DeepJudge is an AI search and workflow platform that helps a law firm find and use what it already knows. It indexes a firm's documents where they sit, across the document management system, email, SharePoint, OneDrive, HighQ, intranets and experience systems, without moving them. Search works by intent. It sorts documents into a taxonomy, spots near duplicates and redlines, and links information on clients, matters, documents and people.
A lawyer can ask whether opposing counsel ever conceded a point in an earlier negotiation, or build a timeline of a matter. Access permissions sync continuously from the source systems, so existing ethical walls carry over, and the firm controls what is indexed. A workflow builder lets the firm assemble AI agents that pull from several sources and act, using whichever commercial or open source models it chooses. DeepJudge can run in its cloud, the firm's cloud or fully on premises, with a hybrid option where only model processing leaves the firm.
Its published terms say customer data is not used to train the models. Named customers include Freshfields, Greenberg Traurig, Holland & Knight, Gunderson Dettmer, Cozen O'Connor, Homburger and Lenz & Staehelin. DeepJudge AG of Zurich was founded by former Google Brain researchers with AI doctorates from ETH Zurich.
Side by side on the same 15 capability axes and 12 legal signals: where the two part company, and what neither one publishes.
Vendors in the same categories as DeepJudge whose graded profiles look most alike, with a line on what each product does.
Similar is not the same, so check that each one does the job you need. No vendor pays for placement.
Ordered by how closely the graded profiles match across the 15 capability axes, inside the same categories. See how each axis is graded in the methodology.
All 15 axes, graded from public sources on the date shown. Hover a grade to see what the letter means on that axis.
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.
No content or conventional product sits underneath DeepJudge's AI. It licenses and publishes nothing of its own. The material it works on belongs to the customer and stays in the customer's systems. What DeepJudge supplies is the retrieval: search by intent that understands content, context and relevance, automatic sorting into a taxonomy, and detection of near duplicates and redlines across what it describes as billions of data points.
Without the models, what remains is connectors and a permissions sync, plumbing a firm would not buy on its own. The founders say search, rather than models, is their wedge. That is a claim about which machine learning matters, not that machine learning is peripheral.
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.
Grounding is built into the product, but no accuracy is measured. Every result is a document the firm already holds, in its own system, which the user can open and read in full, so checking is direct. The retrieval method is described: search by intent, automatic taxonomy sorting, detection of near duplicates and redlines, and linking of client, matter, document and people data. Section 4(j) of the subscription terms requires the customer to verify the accuracy, reliability and fitness of AI features, including through human review.
No accuracy figure, evaluation, test set or benchmark appears anywhere. Nothing addresses hallucination in the workflow and agent layer, which is where new text, rather than retrieved documents, is produced. The one nearby figure is satisfaction: 90 percent of users say they find results faster than with existing tools, which is not an accuracy claim.
What the system decides on its own, what a lawyer must approve, and whether the vendor documents where the review point sits. A tool that drafts under review and a tool that files without one are different products and different risks.
Human review is a contract requirement, and the excluded uses are named. Section 4(j) requires the customer to verify the accuracy, reliability or fitness of AI features through human review. The same clause bars use in judicial decision making by judges, in the work of lawyers serving as arbitrators, mediators or other dispute resolution neutrals, and in any use restricted under the EU Artificial Intelligence Act. Naming who may not use the product is more useful than the usual disclaimer.
What is not published is the operating boundary. The workflow layer is described as agents that carry out complex tasks in several steps, retrieving from many sources and taking actions. The customer builds those workflows itself with a low code builder, so the point at which a workflow acts without a person is set by the firm and described nowhere. Governance of agents is offered as a feature with no account of the controls.
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.
DeepJudge names many customers and publishes figures, but most figures are not tied to a named customer. Sixteen firms are named on the home page. They include global firms such as Freshfields, Greenberg Traurig, Holland and Knight and Gunderson Dettmer, and US mid market firms such as Cozen O'Connor, ArentFox Schiff and Vorys. Swiss and Austrian firms include Homburger, Lenz and Staehelin, Wenger Vieli, Advestra, Schoenherr and CMS Switzerland, plus an in house team at SBB.
