Eudia vs Harvey: how they compare in 2026
Eudia and Harvey both sell an AI platform to the in house legal department, and Eudia positions itself directly against Harvey for that buyer. The difference is certification and measurement. Harvey holds ISO/IEC 42001 with a published statement of applicability, and an AIUC-1 certification conducted by Schellman. It names Schellman again as the auditor behind its SOC 2 Type II and ISO 27001, and names NCC Group and Bishop Fox as its penetration testing and red teaming partners. Its accuracy has also been measured by a third party. In the February 2025 Vals Legal AI Report it scored between 65.0 and 94.8 percent across six tasks and beat the measured lawyer baseline on five of them. Eudia answers on customer evidence and architecture. It names Duracell, Graybar and Cargill with a figure against each, and it holds no repository of its own, connecting instead to the systems where the work already sits.
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.
Eudia has no conventional product underneath its AI. It holds no repository, matter management or contract lifecycle system of its own. Records stay in Google Drive, SharePoint, Box, Salesforce, DocuSign, Onit and LinkSquares, and Eudia connects to them. What it sells is the intelligence layer, turning a company's own data and know how into legal agents, marketed as expert digital twins and an Enterprise Brain. Remove the models and nothing is left to sell. One qualification: delivery is partly human. Forward-Deployed Engineers is one of the seven listed solutions, and the separate Eudia Counsel arm runs on the same platform. Neither replaces the models as what the buyer pays for.
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.
One grounding claim is made, with nothing behind it. Eudia's home page says it connects systems, documents and workflows so every decision is grounded in current, company specific context. No retrieval method, source linking, accuracy figure, test set or evaluation is published. Eudia's trust center lists AI Security and AI Risk Management documents behind an access request.
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.
Eudia is explicit about autonomy and says little about oversight. It sells Decision Self-Service, described as delivering governed intelligence across the enterprise, and says the Enterprise Brain makes the business self sufficient without adding headcount. That means business users getting legal answers without the legal team involved. Oversight appears only as adjectives: intelligence is described as secure, governed and auditable. Nothing names a review point, approval step, confidence threshold, a point where an agent stops, or what happens when an agent is wrong. Human in the loop appears as a phrase, not a described control.
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.
Three large customers are named, each with a figure and its own case study. Duracell reports 50 percent savings on contracting costs using the AI Contracting MIND. Graybar reports 98 percent faster diligence, with contract review falling from four hours to five minutes, using the AI M&A MIND. Cargill reports a 50 percent cut in contracting research time using Contract Insights, and its Chief Legal and Compliance Officer, Rishi Varma, is quoted by name and title on the home page. These are Fortune 500 legal departments, which is real deployment evidence. The success stories index gives no dates and no basis for any of the three figures.
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.
How client confidences are handled goes unstated before signing, for a platform sold to Fortune 500 legal departments. No customer agreement, master subscription agreement or data processing addendum is published. The only agreement on the site is a website terms of use. The privacy policy says expressly that it does not cover personal data stored for customers, which is governed by contracts with each customer. So every key question is unanswered in public: there is no training commitment for customer content, no privilege or work product treatment, no matter or tenant separation, no retention or deletion terms, and nothing on what the named model providers may keep. Related documents exist behind an access request on the trust center: a Data Protection Policy and an AI Training Data and Bias document.
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.
The only disclaimers sit in the website terms of use and apply to the website. Section 5.1 says site content is for information and that the user alone must verify its accuracy. Section 11 disclaims all warranties for the Site and the Technology that runs it. Nothing says platform output is not legal advice, and nothing addresses supervision, competence or jurisdiction limits. The marketing, meanwhile, describes the product in decision terms: Decision Self-Service delivering governed intelligence across the enterprise, agents encoding expert judgment, and an Enterprise Brain that makes the business self sufficient. That points at business users outside the legal department, so the question of who relies on the output is live. No bar or ethics guidance is engaged.
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.
ISO/IEC 42001:2023 is listed as a current certification in the public section of Eudia's trust center, alongside SOC 2 Type 2, ISO/IEC 27001 and GDPR. That is an independently audited standard for managing AI, not a self written principles page. The trust center also has an AI section listing AI Training Data and Bias, AI Security and AI Risk Management, so Eudia has prepared bias material; its contents sit behind an access request. Nothing public names a person or team accountable for model behavior, and no test result or bias finding is disclosed.
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.
