Harvey vs NetDocuments: how they compare in 2026

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.

NetDocuments is a cloud document and email management system sold for twenty five years. The ndMAX suite adds a Legal AI Assistant, AI Search, Tabular Review, an App Builder and a library of ready made legal apps. Smart Answers, which began rolling out to ndMAX Enterprise customers on 31 March 2026, answers questions from the firm's repository with citations. The vendor frames this as bringing AI to the content rather than taking content to AI. Its press material says more than 800 firms began using the AI capabilities in 2025 and over forty percent of new customers choose AI at purchase. Without the models, the document management, email management, search, security and collaboration product remains, which is what most of the 7,000 organizations bought.

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. Through Ask LexisNexis, cases are validated with Shepard's Citations and the Shepard's Signal is shown beside them. Harvey's own United States case law source, more than 9 million CourtListener opinions, runs on Harvey Graph, which maps who cites whom and how each case was treated, so an agent checks treatment before it answers. Harvey says precedent flags and cite checking for that source will follow.

The Legal Context Engineering Benchmark, published 18 August 2026 by the VP of Product, reports quality scores out of 100 for the vendor's own AI answers across three model tiers. Without the Legal Context Graph they are 64.0, 67.4 and 73.9; with it, 65.1, 69.0 and 75.4. Correct answer counts out of 300 and the cost of a correct answer are also given. The test set has 300 questions across ten real matters, five transactional and five litigation, built from public filings and court dockets partly outside the models' training cutoff. It covers 874 documents and about sixty million characters, with nothing synthetic, six named question types including Unanswerable, and a rubric for each question. Citations are checked automatically against the documents, with human spot checks. A worked example shows the product's own answer failing all three criteria without the context layer and still failing one with it. The method is published for firms to rerun. It is the vendor's own benchmark, not third party validation, and an appendix with per matter scorecards is published with it.

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. Legal Work Orchestration, launched in October 2026, sets defined review gates inside a matter. Each task carries its scope, inputs and reviewer, and when an agent finishes a review Harvey notifies the designated partner. The task completes only once that partner approves it, and the next step then begins. Intake applies routing rules the legal team defines to send each request to an approved agent, to Contract Intelligence or to a lawyer. In the Connector Library, admins choose which tools are enabled, who may use them and which require human confirmation before they run. Reports and reviews can run on a schedule, and admins govern access through roles and workspace permissions. What happens to an agent's work when a reviewer rejects it, and what Harvey commits to when that work is wrong, are not stated.

The vendor says the best AI apps run automatically in the background based on a user's actions, and MCP connectivity lets external agents run workflows across systems. On the oversight side, answers carry citations a lawyer can open, and Exhibit B commits to document level activity logs reviewable for ninety days. Administrators control which users and applications reach what through the Customer Service Account. The benchmark report tells readers to read some of the answers themselves before believing any of the numbers. Nothing on the Legal AI page, in the benchmark report or in the agreement says the models work alongside a supervising lawyer, names the point at which an app acts without one, or says what happens when an output is wrong. The Legal AI page also carries an Our Approach to AI video.

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, and The Adecco Group's general counsel says more than 80% of its legal professionals report real gains in quality and speed. In October 2026 CVS Health expanded Harvey across its corporate affairs functions after a pilot, with the in house legal team using it for document review, research, drafting and knowledge retrieval. Cozen O'Connor and Spencer West chose Harvey in November 2026. Other named customers include A&O Shearman, Gleiss Lutz, Cuatrecasas, Macfarlanes, KKR, PwC, Deutsche Telekom, HubSpot, Bridgewater and Carvana. The figures are usage counts and time savings as each customer reports them, with no stated method for how they were measured.

On the Legal AI page, Jared Gullbergh of Buchanan says a single AI app was calculated to save over 1,500 hours of work a year while producing better work product. Charly Duffy of Coghlan Duffy speaks on the App Builder, and Jeff Sabado, Director of Knowledge Management at Davis Wright Tremaine, says the tools already respect the firm's client and matter based security and ethical walls. Company figures are published without reference to any deployment: more than 7,000 organizations served, more than 800 firms starting to use the AI in 2025, over forty percent of new customers choosing AI at purchase, and more than 150 integrations. No deployment is dated, and the basis for the 1,500 hour figure is not published.

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.

Under the agreement, Customer Repository Data is confidential and NetDocuments disclaims ownership of it (5.2.1). Use is limited to providing the Services, supporting third party applications the customer selects, help desk support the customer starts, and legal demands (5.2.2). Ethical walls are a documented product with access controls at user, document and workspace level, and the vendor says the AI works inside a firm's existing permissions and walls. The ndMAX FAQ says inputs and outputs are not kept beyond producing the answer and are not used to train the model. The position on the model provider is explicit, including an agreement with Microsoft that keeps customer content out of abuse monitoring. Privilege and work product are not addressed in the agreement or on the product pages. The training statement covers the LLM rather than NetDocuments' own use.