Three named people speak on the record with title and firm, including two chief innovation officers. Published figures: 65 hours saved per user per year on searching, 85 percent adoption within two months, 23 queries per user per day, a fourfold return in the first year, and 90 percent reporting faster results. One figure is tied to a customer: a named M&A partner at Homburger says more than 80 percent of the firm's legal professionals use it in their work.
The headline figures are aggregate, with no method, and no dates appear on the home page. The individual customer story pages were not opened.
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.
Most confidentiality needs are met by contract and by design, but privilege is never addressed. Section 5(b) of the subscription terms says DeepJudge does not use customer data to train the AI or machine learning models behind the services, a contractual ban that names training. Separation is the product's design: permissions sync continuously with the firm's source systems so existing ethical walls hold at matter level.
Section 5(e) adds that a customer's own workflow logic will not be disclosed to or used for any other customer. Customer data is deleted within sixty days of termination. The firm chooses the models, and the hybrid deployment clause says exactly when processing leaves the firm's own infrastructure. Privilege and work product are not mentioned anywhere, and ethical walls address conflicts, not privilege.
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.
DeepJudge sets out its professional responsibility position in the operative agreement. Section 4(j) says DeepJudge is not a law firm and does not practice law, and the customer accepts it is not relying on the services for legal advice. It says who may not use the product: judges in judicial decision making, and lawyers serving as arbitrators, mediators or other dispute resolution neutrals. It addresses competence and supervision through a positive duty, requiring the customer to verify the accuracy, reliability or fitness of AI features, including through human review.
The legal boundary is stated twice. Section 4(a) limits use to what complies with the duties that apply where the customer and its users are located. Section 4(j) excludes any use prohibited or restricted under the EU Artificial Intelligence Act. The product searches the firm's own documents, not any country's law, so jurisdictional reach of legal content does not apply. No bar or law society guidance is engaged.
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.
One binding governance commitment exists, with nothing around it. Section 4(j) of the subscription terms makes compliance with the EU Artificial Intelligence Act a contractual limit on use, naming the law and excluding the uses in the administration of justice the Act treats as high risk. That is more than a principles page, and the customer can enforce it. The workflow layer is also sold partly on governance, offering to build, deploy, orchestrate and govern AI agents.
No content sits behind either. There is no responsible AI page, governance framework, accountable person or team, account of testing before release, or certification such as ISO 42001. Nothing addresses uneven output across matter types, practices or populations. ISO 27001 and SOC 2 Type II cover security, a different subject.
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.
DeepJudge publishes all five elements in specific terms, in its subscription terms and data processing addendum version 1.1 of May 2026. Retention and deletion: section 3(a) deletes all customer data, including training data and customer models, within sixty days of termination. DPA clause 11 adds prompt destruction or return on request, including data held by subprocessors, with confirmation. Access control: permissions sync continuously from source systems on least privilege principles, access is limited to authenticated users, audit logging is configurable, and the firm controls what is crawled at all.
DPA Annex 2 adds multifactor authentication, role based access control, central event logging and separate test and production environments. Subprocessors: Annex 3 names five entities with purpose and location. Clause 5 requires equivalent written terms, due diligence, DeepJudge's liability for their acts, thirty days' notice before a new subprocessor processes anything, and an objection right with termination without liability.
Incidents: clause 7 commits to notice without undue delay and within 72 hours of discovery, covering breaches at subprocessors too. The definition at 1.13 expressly includes unauthorized internal access, and DeepJudge must help the customer meet its own regulator and data subject duties. Encryption is TLS 1.2 or later in transit and AES-256 at rest, with annual penetration testing and SOC 2 Type II audited controls. Clause 3.4 bars subprocessors from selling or sharing data, processing it for other purposes, keeping or disclosing it outside the direct business relationship, or combining it with data from other sources.
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.
Loss is allocated specifically and mutually. Section 7(a)(ii) caps each party's total liability at the fees paid in the twelve months before the event. Four exceptions sit outside the cap: the customer's payment duties, breach of the use restrictions at section 4(e), fraud or willful misconduct, and intellectual property infringement. Section 7(a)(i) excludes indirect and consequential damages for both parties, with the same exceptions.