Of the five areas that matter for data stewardship, Eudia publishes one. Its subprocessor list is public and specific, naming Elastic, Amazon Web Services, Anthropic, OpenAI and Microsoft Azure. Retention and deletion of customer content are addressed nowhere. The privacy policy's retention section covers personal data Eudia holds for itself, says only that it is kept as long as necessary, and expressly excludes data held for customers. No incident or breach notice commitment is published, and no access control detail beyond the fact of SOC 2 Type 2 and ISO 27001. Eudia's trust center shows the material exists: it lists Data Breach Notifications, Certificates of Destruction, Data Asset Classification, Access Monitoring, a Data Protection Policy and a Backup Policy, all behind an access request.
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.
Liability terms reaching Eudia's product are not published. Its only published agreement is a website terms of use, which by its first sentence governs content at www.eudia.com. It caps total liability at five hundred dollars at section 13, has an indemnity running only from the user to Eudia at section 12, and disclaims all warranties at section 11. Section 5.2 says access to specific areas of the Site is subject to a further agreement called an Annex. That is where platform terms sit, and it is not published, so no indemnity, cap, exception or output warranty for the Enterprise Brain is public. Eudia lists cyber insurance as a document on its trust center, behind an access request. The five hundred dollar cap governs the website, not the platform.
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.
Seven integrations are named, with nothing on what they move. They are grouped by function: document storage, contract management, revenue and business systems, and signing. The seven are Google Drive, SharePoint, Box, Onit, Salesforce, LinkSquares and DocuSign, including two legal tools in Onit and LinkSquares. Each has one line of marketing describing an outcome, such as the Salesforce connection linking legal workflows with deal and customer data. Nothing says what syncs, in which direction, on what trigger, or what a customer must set up. No developer documentation, API reference or implementation guide is published.
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.
Delivery is clearly cloud based, but neither region nor tenancy is stated. Eudia's trust center names Azure for infrastructure and the subprocessor list adds Amazon Web Services, so the hosting providers are known. That identifies infrastructure, not where a customer's data sits. No region or residency option is published. The privacy policy says only that Eudia is based in the United States, works with service providers in the US and other countries, and may store information outside the reader's home country. No tenancy model is stated: nothing describes the platform as shared, single tenant or privately deployed.
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.
The trust center is public and detailed, but its reports sit behind an access request. A SafeBase trust center at trust.eudia.com lists four current standards by name: SOC 2 Type 2, ISO/IEC 27001, GDPR and ISO/IEC 42001:2023. It lists thirteen documents, including a SOC 2 Report, a Pentest Report, a Security Whitepaper, a Network Diagram and a Data Flow Diagram, plus third party security ratings from SecurityScorecard and Qualys SSL Labs. The list is public, but the reports sit behind a Get access flow, which the portal calls starting a security review and requesting access, without saying whether access comes after an email and a click through or through a sales conversation. No auditor is named and no coverage period appears on any public page.
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.
Model providers are named publicly, but the models are not. The trust center's subprocessor entry lists Anthropic and OpenAI as model providers, alongside Elastic, Amazon Web Services and Microsoft Azure for infrastructure, and can be read without an access request. The models themselves are not named, only the companies that make them. Nothing commits to telling customers when the supply chain changes, and no change notice term is published, since no customer agreement or data processing addendum is published at all. Where the models run is not stated.
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.
Pricing is absent at every level, down to the unit of charge. There is no pricing page in the navigation, footer, solution pages, government page, integrations page, success stories, terms or privacy policy, and the only commercial route is to schedule or book a demo. Solution areas are named, and so are three products, the AI Contracting MIND, the AI M&A MIND and Contract Insights. But a product list is not a price list: no tier, seat or usage unit, term or figure appears.
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.
Eudia describes its coverage by practice area and buyer, but not where it stops. It names five solution areas: marketing compliance, contracting and deal speed, self service decisions, mergers and acquisitions, and litigation. It separates an enterprise track from a government track built for government acquisition teams. Named customers at the top of the market, Duracell, Graybar and Cargill, support the enterprise claim. The boundary is left open. Nothing says what the platform does not support, no minimum company size is given, and no jurisdiction or geography is named. The link between the platform and the separate Eudia Counsel services arm is not explained, so where the software stops and the service begins is not clear.
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?
Nothing published addresses whether customer content trains models, in either direction. The quoted line is real, but it sits in the privacy policy's table of legal bases for European residents, and that policy says at the outset that it does not apply to personal information stored or maintained on behalf of customers, which is governed by contracts with each customer. So Training AI System is disclosed as a processing purpose for the data Eudia holds as controller, which is website, marketing and account information, and no improvement right over customer content is granted in anything published.
No customer agreement or data processing addendum is published. An AI Training Data and Bias document on the trust center sits behind an access request and was not available to read.
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?
