Doctrine vs Ordalie: how they compare in 2026

D
Doctrine profile
O
Ordalie profile
Last verifiedNovember 16, 2026
Verdict

Doctrine and Ordalie are both Paris legal AI platforms that answer questions on French law with sources a lawyer can open, and both work inside Microsoft Word. They differ in what sits underneath and in how they are bought. Doctrine is the larger research database, a Legal Graph of 80 million interconnected French legal documents, and since September 2026 it has belonged to RELX's LexisNexis. Its Assistant agent crosses a client file with case law and statutes, and paid modules handle litigation files, contracts, employment terminations and tax. Ordalie is built as an AI workspace, with 23 connectors including iManage, workflows for timelines and contract review, and a free plan. Its PRO plan costs €75 per user a month, while Doctrine publishes no price and sells individual subscriptions after a seven day trial. Ordalie hosts data in France with S3NS and keeps inference in France under contract. Doctrine hosts on AWS in Frankfurt and runs OpenAI and Gemini models in Europe.

Which one fits

Choose Doctrine if

  • Research depth comes first. Doctrine's Legal Graph links 80 million French legal documents, from first instance decisions to the Court of Cassation.
  • Your firm runs heavy litigation. Flow Litigate turns a contentious file into a classification, a chronology, a statement of facts and a bordereau.
  • You advise on employment exits. Jobexit simulates the end of an employment contract in under two minutes, with indemnities, risks and timing.
  • You want a LexisNexis company behind the tool. RELX completed its acquisition of Doctrine in September 2026.
  • Your practice spans several European countries. Doctrine serves legal professionals in France, Italy, Germany, Spain and Luxembourg.

Choose Ordalie if

  • You want to try before buying. A free plan needs no card, and PRO costs €75 per user a month and MAX €99, excluding tax.
  • Your documents sit in iManage or a French practice system. Ordalie connects to iManage, SharePoint, Secib Neo and 20 other systems and reads documents in place.
  • Data must stay in France. S3NS, a Thales company, hosts it in France, and inference runs on Azure France under contract.
  • You are a notary. Ordalie offers workflows for French notaries, and the Congrès des Notaires de France is among its named users.
  • Procurement asks for attestations. Ordalie claims ISO 27001 and SOC 2 Type II certification, and its trust center releases the documents on request.

At a glance

Category
DoctrineLegal Research
OrdalieGeneral Legal Assistants
Founded
DoctrineNot published
OrdalieNot published
Headquarters
DoctrineParis, France
OrdalieParis, France
Last verified
DoctrineNov 16, 2026
OrdalieNov 16, 2026

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.

Doctrine
BB on AI CentralityThe models are the engine of a core capability, layered on a product that would still function without them as a document or workflow system.

Underneath the models is what the vendor describes as 80 million interconnected French legal documents, searchable by keyword, court, company, advocate, date and decision number, which was the product for years before generative features arrived. The model work on top includes a legal chatbot answering in natural language from the legal corpus, a documents chatbot answering only from files the user supplies, and a company chatbot over corporate and litigation records. It also includes decision and instrument summarization, extraction, comparison, conformity checking and translation. Flow Litigate assembles classification, chronology, statement of facts and bordereau for a contentious file, and Flow Counsel analyzes contracts for risk. Strip the models out and a large searchable legal database remains.

Ordalie
AA on AI CentralityThe artificial intelligence is the product. Remove the models and there is nothing left to sell.

Without the models there is nothing left to sell. Terms of Sale clause 2.1 defines the subscription as three services, all run by AI: an assistant that answers legal questions and analyzes documents, a search engine that uses AI to understand links between legal documents, and generation of contracts, memos and analyses. The only other item is training and support. The legal texts underneath are free on Legifrance, and Ordalie positions itself against manual Legifrance research, so what a customer pays for is the synthesis and the citation. Pricing follows the same logic: tiers differ by which AI models are available and by query volume.

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.

Doctrine
BB on Citation Accuracy and Hallucination DisclosureGrounding is real and documented, with linked primary sources and a described retrieval method, short of published accuracy figures an outsider can test.

The vendor's own blog states that the Chatbot Juridique answers exclusively on the basis of the legal corpus, while the Chatbot Documents relies only on the documents supplied, and that both always cite their sources. It sets out the reasoning too: a professional legal tool must let the user verify the origin of every assertion, and the way to get the best from these systems is to confine them to precise legal sources they can cite. The vendor also publishes, on its marketing blog, an analysis by Damien Charlotin, the researcher behind the AI Hallucination Cases database, of the first French decision involving AI hallucinations. And it argues that general purpose AI is unsuitable for legal work because a fabricated authority or a stale rule has concrete consequences. No accuracy figure, test set, evaluation or error rate is published.

Ordalie
BB on Citation Accuracy and Hallucination DisclosureGrounding is real and documented, with linked primary sources and a described retrieval method, short of published accuracy figures an outsider can test.

Grounding is real and the reader can open it. Every answer shows its sources, with direct links to the original text, references to the article, judgment number or page, and the passages relied on; where possible the link goes to Legifrance. The documentation describes the retrieval: the assistant selects sources, shows which it chose, then sorts and synthesizes them into an answer with matching citations formatted as it writes. The user controls the scope by switching source families and jurisdictions on and off. What is missing: no accuracy figure, error rate, test set or evaluation is published, citation status is not checked, and nothing says what happens when no support is found. The marketing claims no hallucinations.

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.

Doctrine
BB on Autonomy and Oversight ModelA written commitment that the models work alongside a supervising lawyer, with real review surfaces, short of the full control structure: commonly the threshold at which the system stops or what happens after it is wrong.