Section 7(b) gives the customer a real indemnity: DeepJudge defends third party claims that the services or software infringe intellectual property. The remedies are set in order: secure the right to continue, modify or replace the software, or terminate the affected part. There is also a service commitment. Section 1(a)(iii) sets a 99 percent quarterly uptime target, with downtime defined and six named exclusions. No warranty covers output, since section 6 disclaims accuracy and completeness, and no insurance is mentioned.
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.
For DeepJudge, integration is the product, and the depth is documented. It connects to the systems firm work lives in: the document management system, email, SharePoint, OneDrive, HighQ, intranets, and experience and metadata systems. Data flow is described both ways. Documents are crawled and indexed in place, with no migration or upload, so the data stays in its source system. Permissions flow the other way, synced continuously from those systems into the index, and the firm decides what is crawled at all.
Setup duties are in the contract. Section 1(d) requires continued access to the customer databases named on the order form and a named technical contact, and warns that withholding access degrades the service. The on premises clause makes the customer responsible for stated hardware and configuration requirements. The platform also gives governed access to other AI systems through the Model Context Protocol and publishes an open specification for handing off between agents.
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.
Three deployment models are defined in the agreement, with where data flows spelled out for each. Cloud deployment hosts the services in a location named on the order form. On premises deployment runs inside the customer's own infrastructure, with DeepJudge given access to install and maintain it and the customer meeting stated system requirements. Hybrid deployment is on premises with some processing, expressly including model processing, outside the customer's infrastructure.
The clause says customer data may then be processed in the cloud hosting location and that model processing may move customer data outside the customer's systems. That separates processing from storage, stated by DeepJudge rather than inferred, and names the one case where an on premises customer's data leaves. The security page adds that the customer runs its choice of model in its choice of residency. No regions are listed: residency is chosen and set on the order form rather than published.
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.
Two certifications are claimed with their scope, but no evidence a prospect can reach is offered. SOC 2 Type II is stated to cover security, availability and confidentiality, and ISO 27001 to cover the information security management system. The data processing addendum, version 1.1 of May 2026, records annual penetration testing in Annex 2 and says controls are audited for SOC 2 Type II. Missing: no auditor, certificate number, issue date, report date or coverage period is published, and there is no trust center or portal.
Evidence is for customers. DPA clause 9 says certifications and audit reports or written questionnaires come first. A customer audit is allowed only where those cannot show compliance, at the customer's expense, once a year, with no access to systems, hosting sites or infrastructure. That helps a signed customer but gives a buyer still deciding nothing to read.
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.
AI providers are named at the platform level, but no model vendor or model. Annex 3 names Microsoft Ireland Operations Ltd, Google Cloud EMEA Ltd and AWS EMEA SARL, each described as AI provider and cloud infrastructure, with location chosen by the customer. Clause 5 commits to at least thirty days' notice before any new subprocessor processes customer data, with a right to object on reasonable data protection grounds and termination without liability if the objection is not resolved.
These three are cloud platforms, not model vendors. No model house, model or default is named, so a reader learns which cloud estates run the AI but not whose model produced the output. The architecture and geography are well disclosed. DeepJudge says its architecture works with any model, and the customer chooses from and may combine any commercial or open source models, in its choice of residency. The hybrid deployment clause discloses that model processing may move customer data outside the customer's own infrastructure.
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.
The unit and the structure are published in the subscription terms without any figure. Fees sit on an order form and are stated to be in United States dollars unless otherwise agreed. The term structure is set out: subscriptions renew automatically for successive one-year periods unless either party gives three months notice before the end of a term, invoices issue within ten days of the effective date and each subsequent term and are payable within thirty days, and late payment attracts the greater of five percent per annum or the statutory maximum.
Professional services are separated cleanly, provided only under a mutually agreed statement of work that becomes part of the agreement, which tells a buyer that implementation is priced apart from subscription. Most usefully, the consumption unit is published: section 4(g) sets a hosting threshold of 500,000 documents per user or 250 gibibytes per user, above which DeepJudge may seek volume reduction or charge at standard rates. What is absent is any figure, any tier, and any pricing page; the only commercial route on the site is Book Demo.
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.
The customer list shows who buys the platform. It includes global firms such as Freshfields, Greenberg Traurig and Holland and Knight, and US mid market firms such as Cozen O'Connor, ArentFox Schiff and Vorys. It also includes technology specialist Gunderson Dettmer, Swiss and Austrian firms such as Homburger, Lenz and Staehelin, Wenger Vieli, Advestra and Schoenherr, and an in house legal and compliance team at SBB.
