E
Enidia
Enidia is document review AI from Enidia AI of Esch sur Alzette, Luxembourg, built to fill a fixed list of questions from documents and show where every answer came from. Law firms use it for due diligence lists, closing checklists and playbook reviews, fund managers for CSSF questionnaires, AIFMD Annex IV reports and SFDR annexes, and in house teams to test supplier contracts against DORA or NIS2.
The AssociAte sorts each answer into found, with the page and passage shown, inferred, with both passages and a label, or absent, reported as a gap. Every run ends in a certificate of each element, its state and its verbatim source. The Librarian builds a knowledge base from a firm's own documents, and the Clerk builds counterparty records from documents and registries such as RCS, GLEIF and Companies House.
In production Enidia runs on a workstation the firm owns, about €5,000 of hardware, so no document reaches an outside model, while a free trial runs in Enidia's cloud. Enidia reports a deployment verified at a Global 10 law firm and a checklist cut from 25 hours to 4 in its own timed run. Research grants of €600,000 from Luxembourg's FNR back the work, and the license is priced to each firm.
Alternatives to Enidia
Other products that do the same kind of work as Enidia, and what each one does.
- NNasdaq LensNasdaq Lens is a disclosure research, benchmarking and drafting platform for public company legal, finance, sustainability and risk teams, launched in 2025 from Nasdaq's Sustainable Lens.
- SSixFiftySixFifty is an employment law compliance platform from SixFifty Technologies, LLC of Salt Lake City, sold to HR teams, in house legal departments, HR consultants and PEOs, and startups.
- BBleeBlee is an AI marketing compliance platform from Blee, Inc. of New York, sold to legal, compliance and risk teams that review marketing, product and sales content, with marketing and…
- CCorlyticsCorlytics sells a regulatory intelligence platform to banks, insurers, asset managers, payments firms and regulators.
- LLuminos.AILuminos.AI is a Washington, DC company, founded in 2023 by Andrew Burt and Mike Schiller, that sells an AI governance and evaluation platform to legal, privacy and AI governance teams and…
- RRedacticRedactic is document redaction and masking software from KS Agents, a software company in Trieste, Italy, launched as its own product in June 2026.
Capability grades
All 15 axes, graded from public sources on the date shown. 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.
Every product is model work on documents. The AssociAte reads a document set and fills a checklist, form, survey or template with found, inferred or absent answers, the Librarian answers questions over a firm's own documents, and the Clerk extracts counterparty facts into structured records. Without the models nothing remains to sell.
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.
Every found answer carries its page with the supporting passage framed, inferred answers show both passages and are never marked found, and missing items come back as absent. The AI Verification Standard names hallucination and incompleteness as the two failure modes and commits Enidia to posting benchmark documents, full outputs and source attributions for outside reviewers to test. No benchmark results are available yet, and the Global 10 verification comes without figures or the name of the verifier.
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.
The AssociAte works through a fixed list a lawyer sets, and the rule that an inferred value is never marked found applies to every run. Each answer lands in one of three visible states, a user can challenge an answer or teach a new term in plain words, and every run ends in a certificate listing each element with its state, verbatim source and location so a reviewer can audit it without the software. Fields it cannot verify are reported as unverifiable rather than filled.
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.
Enidia reports a deployment independently verified at a law firm in the Global 10, naming neither the firm nor the verifier. In its own timed run, a limited partnership agreement checklist that took 25 hours by hand was done in 4, a result no firm has confirmed. Savings run from about 95% on standardized forms such as a DICI to about 70% on bespoke agreements, with most documents between 73% and 85%. No named customer stands behind the figures.
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.
In production the software runs on hardware inside the firm and nothing goes to a third party model, so privilege is not exposed to an outside service, and documents under embargo or contractual bans on external processing can be handled. The free trial runs in Enidia's cloud on public or anonymized files, with a promise that nothing sent is used for training, deletion when the trial ends and an NDA on request. Separation between matters and the handling of privilege inside the software are left out, and no customer agreement is posted.
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.
Enidia says plainly that it is not a general legal assistant and does one task, filling fixed lists from documents for law firms, fund professionals and in house legal and compliance teams. Its verification standard holds that a practitioner cannot meet a professional duty by pointing to output that cannot be traced to its source, and that the duty to look for omissions falls back on the practitioner when a tool hides them. No jurisdiction limits or named bar rules accompany that position.
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.
The AI Verification Standard, a July 2026 draft for public comment written by Tomer Libal under a Creative Commons license, sets eight requirements for verified output, from source attribution and three state classification to explicit omission reporting. It adds a benchmark method open to outside reviewers and a dispute panel of a practitioner, a technologist and an academic, and Enidia commits to post its own results and test other tools that submit output. Results across document types, languages and markets have yet to appear.
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.
Production data stays on the firm's own machine, so retention and access follow the firm's controls. Documents sent to the hosted trial are deleted when it ends, with no trial length stated. The privacy policy of September 2, 2026 covers inquiries, demo bookings and analytics, keeping inquiry records for three years, and leaves out trial documents, access control, the trial host and incident practice.
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.
No license agreement, terms of service or limitation of liability is posted, and nothing covers a wrong answer or a missed element with a warranty or remedy.