Assistant, Doctrine's legal agent, reads a user's documents, crosses them with case law and statutes, and returns a sourced, editable work product in Doctrine or Word. The agent breaks the work into steps, explains each one and does not decide in the user's place, and each proposed change is accepted or refused by the user. Flow Litigate produces classification, a chronology, a statement of facts and a bordereau from a contentious file, and Flow Counsel analyzes contracts and returns legal recommendations. Doctrine's stated position is that the user must be able to verify the origin of every assertion. No point is published at which the system defers to a person, and nothing covers what happens when a chronology, summary or recommendation is wrong.

Ordalie
BB on Autonomy and Oversight ModelA written commitment that the models work alongside a supervising lawyer, with real review surfaces, short of the full control structure: commonly the threshold at which the system stops or what happens after it is wrong.

Review tools come with a written position that a lawyer stays in charge. Every answer carries clickable source markers, the editor tracks every change so it can be commented on or reversed, and the Ledger documentation says its filters help review and do not replace checking the source. Terms of Use 5.2 says Ordalie cannot provide legal services, that its analyses are not legal advice, and that users should consult a qualified professional. Human oversight is one of the CNB dimensions it claims. What is missing: the same site markets workflows that return a finished deliverable ready to send to a client, and press material claims automation of up to eighty percent of repetitive tasks. No threshold, confidence boundary or type of work is published at which the system runs without a reviewer, and nothing says what happens after it is wrong.

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.

Doctrine
BB on Operational and Outcome EvidenceReal deployment evidence with substance, short of full attribution or measurement: a named customer without figures, or figures without the named customer.

The RELX completion release of 3 September 2026 states that Doctrine is used every day by 27,000 legal professionals across France, Italy, Germany, Spain and Luxembourg, and describes its customers as law firms of all sizes, large multinational corporations and public sector organizations including French government ministries, local authorities and universities. An institutional partnership with the Paris Bar gives members preferential subscription terms. A named advocate in company law is quoted on the record, with a linked interview, saying he uses the platform systematically in both advisory and contentious work to get a precise and exhaustive picture of a company's position. No figure for time saved, matters handled or outcomes changed is published, and no method sits behind any claim.

Ordalie
BB on Operational and Outcome EvidenceReal deployment evidence with substance, short of full attribution or measurement: a named customer without figures, or figures without the named customer.

Customer names and usage figures both appear on Ordalie's site, but never together. Ordalie claims 63,740 users, more than 800 client organizations, 32 hours of work saved a month on average and 98 percent retention. Named customers include the French Ministry of Justice and Ministry of the Interior, the Congrès des Notaires de France, SNCF Connect and Tech, the Paris Bar, Crédit Mutuel Arkéa, ESSEC Business School and Mondial Relay. The legal teams of France's prison administration have used Ordalie since May 2026, according to a September 2026 announcement. No figure is tied to a named customer, and the averages carry no stated method.

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.

Doctrine
BB on Privilege and Confidentiality PostureSubstantive published commitments on confidentiality and training use, short of the full picture: commonly silence on segregation between users or matters, or on what the underlying model provider may retain.

The security FAQ for documents states that documents and data are never used to train Doctrine's algorithms or AI models, that no Doctrine employee can access files except developers and product staff correcting an error or studying product use, that files go to no third party beyond the subprocessors the service needs, that documents pass through Microsoft Azure's European servers and are deleted immediately, and that Google and Microsoft are instructed not to retain them. The vendor blog adds strict isolation between customers and hosting only in the European Union, and its Document Analyzer post (2020, updated May 2026) says a confidentiality agreement in the CGV protects the customer's secret professionnel. Isolation between customers is asserted without detail, segregation between matters within a firm is not addressed, and the access exception extends to studying product use.

Ordalie
BB on Privilege and Confidentiality PostureSubstantive published commitments on confidentiality and training use, short of the full picture: commonly silence on segregation between users or matters, or on what the underlying model provider may retain.

Training, separation, privilege and model providers are covered in the contract, while retention during a subscription is left open. Terms of Sale 9.1, in force from 22 October 2026, says client data is not used to train models, and Ordalie adds that documents train no model, neither its own nor those of its AI providers. Case based spaces, projects per matter, role based permissions and private personal connections keep work apart, and a user's own sources are never visible to others. Clauses 7.1 and 9.3 bind Ordalie and its staff not to use or reveal transmitted documents and accept liability for breach. One narrow exception lets developers fix a processing error, the customer can refuse it, and the copies are destroyed within 24 hours. Inference providers must process in France, not train on client data, and not keep prompts, outputs or metadata. Data is purged 30 days after the contract ends, but no period is stated for uploaded documents during the contract.

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.

Doctrine
CC on UPL and Professional Responsibility PostureA boilerplate disclaimer sits in the terms while the marketing describes the product in advice terms, or the intended audience is left ambiguous.

Flow Counsel is described as analyzing contracts, identifying risks and returning legal recommendations; the legal chatbot answers substantive questions of French law; and the marketing addresses advocates and in house counsel directly. The vendor's published position on verification places responsibility with the user: it argues that a professional tool must let the user check the origin of every assertion. The product pages carry no statement that an output is not legal advice, that the subscribing lawyer remains responsible for the determination, or what the tool does not cover. The audience is unambiguous and professional.

Ordalie
BB on UPL and Professional Responsibility PostureA real position is published on advice versus tooling, short of full treatment: commonly a disclaimer without the supervision and competence dimension, or silence on jurisdiction limits.