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.

The product's AI apps analyze NDAs, assess contract risk, produce judge analytics and answer questions about a matter. The Software as a Service Agreement covers use restrictions, AI restrictions on the customer, warranties and liability without touching advice, competence or supervision. Section 9.4 disclaims that results will meet any requirement, which is a warranty disclaimer, not a statement of what the product is and is not. The Legal AI, App Builder and Ethical Walls pages carry no disclaimer. The audience is clear: every segment page addresses law firms, corporate legal departments and public sector legal teams, and the agreement limits use to internal business purposes by named users, so there is no consumer surface. NetDocuments also publishes Terms of Use.

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.

The Legal Context Engineering Benchmark is a signed technical report by the VP of Product describing what is tested and how. It includes a test harness that compares the two tool setups before a run and refuses to proceed if they differ in anything but the capability under test. Citations are checked automatically against the source documents, with human spot checks, and the method is published for others to apply. The report also states its own limits. It does not yet measure change over a matter's life, it withholds the context build cost because that varied about sixfold across ten matters, and it was drafted with AI assistance. No responsible AI framework, policy or standard such as ISO 42001 is published. Nobody is named as accountable for AI governance, as distinct from authoring the benchmark, and nothing addresses uneven output across matter types or populations. An appendix PDF with per matter scorecards accompanies the report.

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.

Under Exhibit B of the Software as a Service Agreement, NetDocuments must report confirmed Unauthorized Access to the customer without unreasonable delay and within 24 hours, with a root cause analysis and mitigation plan. Customer Repository Data is encrypted in transit and at rest, and access follows least privilege, removed when employment ends. Independent penetration tests run at least twice a year, with monthly internal and external vulnerability scans. Document and administrative activity logs are reviewable by the customer for ninety days, and NetDocuments keeps log information for at least a year. Data is stored across two or more geographically separate data centers, with a two hour recovery point and six hour recovery time. Under section 7.4 the customer is responsible for export, and NetDocuments may destroy remaining data thirty days after the term ends. Section 12.2 permits Common Subcontractors and promises notice of material changes without naming them, so no subprocessor list is published; Microsoft, for the AI, is the only processor named anywhere.

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.

Section 11.1 of the Software as a Service Agreement gives a defense and indemnity against third party claims that the Services infringe intellectual property, with three named exclusions and modify, license or terminate remedies at 11.3.3. Section 10.3 caps each party at the annualized subscription fee paid or payable at the time of the event, with indemnities carved out of the cap. Section 10.1 excludes indirect and consequential loss, and 10.4 preserves liability for death or personal injury caused by negligence and for fraud. Section 9.2 warrants professional and competent performance, and Exhibit A publishes a 99.9 percent availability target with graduated service credits. No insurance is stated. Section 10.1 also removes every liability limit for a customer's breach of the AI Restrictions in section 3.5, so NetDocuments may recover lost profits and all indirect and consequential damages without a cap, while the customer's own recovery stays capped at a year of fees.

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. Since October 2026, Harvey also applies the ethical wall policies kept in iManage Security Policy Manager to Threads, Vaults, Review Tables and Spaces, with Security Policy Manager remaining the system of record. 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. The Connector Library, generally available since October 2026, connects Harvey through MCP to tools such as Ironclad, Docusign, Intapp, Aderant, Everlaw, LawVu, Box, Gmail, Slack and Jira, and customers can add their own MCP connectors. Admins choose which connectors are enabled, who may use them and which tools need human confirmation, and each connector passes Harvey's security review before launch. LexisNexis content reaches research through Ask LexisNexis.

The vendor says the platform integrates with more than 150 technologies, including Microsoft 365, Teams and DocuSign, and publishes an App Directory and partner locator. It ships ndOffice and mobile access software defined in the agreement. In 2026, Model Context Protocol connectivity let external AI applications and agents reach NetDocuments content and run workflows without downloads or manual transfers. The vendor names ChatGPT, Claude, Claude Cowork and its legal plugin, Microsoft Copilot, Perplexity, Google Gemini Enterprise, Harvey and Legora, and says this works within existing permissions, ethical walls and audit controls with no content leaving the platform. The agreement sets out API access: it is not included in the subscription, is granted on request after registering an application, may not be shared, and may carry further terms. The Legal AI page, the press material and the agreement do not say what moves in which direction for each integration, and NetDocuments also publishes support documentation.

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, and sign in through a separate US, EU or AU login. 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. Federal customers can use Harvey within Knox Systems' FedRAMP Class D certified cloud environment. What changes between plans or tiers is not stated.