No public material says how long prompts, outputs or uploaded documents are kept. The privacy policy has a retention section, but it governs personal information Eudia holds as controller, says only that data is kept as long as necessary to fulfill the purpose of collection, and expressly excludes data held on behalf of customers. Eudia's trust center lists Certificates of Destruction and a Backup Policy, which shows a retention and destruction practice exists, behind an access request.
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?
Nothing addresses walls, matter separation or whether retrieval respects the access model of the systems it draws from. That matters because the platform indexes across Google Drive, SharePoint, Box, Salesforce, Onit and LinkSquares, so whose permissions govern at query time is a live question, and the marketing answers it only with the assertion that intelligence is secure, governed and auditable, meeting enterprise requirements for access control.
Eudia's trust center lists Environment Segregation under infrastructure, which concerns separation of environments rather than of matters or users, behind an access request.
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?
Eudia's privacy policy says it may disclose data when compelled, and says nothing about telling the customer. It says Eudia will use personal information as it believes necessary to comply with lawful requests and legal process, including subpoenas, and may disclose data to law enforcement, government authorities and others in response to valid requests by public authorities. No commitment to notify the customer, and no reserved discretion over notice, appears.
The policy covers data Eudia holds for itself and expressly excludes customer content, for which no agreement is published.
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?
No corpus is identified. Eudia works on the company's own contracts, policies, precedent and know how, drawn from its existing systems, not on a published body of law, so questions about coverage and title to a legal corpus do not arise in the usual way. Nothing suggests the product retrieves primary law, and no statement of jurisdiction coverage is published.
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?
No public material addresses subsequent history, treatment flags or citator coverage, including on the litigation solution page. The product is built on enterprise documents rather than primary authority and claims no citator function, and no weaker form of checking is described.
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?
Nothing describes what the platform does when it cannot ground an answer. No abstention path, no way of saying it has no answer, and no confidence or grounding score is documented. The only related statement is that decisions are grounded in current, company specific context, which asserts grounding without saying what happens when the context lacks the answer.
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, and reporting based on it, records no court order, opinion or disciplinary record naming Eudia or its corporate name, Cicero Technologies. The same company operates a legal services arm and law firm under the Eudia Counsel brand, which this record does not cover.
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?
No public material engages with ABA Formal Opinion 512, any state bar opinion, or any other named ethics guidance, and nothing addresses professional responsibility even in general terms. That stands out against the product's positioning, which has business users outside the legal department getting legal answers through Decision Self-Service.
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?
Eudia publishes savings claims: 50 percent savings on contracting costs at Duracell, 98 percent faster diligence at Graybar, and a 50 percent cut in contracting research time at Cargill. The Duracell story describes a shift from billed hours to billed outcomes, which concerns how the customer pays for legal work, not how AI work is recorded or disclosed, and no record of AI work per matter is described. Eudia's buyer is the legal department that receives bills, not the firm that sends them, so the savings it markets fall on outside counsel spend and internal capacity.
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?
Eudia publishes a current subprocessor list on its trust center, readable without an access request. It names Anthropic and OpenAI as model providers, alongside Elastic, Amazon Web Services and Microsoft Azure for infrastructure, so a legal department asked which providers see its content can answer. Material for clients that can be forwarded is missing: the Data Processing Agreement, Security Whitepaper and Data Protection Policy all appear on the trust center behind a Get access flow, and none is published or available without a request.
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?
Nothing addresses judicial standing orders, disclosure of AI use or verification certification, including on the litigation solution page, and no exportable record of model use, sources retrieved or human review is described. The marketing says intelligence is auditable, but no audit record is documented and nothing indicates one could be produced for a document. The platform is sold to in house departments, so a filing disclosure duty would usually fall on the outside counsel handling the matter rather than on the buyer.
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.
- Commercial Transparency
- Good Law Verification
- Refusal and Uncertainty Behavior
Which one fits
Choose Eudia if
- You want named enterprises with figures attached rather than a logo wall. Eudia publishes Duracell at 50 percent savings on contracting costs, Graybar at 98 percent faster diligence with contract review falling from four hours to five minutes, and Cargill at a 50 percent reduction in contracting research time, with Cargill's chief legal and compliance officer quoted by name and title.
- You are not moving your systems of record. Eudia holds no repository, no matter management and no contract lifecycle system of its own, and connects instead to Google Drive, SharePoint, Box, Salesforce, DocuSign, Onit and LinkSquares, so the documents stay where they are and what the platform adds is the layer over them.
- You need to name the model providers today, without asking. Eudia's trust center renders publicly and its subprocessor list is readable without completing any access request, naming Anthropic and OpenAI as model providers alongside Elastic, Amazon Web Services and Microsoft Azure, with SOC 2 Type 2, ISO/IEC 27001, ISO/IEC 42001:2023 and GDPR listed as current.