Ordalie draws the advice line in its terms. Terms of Use 5.2 says it is not authorized to provide legal services, that its texts, positions and analyses are not legal recommendations or advice, and directs users to a qualified professional. It also claims alignment with the CNB's guidance on AI use by lawyers, naming human oversight and traceability, which touches the supervision duty. What is missing: the free plan is open to anyone with an email address, and nothing says the product is for professionals only or addresses use by people who are not lawyers; clause 3.1 only bars registering under a false professional status. No jurisdiction limit on reliance is stated for a product answering across several jurisdictions, and nothing explains how a supervising lawyer meets a competence duty over its output.

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.

Doctrine
CC on AI Governance and Bias DisclosureResponsible AI principles are published without a mechanism, a testing regime, or anything a buyer could audit.

The vendor writes publicly about the failure modes of legal AI. It argues that hallucinations have concrete consequences for a strategy or a file, that general purpose models are unsuitable for professional legal work, that reliability depends on the quality of the sources a system draws on, and that context management is key to useful output. It publishes an outside researcher's analysis of the first French hallucination decision on its own blog. These are positions on how legal AI should behave rather than on governance of Doctrine's own models. The Trust Center offers a Code of conduct on GenAI and, on request, a statement of compliance with the Paris Bar's White Paper on Artificial Intelligence. Nothing names who is accountable for the chatbots' behavior, describes what is evaluated before a change ships, or publishes any result. Nothing discloses whether output quality is uneven across areas of law, jurisdictions or the five languages the platform serves, and no model card or bias statement is published.

Ordalie
CC on AI Governance and Bias DisclosureResponsible AI principles are published without a mechanism, a testing regime, or anything a buyer could audit.

The product is designed, Ordalie states, to align with the French national bar council's (CNB) requirements for AI use by lawyers, naming transparency, confidentiality, human oversight, data protection and traceability. That is a claim of alignment with an outside framework, with no mapping, assessment or supporting document published. SOC 2 Type II and ISO 27001 are security certifications, not AI governance. No one is named as accountable for AI, nothing is published about testing before a model or workflow ships, and no incident has been disclosed. Bias is not addressed. On this product the risk is concrete: answers draw on source families the user switches on and off across jurisdictions of very different depth, so quality is likely to vary by practice area and country, and nothing examines that.

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.

Doctrine
BB on AI Safety and Data StewardshipSubstantive published policy covering most of the ground, short of the full set: commonly no named subprocessor list or no stated incident practice.

The security FAQ for documents says documents are held only for the time the provider needs to return its analysis, are deleted immediately from the Azure servers they pass through, and that Google and Microsoft are instructed not to keep them; the vendor blog lists deletion of customer data as available. No Doctrine employee can access files, save developers and product staff correcting an error or studying product use. Encryption of files and transfers and strict isolation between customers are stated on the blog, with hosting on AWS in Frankfurt. The Trust Center at trust.doctrine.com, on Vanta, lists the subprocessors: Amazon Web Services in Germany for hosting, Auth0 in Europe for authentication, Google Cloud in Europe for data processing and Azure in Europe as the OpenAI model hosting provider, which serves the OpenAI GPT models alongside Gemini through Google Cloud. No incident notification practice is published; the blog says only that no security incident has occurred.

Ordalie
BB on AI Safety and Data StewardshipSubstantive published policy covering most of the ground, short of the full set: commonly no named subprocessor list or no stated incident practice.

Data is encrypted with AES-256 at rest and TLS 1.3 in transit, with keys kept apart and rotated every 90 days. Access is logged, limited to authorized people and governed by role and group, with SSO on MAX. The privacy policy, updated 22 October 2026, names S3NS for hosting, Scaleway for internal models and backup copies, AI inference providers including Azure France, and Stripe for payment. Ordalie commits on its home page to report any incident within 24 hours, and Terms of Sale 9.4 to notify a personal data breach without undue delay. Content is erased from production when deleted or when its period ends, with copies left in encrypted backups until overwritten, and data is purged 30 days after the contract ends. No period is stated for uploaded documents during the contract, and the documentation keeps conversations for thirty days while the pricing table sells unlimited history. The inference providers are barred by contract from keeping prompts, outputs or metadata.

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.

Doctrine
CC on AI Liability and RecourseLiability is addressed only through a standard limitation clause that disclaims the exposure the product creates.

The general terms of sale (conditions générales de vente, in force from 23 January 2026) are published in French. Article 7.1 commits Doctrine to strict confidentiality over documents and data. That includes a ban on training or fine tuning any AI model on them and notice within two working days of any foreign authority's request for access, and breach may engage Doctrine's liability. Article 7.3 commits Doctrine to carry professional liability insurance, without stating the amount. Article 10.3 warrants that Doctrine holds the rights needed to provide the services and guarantees their peaceful enjoyment, the French law counterpart of an intellectual property warranty. The same article says Doctrine cannot be held civilly liable for any damage arising from use of the document services, and article 9 excludes liability when generative AI features are unavailable because of its AI providers. The site terms limit any liability to proven direct damage and say AI output never replaces a professional's analysis and must always be checked. No cap figure, indemnity or remedy for wrong output is stated.

Ordalie
BB on AI Liability and RecourseA real published position on liability, short of the full picture: commonly a stated indemnity without scope or caps.