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.
The Librarian works over a firm's document collection with no IT integration required, and the Clerk draws on the RCS, GLEIF and Companies House registries. Forms come from regulators' own download pages, and results go back as filled templates and certificates. No document management, matter system or Word connection is named.
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.
Production runs on a workstation the firm buys, about €5,000 of hardware, so documents are stored and processed inside the firm and nothing reaches an outside model. The trial runs apart from production in Enidia's own cloud, location unstated, on files the firm can share. Scaling means adding another machine the firm owns.
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.
No ISO 27001 certificate, SOC 2 report or trust center is offered. The security case rests on running inside the firm with no external processing, and the Clerk's AMLD6 readiness is a claim rather than an attestation.
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.
Models run locally on the firm's hardware with no third party model involved, and the tool is promised to keep working when a model vendor changes course. Enidia's CTO makes the case for small local models a firm can run itself, with no model family or version given for the product.
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.
Firms buy the machine, about €5,000 once, and license the tool at a price fitted to their volume and matter mix, with no per seat or metered charges, so heavy months cost the same as light ones. The Librarian is free for law firms with 20 or more fee earners and for fund and in house teams of that size, and the trial is free. A calculator estimates hours and fees released by document type, volume and rate, while the license figure and term come on request.
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.
Three kinds of buyer are served: law firms, fund managers with AIFMs and administrators, and in house legal and compliance teams. Templates follow EU rules and national ones in Luxembourg, France, Germany and the Netherlands, from CSSF questionnaires and DICI checks to German investment management agreements, Dutch Wwft dossiers and MiFID II suitability reports, alongside limited partnership, loan and shareholder agreements, processor agreements and CSRD disclosures. General legal assistance is out of scope by design.
3 public documents
The public pages on file for Enidia, with the recorded signals each one supports and the date it was last read. Open any of them and check the reading against the record.
-
enidia.ai2 signals
Client Data in Training, Prompt and Output Retention
Read Nov 21, 2026
-
Refusal and Uncertainty Behavior
Read Nov 21, 2026
-
Fabricated Citation Record
Read Nov 21, 2026
No published figure
- Enidia is bought as a license plus a machine the firm owns.
- The machine costs about €5,000 once, and the license price depends on the firm's volume and work.
- There is no charge per user and no charge per page or question.
- The trial is free, and the Librarian is free for larger law firms and for fund and in house teams.
A license priced to the firm's volume and matter mix, plus hardware of about €5,000 that the firm buys and owns. No per seat billing and no metered usage, so heavier months and more users do not change the bill; scaling means adding another machine. The Librarian is free for law firms with 20 or more fee earners and for fund and in house teams.
Implementation: Not stated. The free trial runs on Enidia's hosted environment before anything is bought.
Note: The license terms, contract length and renewal are not posted.
Legal Signals
What each signal meansA signal records what public sources say on the date shown. It is not a grade and it is not a recommendation. Where a signal reads Not addressed, it means the index did not locate the material in public sources on that date, which is a statement about disclosure rather than about the product.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
A public policy or trust page states no training on customer content, with no matching term located in the published agreement.
Nothing sent to the hosted trial is used for training, a promise on the website rather than in a contract, and in production the models run on the firm's own hardware with no outside model.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
Retention is acknowledged in public materials with no stated period.
Trial documents are deleted when the trial ends, with no trial length given, and in production data stays on the firm's machine.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
No located public material addresses walls or matter level segregation.
How matters or users are kept apart on the firm's Enidia machine is not described.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
No located term or policy addresses third party requests for customer data.
In production Enidia holds no client documents, and no term covers legal demands for documents sent during a trial.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
Sources are identified without stating the license or rights basis.
Answers come from the firm's own documents, and the Clerk adds registry data from RCS, GLEIF and Companies House. Terms for using those registries are not given.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
No located public material addresses whether authority is checked for subsequent history.
Answers cite the firm's own documents, not case law, so no authority is tested for later history.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
The vendor documents an explicit no answer path and it is observable in the product or in published evaluation.
Items the documents do not answer are reported as absent and fields that cannot be verified as unverifiable, and Enidia's demonstration films show four questions coming back with no answer and saying so.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
No court order, opinion or disciplinary record addressing fabricated or hallucinated legal citations produced by this product has been located as of the date shown. This is a statement about the public record on that one subject, not a finding about the product, and this signal is not a litigation history.
No entry in the AI Hallucination Cases database maintained by Damien Charlotin names Enidia.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Public materials refer to professional responsibility in general terms without naming guidance.
The AI Verification Standard ties auditable, complete output to a practitioner's professional duty in general terms, without naming a bar opinion.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
A usable record of AI assisted work exists with no published fee guidance.
Every run produces a certificate of each element, its state and source, a record of the AI assisted work. Enidia suggests hourly firms read its time savings as released capacity, with no guidance on billing or disclosing the work to clients.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
No located public material supports a client side disclosure obligation.
Production adds no processor or outside model. The trial host is left out, and the processors in the privacy policy serve only the website, contact form, email and scheduling.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Some elements of the record are available, short of a document level export.
Each run's certificate lists every element with its state, verbatim source and location and can be handed to a reviewer. It does not name the model used or record a person's sign off.