Choose Harvey if
- You want the AI itself certified and adversarially tested. Harvey holds ISO/IEC 42001 with a published statement of applicability identifying which controls apply, and an AIUC-1 certification conducted by Schellman that the vendor describes as validating adversarial testing of its AI security program, with EU AI Act conformity listed separately.
- You want accuracy measured by somebody else. Harvey Assistant took part in the February 2025 Vals Legal AI Report, an independent benchmark against a measured lawyer baseline, scoring between 65.0 and 94.8 percent across six tasks and exceeding the baseline on five of them. It separately publishes BigLaw Bench with task categories and grading rubrics on a public repository.
- Your firm's walls have to hold when the AI reads. Harvey states that it follows existing NetDocuments permissions so a user sees only what they can already open, documents the iManage connection as a direct OAuth integration respecting permissions and ethical walls, and publishes admin documentation for syncing the firm's own ethical walls provider.
In summary
Eudia
Eudia is an AI platform for large in house legal departments. It codifies a company's own contracts, policies and institutional knowledge into what it calls an Enterprise Brain and runs legal agents on top of it. The agents cover marketing compliance, contracting, decision self service, mergers and acquisitions and litigation, with a separate track for government acquisition teams. According to the AI Legal Index, Eudia holds no repository or matter system of its own and connects to the systems the work already sits in. Duracell, Graybar and Cargill are published as named customers with a figure against each. No customer agreement, training position for customer content or price is published.
Harvey
Harvey is an enterprise legal AI platform for law firms, in house departments and professional services firms. Its agents plan and run multistep work for a lawyer to approve, Vault reviews documents in bulk, and Knowledge researches with citations across more than 500 legal sources. According to the AI Legal Index, Harvey holds ISO/IEC 42001 with a published statement of applicability alongside an AIUC-1 certification conducted by Schellman. Its accuracy has been measured independently in the February 2025 Vals Legal AI Report. It documents its connections into the systems a firm already runs, including an iManage integration that respects permissions and ethical walls. A published Platform Agreement carries an intellectual property indemnity and a liability cap, and no price appears at any level.
Questions buyers ask
Eudia vs Harvey: which is better for an in house legal team?
The difference is certification and measurement rather than ambition. Harvey publishes two AI specific certifications, named auditors and independent accuracy figures. Eudia publishes three named Fortune 500 customers with figures attached and a subprocessor list a buyer can read today. Neither publishes a price, and only Harvey publishes liability terms that cover its platform. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
Has either had its accuracy measured independently?
Harvey has. Harvey Assistant participated in the February 2025 Vals Legal AI Report, an independent benchmark measured against a lawyer baseline, scoring between 65.0 and 94.8 percent across six tasks and exceeding that baseline on five of them. It also publishes its own BigLaw Bench framework with task categories and grading rubrics. Eudia makes one grounding claim, that decisions are grounded in organization specific context, and publishes no accuracy figure, test set or evaluation. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
Do either say whether client content trains their models?
Harvey's Platform Agreement bars Harvey and its subprocessors from training any AI model on customer documents, prompts or answers, and its model providers work under zero data retention. On Eudia, nothing addresses customer content either way: its privacy policy lists training AI systems as a processing purpose, and that policy states expressly that it does not apply to information held on behalf of customers. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
Which one respects your document management permissions?
Harvey does, and documents it. It states that it follows existing NetDocuments permissions, that a user sees only the cabinets, matters, folders and files they can already access, and that it does not expand or modify permissions, with the iManage integration documented as respecting permissions and ethical walls and a separate admin route for syncing a firm's walls provider. Eudia indexes across Google Drive, SharePoint, Box, Salesforce, Onit and LinkSquares, and nothing published states whose permissions govern at query time. From the AI Legal Index, based on each vendor's own published materials as of September 3, 2026. No vendor pays for placement.
What do Eudia and Harvey both leave unpublished?
Neither publishes a price, a tier structure or a unit of charge. Neither names an ethics opinion, including ABA Formal Opinion 512. Neither documents what the product does when it cannot ground an answer. And neither addresses legal professional privilege or work product directly. From the AI Legal Index, based on each vendor's own published materials as of September 3, 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.
Liability terms read differently on the two sides. Harvey's Platform Agreement, last updated 9 January 2026, carries an intellectual property indemnity and caps liability at the greater of a year's fees or $250,000. Eudia's only published agreement is a website terms of use. By its own first sentence it governs the website, it carries a $500 cap, and it points platform terms to an unpublished annex. Eudia also lists a data protection policy, an AI training data and bias document and a cyber insurance certificate on its trust center behind an access request, so their contents are not public. Neither vendor reviewed this page.