Ordalie publishes a liability position that goes beyond a disclaimer, without reaching AI accuracy. Terms of Sale 7.2 commits it to carry professional liability insurance. Clauses 7.1 and 9.3 accept that Ordalie can be held liable for breaking its confidentiality undertaking. Terms of Use 5.4 limits liability to direct and proven damages rather than excluding whole categories. What is missing: no indemnity, cap figure, insurer or cover limit, and no warranty of output. Terms of Use 5.1 supplies the service on a best efforts basis at the user's own risk, and Terms of Sale 8.2 makes the user solely responsible. The 99.95 percent availability target is stated not to be a binding obligation, though the top tier is sold with an SLA. A customer relying on a wrong answer has no published recourse.

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.

Doctrine
BB on Practice Systems Integration DepthReal integrations exist and are documented, short of depth: named connections without a description of what they actually move.

Doctrine runs inside Microsoft Word through a plugin, where Assistant works on the open document, clauses and case law can be pulled in, and Flow Litigate manages exhibits and the bordereau. SharePoint and Google Drive are offered as connected document stores, with Dropbox marked as coming soon, and Doctrine says it also works with the user's mailbox. The GED connector brings a firm's own precedents and files into research alongside the public corpus, which the vendor frames as activating knowledge already stored in the document management system. No direction, sync behavior, configuration requirement or implementer documentation is published.

Ordalie
AA on Practice Systems Integration DepthDocumented, verifiable integrations into the systems legal work already lives in, with the depth described: what syncs, in which direction, and what a firm must configure.

Integrations come with their direction and setup spelled out. Ordalie lists 23 connectors, with custom ones on request, among them SharePoint, OneDrive, Outlook, Gmail, Google Drive, Dropbox, Box, iManage, Nextcloud, Synology, Nuxeo, Odoo, Tomorro, Agestia, Secib Neo and Dastra. The documentation covers setup connector by connector. iManage connects read only through OAuth with credentials from iManage Control Center, while Dropbox and Box can be given write access so output is saved back. Setup fields are published, such as tenant, site and client credentials for SharePoint. Since September 2026 the Ordalie chat also works inside Word and Outlook. There is no public developer API, and two way sync sits in the separate Vault module.

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.

Doctrine
BB on Deployment Model and Data ResidencyDeployment model is stated clearly with partial residency detail, or residency is offered without the processing location being addressed, or the tenancy model is stated on its own with no residency detail published.

The vendor blog says data is hosted only in the European Union, on AWS servers in Frankfurt, and the security FAQ for documents says the AI models that analyze documents run on Microsoft Azure and Google Cloud servers in Europe and that data never leaves the European Union, so processing is stated separately from storage. Subprocessors named on the Trust Center are AWS in Germany and Google Cloud, Azure and Auth0 in Europe. The platform is delivered as a single web service; no alternative deployment, regional choice or tier difference is described, and tenancy is asserted as strict isolation between customers without detail.

Ordalie
AA on Deployment Model and Data ResidencyDeployment options and data residency are published, including the regions available, what changes between tiers, and where processing happens as distinct from where data is stored.

The region, what changes by tier and where processing happens are each stated. Infrastructure and data are hosted in France by S3NS, a Thales company, on a SecNumCloud offering, named in the Terms of Sale in force from 22 October 2026 and in the privacy policy. Scaleway holds Ordalie's internal models and an encrypted copy of backups. The pricing table offers on premises deployment as an option on MAX, with SSO and access governance, and the enterprise route adds on premises or private hosting. Third party inference providers, including Azure France, must by contract process data in France, so the model call stays in France too. France is the only hosting region offered.

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.

Doctrine
BB on Security Certifications and Trust CenterCertification is real and stated, short of accessible evidence: a named standard without scope, date, or a way to obtain the report.

The vendor blog (published 14 January 2025, updated 7 May 2026) states that Doctrine has been ISO 27001 certified since January 2025 after several months of verification by an independent auditor, that the certification is managed through Vanta, and that its hosting provider AWS holds ISO 27001, 27017, 27018 and 9001 certification. The Trust Center at trust.doctrine.com, on Vanta, lists ISO 27001:2022 and GDPR compliance, offers the ISO 27001 certificate and the Statement of Applicability as direct downloads, and provides its security policies, a penetration test report, an architecture document and a security questionnaire on request. Vanta monitors the listed controls continuously. The site itself does not state the certificate scope or the auditor's name.

Ordalie
BB on Security Certifications and Trust CenterCertification is real and stated, short of accessible evidence: a named standard without scope, date, or a way to obtain the report.

ISO 27001 certification, in its 2022 version, and a SOC 2 Type II report are claimed in Ordalie's own name. An independent SOC 2 audit covered 79 controls with no exceptions, by Ordalie's account, and the report is available on request. A trust center at trust.ordalie.com, run on Bastion, carries both and shows 146 controls under monitoring, with the documents released on request. The claims concern Ordalie itself, not its host. No auditor, report period, certificate number or SOC 2 scope is published outside the request process.

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.

Doctrine
BB on Model Supply Chain DisclosureThe supply chain is partly disclosed: providers named without change notification, or architecture described without the providers.

The security FAQ for documents says uploaded documents are analyzed with AI models hosted in Europe: OpenAI's GPT models supplied through Microsoft Azure on European servers, and Gemini through Google Cloud on European servers, with data never leaving the European Union. The vendor blog on adopting generative AI says users' questions are sent to a version of OpenAI's GPT models hosted only in Europe and that all of its providers are based in Europe. On the Trust Center, Azure in Europe is named as the OpenAI model hosting provider and Google Cloud in Europe for data processing. Assistant answers from the legal corpus and the user's own files, with sources. No model version is named and no notice of changes is committed.

Ordalie
BB on Model Supply Chain DisclosureThe supply chain is partly disclosed: providers named without change notification, or architecture described without the providers.

