained by Damien Charlotin, which records court decisions worldwide that address hallucinated AI content and the tool involved where known, has no entry naming HAQQ or its Justinian engine. HAQQ's answers attach a citation to each claim, and its terms require users to verify citations before relying on them.
The AI Hallucination Cases database, maintained by Damien Charlotin, records court decisions worldwide that address hallucinated AI content and names the tool involved where known. It lists one matter naming Harvey, Hill v. Foundation Media LLC (S.D.N.Y., No. 25-cv-5947). On 22 September 2026 the court directed defense counsel to say whether AI was used on a brief with two questioned quotations and a citation, and a week later it took no further action.
Harvey is named only in counsel's own letter to the court, which says the brief was prepared with Harvey. No order makes any finding about Harvey or its output. A Galicia High Court order of 7 July 2026 lists Harvey AI among verified legal AI tools while setting a €1,800 fine, again with no finding about the product. Harvey's agreement states that output is AI generated and may contain errors.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
A HAQQ guide on AI ethics in law, published in May 2026, walks through ABA Formal Opinion 512 and the duties of competence, confidentiality, communication, candor, supervision and reasonable fees. It sets them beside UN and EU frameworks and the EU AI Act, and notes that the opinion asks lawyers to discuss AI use where a client would reasonably want to know. The terms leave bar, court and client consent obligations with the user. No mapping of the product to guidance from the Gulf, Levant or North African bars HAQQ serves is published.
Harvey's terms call the service a research tool whose output is not legal advice, and the agent pages leave the judgment on each task with the lawyer. No named ethics opinion appears, such as ABA Formal Opinion 512 or a state bar opinion.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
eFirm ties AI work to the matter. Chats, drafts and documents live on the matter record, time is tracked against the matter as the work happens, and invoices are built from the same record. Every action is logged to a named user and can be exported. HAQQ claims 70% less drafting time and a jump from one contract a day to ten or fifteen. It offers a billable hours calculator but no guidance on how firms should bill or disclose AI assisted work.
ROI calculators for law firms and in house teams sit on Harvey's site, and customer stories report time saved, such as Repsol's. Command Center reports adoption by user, practice group and role. Nothing records AI assisted work per matter for billing, or addresses how a firm bills for that work or discloses it to clients.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
Book III commits HAQQ to keep a list of material subprocessors with their function and processing region, and the named list is sent on request through the contact form. The subprocessor categories, the data processing terms and the security overview are published, and an executed DPA or standard contractual clauses can be requested. Model providers appear on the engine page without a list tied to the contract. Customers who need a certificate rather than a control set are asked to say so early, and HAQQ confirms the compliance scope in writing.
The trust center offers a Data Processing Addendum, a Business Associate Addendum and a Data Subject Requests item. It adds the CAIQ v4.0.3, SIG Core and SIG Lite questionnaires, data flow and network diagrams, a HIPAA report and a Security Welcome Packet. Sensitive items are released through a self serve request with bulk download. Statements of Applicability for ISO 27001, 27701 and 42001 show what each certification covers.
The public Subprocessor List names every provider with its purpose and location, and subscribers hear of additions 30 days ahead.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Each answer shows the sources behind it and the credits it used, chats stay on the matter, and the platform logs every action to a named user, with log exports depending on the subscription tier. No export ties a filing to the model used, the sources retrieved and the person who verified it, and no court disclosure template is published.
Audit logs are a standard enterprise control, agents log every step, and every claim links to the source passage behind it. Command Center reports usage by user and feature. No export brings together the model used, the sources retrieved and the human review for one document.
The questions both sides leave open
Derived from the records above rather than written, so it cannot favor either vendor. Take these into both conversations and ask each side the same question.
- Good Law Verification
Which one fits
Choose HAQQ if
- You practice under civil law or Sharia influenced systems. HAQQ detects the jurisdiction of each request and keeps named statutes and codes for 33 countries, many in the Middle East.
- You want practice management and AI in one system. eFirm holds matters, hearings, time and invoices, and the AI works on the firm's own records.
- You want to start without a sales call. HAQQ has a free tier and published per user plans for its chat and practice management products.
Choose Harvey if
- You are a large firm or a global legal department. Harvey sells to law firms, in house teams and professional services networks, with Spaces for work shared across organizations.
- Your documents live in iManage or NetDocuments. Harvey connects to iManage and NetDocuments, as well as SharePoint, Box, Word and Outlook.
- You want to measure AI adoption. Command Center gives leaders analytics, benchmarking and insights on how the organization uses Harvey.
In summary
HAQQ
HAQQ, from HAQQ Inc., is a legal AI platform built in Beirut for law firms, in house teams and individuals, with its deepest coverage in the Middle East and Europe. Legal AI Chat researches, drafts and reviews with a citation behind each claim, and eFirm runs matters, hearings, time, billing and client intake. According to the AI Legal Index, HAQQ publishes its prices, offers hosting in the EU, the United States or the Middle East, and runs its own benchmark across 19 systems. Its interface runs in seven languages with native Arabic.
Harvey
Harvey, based in San Francisco, sells enterprise legal AI to law firms, in house legal departments and professional services firms. Its platform includes Agents, Vault, Knowledge, Spaces, Command Center, Contract Intelligence, Horizon Scanning for regulatory change and a mobile app, and a Memory feature carries a user's preferences across products. According to the AI Legal Index, Harvey connects to iManage, NetDocuments, SharePoint, Box, Word and Outlook and publishes BigLaw Bench, its own public evaluation framework.
Questions buyers ask
What does each sell besides AI?
Harvey sells AI products, from agents and research to bulk review, contract intelligence and regulatory horizon scanning. HAQQ also sells eFirm, practice management with matters, time tracking, invoicing, a financial ledger and a client portal. From the AI Legal Index, based on each vendor's own published materials as of October 12, 2026. No vendor pays for placement.
Which jurisdictions does HAQQ cover in depth?
HAQQ publishes a coverage list with named statutes and codes for 33 countries and general reasoning for more than 200, and runs in seven languages with native Arabic. Harvey's Knowledge product researches legal, regulatory and tax questions across domains for global firms. From the AI Legal Index, based on each vendor's own published materials as of October 12, 2026. No vendor pays for placement.
What does each publish about testing?
Harvey publishes BigLaw Bench, its own evaluation framework, and research notes on models and benchmarks. HAQQ publishes a task benchmark of 19 systems across 11 legal task categories and a cross domain study, and lists that benchmark's limits. From the AI Legal Index, based on each vendor's own published materials as of October 12, 2026. No vendor pays for placement.
Does HAQQ or Harvey publish plan prices?
HAQQ starts free and publishes per user plans with monthly credits, plus per seat pricing for eFirm. Harvey is sold through its sales team and publishes no price. From the AI Legal Index, based on each vendor's own published materials as of October 12, 2026. No vendor pays for placement.
Who are their typical customers?
Harvey's buyers are law firms of every size, in house departments and professional services networks, with industry pages for healthcare, technology and media companies. HAQQ quotes a partner at Yakan Law Firm and the president of the Beirut Bar Association, has signed a partnership with the New Cairo Bar Association, and reports more than 17,000 legal teams. From the AI Legal Index, based on each vendor's own published materials as of October 12, 2026. No vendor pays for placement.
Neither vendor paid for inclusion, placement or a grade, and neither reviewed this page before it published. Everything on it comes from public material on the dates shown. How the index grades.
Harvey is an independent company that announced a $550M funding round. HAQQ is made by HAQQ Inc., which announced a $3M seed round. HAQQ publishes its own comparison of the two products. Neither vendor reviewed this page.