Services Region is a defined term in the Software as a Service Agreement. Section 2.4 commits NetDocuments to store Customer Repository Data in the region on the Order Form and not to transfer or access it except at the customer's or a user's direction or as law requires; sharing and access by users outside the region are excepted where the customer sets them up. Regional entry points cover the United States, United Kingdom, Germany, Australia and Canada, and the contracting entity follows the region. Tenancy is described as the customer's private repository on the ND Network. Customer Managed Storage Locations using ndFlexStore or ndMirror let a customer keep repository data on servers NetDocuments does not host. Processing is addressed separately from storage: section 4.2 says help desk staff or subcontractors may work from regions outside the customer's Service Region, and the customer must say in advance if its policies or law restrict that. Where Azure OpenAI inference runs for a customer in a given region is not published.

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 Level 2, GDPR, CCPA and an EU AI Act item. It publishes Statements of Applicability for ISO 27001, 27701 and 42001. In October 2026 Harvey announced FedRAMP Class D certification, operating within Knox Systems' FedRAMP Class D certified cloud environment for federal agencies and contractors. 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.

Exhibit B of the Software as a Service Agreement commits NetDocuments to an annual SOC 2 Type 2 audit for security, availability and privacy and to maintaining ISO 27001 certification or a comparable successor. It says all data centers fall within the annual ISO 27001 audit and that the listed controls are verified in both. No auditor, audit period or report date is published. The agreement names a trust website, trust.netdocuments.com, a TrustShare portal offering Due Diligence Response material.

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. Harvey Tenet, in research preview since August 2026, is Harvey's own open weight model, post trained from Moonshot's Kimi K3 with Fireworks for long legal tasks. Harvey says it was trained on synthetic, public legal and expert data, with no customer data. Where Tenet runs and which customers can use it are not stated. 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.

The Legal AI page says ndMAX is powered by Microsoft Cognitive Services and Azure OpenAI and their suite of OpenAI models. Only inputs and outputs are shared with the LLM, they are not kept beyond the exchange, and an agreement with Microsoft keeps customer content from triggering Microsoft's abuse monitoring. The same FAQ reserves the right to change what sits underneath. The solutions are described as model agnostic, so better models may be used as they are developed, after internal testing. No specific model is named and no inference location is stated. The only notice mechanism in the agreement is the general notice of material changes to Common Subcontractors in section 12.2, which is not a commitment about models. MCP connectivity, which lets a firm's own ChatGPT, Claude, Copilot or Harvey reach NetDocuments content, reflects the customer's model choice, not the vendor's supply chain.

Licensing is per seat. The help center FAQ, updated November 6, 2026, puts the average cost at several hundred dollars per user per month, and the price depends on the number of seats, the products deployed and the contract terms. No rate card, plan or tier is published, and buying starts with a demo and a quote from sales. Ask LexisNexis, the research add on, is licensed separately from the base service. 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.

Every purchase path on the site ends at Contact Sales or a demo request, and no price, band or unit figure appears on the site. The agreement supplies the structure. Fees are set in an Order Form as recurring Subscription Fees with a Billing Commencement Date. Under section 4.2, a customer exceeding the user, lawyer, document or other license metrics in the Order Form is invoiced for the overage. Section 4.4 caps annual increases at CPI plus five percent, no more than once a year, section 8.1 requires disputed invoices to be raised within thirty days, and Exhibit A sets service credits as a share of the annual fee. Tier names appear in product material, with ndMAX Enterprise named as the plan that receives Smart Answers and expanded MCP connectivity. The unit, the cap on increases and the plan structure are published before a sales call; the price is not.

Large law firms, mid sized firms, in house legal departments and professional services firms are the stated buyers. Since October 2026, federal agencies and contractors can also use Harvey within a FedRAMP Class D environment. Industry pages cover healthcare and life sciences, technology, energy and utilities, media, entertainment and telecom, and the public sector. 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, and its in house page counts more than 1,000 enterprises. 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 site has separate pages for small, midsize and large law firms, corporations, in house legal departments, public sector, federal government, and state and local government and higher education. Regional pages cover Asia, Australia and New Zealand, North America, South America, and the UK and Europe. The vendor says it supports more than 7,000 law firms, corporate legal departments and public sector organizations. Practice coverage is described through tasks rather than areas. The ndMAX Studio library covers automatic profiling, NDA analysis, contract risk analysis, playbook creation and application, deposition preparation, judge analytics, offer letter workflows and translation. The benchmark is split evenly between transactional and litigation matters, the closest thing to a statement of where the AI has been tested. No practice area, jurisdiction or language is named as out of scope, and the eDOCS line for on premises customers has no stated boundary with the cloud product.