The inference provider and Ordalie's own retrieval models are named, but not the model that writes answers. The privacy policy (article 3.4, updated 22 October 2026) and Terms of Sale 9.1 name AI inference providers including Azure France, which must process in France under contract. Ordalie says it trains its own legal models, and Solon Embeddings 0.1, its open embedding models for French text retrieval, are published on Hugging Face. Scaleway hosts those internal models. No generation model or version is named, and the basic and advanced model tiers are commercial labels. No notice of model or provider changes is published, and the subprocessor list is accurate as of its update date only.

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.

Doctrine
CC on Commercial TransparencyPricing is gated behind a demo request while tier names and feature splits are published, so the shape is visible and the number is not.

The pricing page presents three buyer profiles, the individual advocate, the law firm and the in house legal department, rather than a single enterprise funnel. Each subscription is individual and gives unlimited access to all Doctrine Pro features across every area of law, without segmentation by practice area. Flow Litigate, Flow Counsel, Jobexit and Le Fiscal by Doctrine need an additional subscription, according to the pricing FAQ, and prices are stated excluding tax. A seven day free trial needs no bank card and no commitment. The Paris Bar partnership publishes a commercial term rather than a rate. Members get several free months and a discount over a twelve month commitment, with a second license free in the first year for firms of two advocates. No price, price range or unit of charge is published on the vendor's own pages.

Ordalie
AA on Commercial TransparencyA buyer can learn what this costs without entering a sales process: published rates, the unit being charged, and what implementation adds.

The price is published, and buying needs no conversation. The free plan costs nothing. PRO is €75 and MAX €99 per user per month excluding VAT, with ten percent off for annual billing, which makes PRO €67.50 a month. A twenty six row table compares the tiers across models, volume, history, sources, uploads, exports, collaboration, integrations, governance and support. Purchase is by card or SEPA through Stripe, seats and plan changes prorate immediately, monthly plans have no commitment and cancel in a few clicks, and invoices download directly. Enterprise beyond MAX is quoted, covering Vault, on premises or private hosting, custom connectors and pilots. Two inconsistencies: the table includes Vault and on premises deployment in MAX while the FAQ puts both in the enterprise offer, and the Terms of Sale say prices include twenty percent VAT while Ordalie's pricing page says they exclude it.

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.

Doctrine
BB on Firm and Practice CoverageSegment and practice coverage is described with substance, short of the boundaries: what is supported is clear, what is not is left open.

Buyer segments are the structure of the offer itself, with distinct propositions for the individual advocate, the law firm and the in house legal department. The RELX release adds public sector bodies, including French government ministries, local authorities and universities. Jurisdictional reach is stated as France, Italy, Germany, Spain and Luxembourg, which the vendor frames as a global view with local impact. A subscription reaches every area of French law without segmentation, with a tax encyclopedia called out separately, and the legal database is quantified at 80 million interconnected documents. No area of law, jurisdiction or matter type is named as unsupported, no depth comparison is offered between the five countries, and nothing describes what the platform does not cover.

Ordalie
BB on Firm and Practice CoverageSegment and practice coverage is described with substance, short of the boundaries: what is supported is clear, what is not is left open.

Who the product serves and what it covers are both described in detail. Three audiences each have a page: lawyers in private practice, in house counsel and French notaries, with named notarial customers and workflows for notaries. Firms get shared case files, role based permissions and a private space per lawyer. The sources documentation lists the 102 French codes, the courts covered and BOFiP for tax, and shows how a practitioner narrows the scope. Jurisdictions are selectable, and uncovered countries can be requested. What is missing: no firm size, unsupported practice area or government use is addressed. The coverage claim is also inconsistent. The lawyer page and FAQ claim 3.5 million official texts across eight jurisdictions, including Germany, Italy, Spain and Belgium, while the sources documentation lists France, the EU, Luxembourg, Great Britain and Switzerland and two million references.

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?

Doctrine
Never, in the contract

The customer agreement prohibits training on customer documents. Article 7.1 of Doctrine's general terms of sale (conditions générales de vente, in force from 23 January 2026) commits it to strict confidentiality over customer documents and data, including a ban on training or fine tuning any AI model on them. The security FAQ for documents states the same position, that customers' documents and data are never used to train Doctrine's algorithms or AI models, and the vendor blog repeats that no data is used to train its models.

Ordalie
Never, in the contract

The training ban is in the agreement. Terms of Sale 9.1 says client data is not used to train models, and requires the AI inference providers to process in France and not keep prompts, outputs or metadata. Privacy policy article 3.4 bars those providers from reusing client data for training, and article 3.6 says Google Workspace data is not used to develop or train any generalized AI model. The marketing says the same.

Three qualifications: the terms and privacy policy list improving the services as a purpose, without naming training, while clause 9.1 governs; the Google clause covers only generalized models; and no data processing addendum carries the commitment into a signable annex.

Prompt and Output Retention

How long does the product keep what a lawyer typed, and can that be set to zero?

Doctrine
Customer controlled, no zero option

The vendor blog lists deletion of customer data as available. The security FAQ for documents says documents are held only as long as the provider needs to return its analysis, are deleted immediately from the Azure servers they pass through, and that Google and Microsoft are instructed not to keep them; analysis results remain available in the customer's interface. No platform retention period for documents, prompts and results, and no zero retention setting, is published.

Ordalie
Disclosed fixed window

Two pages give different retention periods. The legal research documentation says conversation history is kept for thirty days, while the pricing table sells history as a tier feature, seven days on the free plan and unlimited on paid plans. The Terms of Sale in force from 22 October 2026 defer to the period that applies to the workspace or organization without saying what it is or who sets it. Data is purged 30 days after the contract ends, by Ordalie's account.

Deletion is precise, immediate and irreversible from production, with copies in encrypted backups until overwritten. The inference providers are barred from keeping prompts, outputs or metadata.

Ethical Walls and Matter Segregation

Does retrieval respect the firm’s ethical walls, or can the model read across them?

Doctrine
Claimed, not documented

The vendor blog states strict isolation between customers, and the security FAQ for documents says no Doctrine employee can access files, save developers and product staff correcting an error or studying product use. Nothing published describes permissions within a firm, walls between matters, or how isolation is implemented.

Ordalie
Own model, documented

Ordalie keeps its own documented permission model, which the firm must keep aligned. It has case based spaces with role based permissions and activity history, projects per matter, a private space for each lawyer, private tags, a user's own sources kept private, account isolation, and a split between personal and organization integrations. Access to content is limited and logged. The iManage connector can run as one shared connection or with each member signing in, and only the second applies iManage's own access rules per user.

Because neither is required, keeping walls correct falls to the firm's administrators. Two gaps: no conflicts or ethical wall feature is named, and nothing says whether permissions limit what the assistant retrieves.

Third Party Request and Subpoena Notice

If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?

Doctrine
Notice committed

Article 7.1 of Doctrine's general terms of sale (conditions générales de vente, in force from 23 January 2026) commits it to inform the customer within two working days of any request by a foreign authority for access to customer documents or data. The commitment does not cover requests from French courts or authorities or domestic legal process generally, and no transparency report is published.

Ordalie
Disclosure addressed, notice absent

Data can leave under the terms, with nothing said about telling the customer. Terms of Sale 7.1 gives an irrevocable confidentiality undertaking over data generated by the service, with an exception for a legal obligation or a court injunction. Clause 9.3 repeats the same exception for documents sent to the platform. No commitment to notify the customer of a demand, reserve discretion either way, challenge a demand or cooperate on a protective order is published, and no transparency report exists.

The company and data are French, so a demand would usually run through French courts, but nothing says so and the clause is not limited to French authorities.

Primary Law Corpus Provenance

Where does the law in this product come from, and does the vendor have the right to use it?

Doctrine
Sources named, basis unstated

The sources are named, and a court has ruled unlawful the way part of the database was built. The sources are French case law, legislation and regulatory content, which the vendor quantifies at 80 million interconnected documents. No rights basis for the content is stated. On 7 May 2025 the Paris Court of Appeal found serious, precise and concordant presumptions that Forseti, which publishes doctrine.fr, had obtained hundreds of thousands of first instance decisions without authorization from court registry directors.

On the same presumptions it found that Forseti diverted a Conseil d'État research convention and exploited commercial court data after its Infogreffe partnership ended. It held those acts to be unfair competition against five legal publishers, awarding €40,000 each to Lexbase, Lextenso and Lamy Liaisons and €50,000 each to Éditions Dalloz and LexisNexis. It also ordered an extract of the judgment published on the Doctrine home page for sixty days.

Doctrine did not appeal to the Cour de cassation. The judgment concerns the collection built between roughly 2016 and 2019, before judicial open data, and says nothing about current sourcing.

Ordalie
Sources named and licensed

Sources, rights and update cadence are published. The sources documentation covers the 102 complete French codes, non codified laws and ordinances from the Official Journal, case law court by court, BOFiP for tax, EU law and judgments from EUR-Lex, and Luxembourg sources when switched on, and a November 2026 changelog adds Dutch legislation. Terms of Use 4.1 states that court decisions and legal texts carry no intellectual property rights, and 4.2 says third party commentary is linked rather than reproduced.

The corpus is kept constantly up to date. No publisher license is named, and Ordalie's marketing names eight jurisdictions, including Germany, Spain, Ireland, the United Kingdom and Switzerland, which the sources documentation does not describe.

Good Law Verification

Does the product tell you when the authority it just cited has been overruled?

Doctrine
Not addressed

Nothing published describes a treatment signal, a subsequent history flag or an alert when a cited decision has been reversed or a provision repealed. The product sells search from first instance to cassation, and the vendor's material describes retrieval, citation of sources and conformity checking of documents. The blog argues that a stale rule can have concrete consequences for a file.

Ordalie
Prompts the user to verify

The reader is sent to the authority and told to check it, with no treatment analysis done in the product. Answers link to the source, often on Legifrance, with exact references, and the Ledger documentation says its filters do not replace checking the cited source. Nothing indicates whether a cited decision was overturned, confirmed or departed from, and no treatment flag is described. The citator is a common law tool with no exact French equivalent.

The nearer French question, whether a code article is the version in force, is partly covered by the statement that the codes are kept continuously current, which describes the corpus rather than a check shown on an answer.

Refusal and Uncertainty Behavior

What does the product do when the answer is not in the corpus?

Doctrine
Not addressed

The legal chatbot responds exclusively on the basis of the legal corpus and the documents chatbot only from the files supplied, both always citing their sources. The vendor explains the design reason: confining a model to precise sources it can cite is the best available way to suppress hallucination and make verification easy. Nothing published states whether either chatbot declines a question the sources cannot support, reports that it found nothing rather than assembling an answer from adjacent material, or exposes any confidence or coverage signal.

Ordalie
Not addressed

Nothing published describes what the assistant does when it cannot answer. The legal research guide, the sources documentation, the Ledger documentation and the lawyer FAQ, including its answer on how Ordalie reduces hallucinations, all answer with grounding: every answer cites verifiable official sources with links. None says a question outside the corpus is refused, a thin basis is flagged, a confidence indicator is shown, or an ambiguous question is put back to the user.

The only related behavior is in export, which stops and reports a missing attachment. The marketing says the product does not hallucinate, which makes the missing no answer path matter more.

Fabricated Citation Record

Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?

Doctrine
None located

The AI Hallucination Cases database maintained by Damien Charlotin holds no entry naming Doctrine or Forseti as the source of fabricated authority. Forseti's name appears in French court reporting from May 2025 on unfair competition in how its corpus was assembled and on a former employee's unauthorized access to a court registry system, and neither concerns invented authority in a filing.

Ordalie
None located

Trackers and coverage of decisions on AI fabricated citations, in French and English, record no decision naming Ordalie or its company. That record covers fabricated citations only. French courts have started meeting the problem: the administrative court of Grenoble in December 2025 described a claim as drafted with generative AI and citing fanciful case law, and the administrative court of Orleans found cited decisions that do not exist; both involved self represented litigants and named no product.

A March 2026 practitioner survey found no French court had yet sanctioned a lawyer for AI fabricated authority, so an empty French record carries less weight than an empty US one.

Bar Guidance Alignment

Has the vendor engaged in public with the ethics opinions its buyers are bound by?

Doctrine
Generic reference

The vendor publishes that a professional legal tool must let the user verify the origin of every assertion and that a general purpose model is not sufficient for legal work. It adds that an invented authority, a badly cited source, a stale rule or an approximation about a legal regime can have concrete consequences for a strategy or a file. It hosts on its own blog an analysis by Damien Charlotin of the first French court decision involving AI hallucinations, addressed to practitioners.

No deontological rule or CNB position is cited on the public pages, and the Paris Bar's White Paper on Artificial Intelligence is named only in the title of a Trust Center document available on request. Nothing addresses the position of an advocate who relies on a generated chronology, bordereau or contract recommendation in a matter.

Ordalie
Named guidance addressed

Ordalie states on its lawyer page that it complies with the seven requirements set by the Conseil national des barreaux, the French national bar council, for lawyers' use of AI. An FAQ answer says the product is designed to align with that guidance and names five dimensions: transparency, confidentiality, human oversight, data protection and traceability. Two limits: only French guidance is engaged, though the product covers other jurisdictions and serves notaries, whose rules are separate; and no mapping of features to the seven requirements is published. Ordalie also has a partnership with the Paris Bar to promote AI in law firms and offers its members a discount.

Billing and Fee Posture

Does the vendor address what happens to the bill when the work takes an hour instead of six?

Doctrine
Savings claims only

Doctrine sells to individual advocates and law firms whose work is billed to clients, and its stated value is speed: taking hold of a contentious file in seconds with classification, chronology, statement of facts and bordereau produced automatically, understanding a corporate instrument in seconds rather than reading it, analyzing contracts and obtaining recommendations in a few clicks, and gaining time, files and clients.

Nothing published addresses what happens to the client's bill when that work compresses, whether a matter carries any record of AI assisted work, or how a firm would disclose the use of the tool in a fee note.

Ordalie
Savings claims only

Ordalie claims specific time savings: eight hours a week on average, a case timeline in three minutes against three hours by hand, an hour instead of a full day, a twelve page contract draft in fifteen minutes, and an onboarding path aimed at reclaiming hours each week. Press material repeats a claim of automating up to eighty percent of repetitive tasks. Nothing addresses the client's bill. Its main buyers bill clients for the research and drafting it speeds up, yet no per matter record of AI assisted work, invoice disclosure or fee guidance is described, including in its CNB alignment claim. Nothing helps a firm decide what to tell clients who were going to be billed for that time.

Outside Counsel Guideline Readiness

Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?

Doctrine
Subprocessors listed

The Trust Center at trust.doctrine.com lists subprocessors with locations: Amazon Web Services in Germany, Auth0 in Europe, Google Cloud in Europe and Azure in Europe as the OpenAI model hosting provider. The security FAQ for documents names OpenAI's GPT models supplied through Microsoft Azure and Gemini through Google Cloud, on European servers, as the AI that analyzes uploaded documents, and the vendor blog names OpenAI's GPT models hosted in Europe for users' questions.

Security policies, a questionnaire and a statement on the Paris Bar's White Paper on Artificial Intelligence are available on request, and the ISO 27001 certificate downloads directly.

Ordalie
Subprocessors listed

A dated subprocessor and model provider list is published, but no pack a firm can forward. The privacy policy (article 3.4, updated 22 October 2026) names S3NS for hosting, Scaleway for internal models and an encrypted backup copy, AI inference providers including Azure France with their contract terms, and Stripe for payments. It goes past infrastructure to say who runs inference and what they may not do with client content.

The SOC 2 Type II report and ISO 27001 certificate sit in a trust center, released on request. No separate data processing addendum is published, and the controller and processor terms sit in Terms of Sale 9.4. No notice of list changes is promised.

Court Disclosure Support

If a judge’s standing order requires an AI disclosure, can the product produce one?

Doctrine
Partial record

Source attribution is the part of a record that exists. Both chatbots always cite the material they relied on, and each is confined to a defined set of sources, so a user can see whether an answer came from the public legal corpus or from the user's own documents, and the vendor's stated design goal is that the origin of every assertion can be verified. No export per item records which model produced an output, what it retrieved and what a person verified before it was used.

Nothing addresses a French court's expectations about disclosing AI assistance in a filing, though the vendor's own blog analyzes the first French hallucination decision, and no certification or template a filer could attach is offered.

Ordalie
Not addressed

Nothing published addresses disclosing AI involvement in work that reaches a court. Ordalie does produce provenance: answers list their sources, citations link to the original text, the editor tracks every change, and a conversation can be exported in full to PDF or Word with citations. A lawyer could reconstruct which authorities an answer used. But nothing records that a machine was involved: no model is identified, exports do not mark generated text, and no disclosure template or certification exists.

The Ledger is a contract review dashboard, not an activity log. This matters in France, where courts have started remarking on AI drafted filings in published decisions.

What neither one publishes

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.

Signals neither addresses in public material
  • Refusal and Uncertainty Behavior
Change log

Product updates

The latest logged changes from each vendor, newest first, each linked to its source.

Doctrine

  • Product / capability

    Doctrine added full document drafting to its two workflow products after a beta earlier in 2026.

    Source
The Doctrine update →

Ordalie

  • Integration / interoperability

    Ordalie users can now add a file from a connected source to a project through the project picker, and the file is opened with that connection's permissions.

    Source
The Ordalie update →

In summary

Doctrine

Doctrine is a legal research and AI platform from Paris, published by Forseti SAS and part of LexisNexis since RELX completed its acquisition on 3 September 2026. It is used by 27,000 legal professionals in France, Italy, Germany, Spain and Luxembourg, and rests on a Legal Graph of 80 million interconnected French legal documents. Assistant, its legal agent, crosses a user's documents with case law and statutes and returns a sourced draft that can be edited in Doctrine or Word. Flow Litigate, Flow Counsel, Jobexit and Le Fiscal are sold as additional subscriptions. According to the AI Legal Index, Doctrine's general terms of sale ban training or fine tuning any AI model on customer documents. In May 2025 the Paris Court of Appeal found that Forseti had unlawfully collected court decisions, a ruling Doctrine did not appeal.

Source: AI Legal Index, 2026

Ordalie

Ordalie is a Paris legal AI workspace for lawyers in private practice, in house teams and notaries. It answers legal questions in French or English with every source linked down to the article or judgment number, drawing on French codes and case law, EU law and Luxembourg sources. Packaged workflows build case timelines, compare contracts and check consistency, and a Ledger view reviews sets of contracts. Its 23 connectors include iManage, SharePoint and Outlook, and the chat works inside Word and Outlook. According to the AI Legal Index, Ordalie's Terms of Sale bar training on client data, data is hosted in France by S3NS, a Thales company, and Azure France runs inference under contract. A free plan exists, and PRO costs €75 per user a month, excluding tax.

Source: AI Legal Index, 2026

Questions buyers ask

How does Doctrine compare to Ordalie?

Both answer French legal questions with linked sources and work inside Word. Doctrine is a research database first, with 80 million interconnected documents, an agent that crosses them with a client file, and paid modules for litigation, contracts, employment and tax. Ordalie is an AI workspace first, with 23 connectors, packaged workflows and published prices. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.

What do Doctrine and Ordalie cost?

Doctrine publishes no price. It sells individual subscriptions after a seven day free trial without a bank card, and its four modules need an additional subscription. Ordalie has a free plan, PRO at €75 and MAX at €99 per user a month, excluding tax, with 10 percent off for annual billing. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.

Do Doctrine and Ordalie train AI on client documents?

Neither does, under its terms. Doctrine's general terms of sale ban training or fine tuning any AI model on customer documents and data. Ordalie's Terms of Sale say client data is not used to train models, and its inference providers may not keep prompts or outputs. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.

Where do Doctrine and Ordalie host data?

Doctrine hosts data on AWS in Frankfurt and analyzes documents with OpenAI models on Azure and Gemini on Google Cloud, both in Europe. Ordalie hosts in France with S3NS, a Thales company, and sends inference to providers including Azure France, which must process data in France. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.

Which legal sources do Doctrine and Ordalie cover?

Doctrine's Legal Graph links French case law, legislation, collective agreements, tax treaties and company data, and it serves professionals in five European countries. Ordalie covers the French codes, statutes, case law and BOFiP, EU legislation and court judgments, and Luxembourg sources, with Dutch legislation added in November 2026. From the AI Legal Index, based on each vendor's own published materials as of November 16, 2026. No vendor pays for placement.

Disclosure

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.

Neither vendor publishes an accuracy figure or test set for its answers. Doctrine publishes no price. Ordalie's documentation keeps conversations for thirty days while its pricing table sells unlimited history, and its Terms of Sale say VAT is included while its published prices exclude tax.

Contact

Correct a record, or ask how something was graded

Every grade and every signal on this index is drawn from public sources and dated. If a record is wrong, out of date, or missing an artifact the index did not locate, send the source and it will be reviewed and the record redated. Vendors are welcome to submit documentation. Nothing on this index is for sale, including a listing, a placement, or a grade.

AI Legal Index

The AI Legal Index is an independent index that tracks changes to AI vendors in legal. It holds 535 vendors across 9 categories, each graded on the same 15 capability axes and recorded against 12 legal signals, from public sources. No vendor pays for inclusion, placement, or rating.

Index Status
Last index update
November 19, 2026
The AI Legal Index is an editorial reference. It is not a regulatory body, not a law firm, and nothing published here is legal advice or a recommendation to retain or avoid a vendor. Records are verified against published sources, bar guidance and public court records. Where a record reads not addressed, the material was not located in public sources on the date shown. See the Methodology page for evaluation standards and limitations.
© 2026 AI Legal Index
3801 N Capital of Texas Hwy, Ste E240 · Austin, TX 78746