Teams are stated to be across North America and Europe, serving clients worldwide. Practice areas are not listed, which matters less here, because the platform searches whatever the firm already holds rather than a defined body of law. What is missing is the boundary. No firm size is named, nothing sets a minimum data estate or user count, and nothing says where the product stops. The clearest published limits are the barred users at section 4(j).
The public pages on file for DeepJudge, 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.
Client Data in Training, Prompt and Output Retention, Third Party Request and Subpoena Notice
Read Sep 2, 2026
Fabricated Citation Record
Read Sep 2, 2026
Ethical Walls and Matter Segregation
Read Sep 2, 2026
Structure and usage unit published, no figure. Fees are set on an order form under section 2 of the subscription terms, in US dollars unless otherwise agreed. Terms renew automatically for one year at a time unless either party gives at least three months' written notice before the term ends. Invoices are issued within ten days of the start of each term and are due within thirty days, with late charges at the greater of five percent a year or the legal maximum.
Professional services are priced separately under an agreed statement of work. The usage unit is set at section 4(g): a hosting limit of 500,000 documents or 250 GiB per User. Above that, DeepJudge may work with the customer to reduce volume or charge for the excess at its standard rates. Deployment is also a commercial choice, with cloud, on premises and hybrid options defined in the terms and the cloud location set on the order form.
There is no figure, tier or pricing page, and the only route on the site is Book Demo, so no entry price is shown. Verified 2 September 2026.
DeepJudge released a ChatGPT plugin that lets users search the firm's internal documents, retrieve source content and run search tasks from ChatGPT. OpenAI's example is bringing prior deals into a comparison.
A 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.
Can material a lawyer puts into this product be used to train a model?
The published terms prohibit training on customer content. Not a policy page, the agreement.
DeepJudge's subscription terms ban training on customer data. Section 5(b), headed No Training, bars using customer data to train the AI and machine learning models behind the services, unless the customer gives prior written consent. It names training and names the models. The consent exception lets the parties agree otherwise. It is not a product setting. Two nearby permissions are narrow. Section 5(c) allows Usage Data to improve the services, and defines it fully as telemetry: hardware and storage use, operating systems, cluster setup and health, uptime and response times, feature use, interactions and error logs, not shared with third parties.
Section 5(e) allows reuse of general workflow patterns only where no customer data is copied or disclosed. Clause 3.4 of the addendum limits processing to providing the services on written instructions. But Annex 1 describes processing as storage and other processing needed to provide, maintain and improve the services, and no clause of the addendum addresses training either way. The ban rests on section 5(b), the only provision that names training. A buyer should weigh Annex 1's wording on improvement alongside it.
How long does the product keep what a lawyer typed, and can that be set to zero?
A specific retention period is published and the customer cannot change it.
A fixed deletion period after termination applies, and the customer cannot change it. Section 3(a) of the subscription terms says DeepJudge deletes all customer data, including training data and customer models, within sixty days of termination. The customer in turn deletes DeepJudge software and data within thirty days and confirms in writing. DPA clause 11 adds deletion on request: prompt destruction or return of data, including data held by subprocessors, with confirmation.
Customer Data, defined at section 1(d)(i), covers data from the customer's databases, user Inputs and the resulting Outputs, so prompts and answers fall under both duties. No retention period during the term is published for Inputs and Outputs. What the customer controls is the footprint: documents stay in their source systems, and indexing controls decide what DeepJudge crawls, indexes and makes available.
Does retrieval respect the firm’s ethical walls, or can the model read across them?
Retrieval enforces the source system access model at query time, per user, and the vendor documents it.
DeepJudge enforces the firm's own access rules rather than building its own, and this is its central design. Permissions sync continuously from the source systems, which DeepJudge says ensures least privilege and maintains ethical walls. The access model enforced is the firm's own, checked against the source systems rather than rebuilt inside DeepJudge. Because documents are indexed in place and never moved, the document management system keeps authority over who may see what.
A customer confirms this on the customers page, describing search based on permissions so users only see what they are authorized to see. Two limits. Nothing says whether permissions are checked per user at query time or by periodic sync, and continuously synchronized fits either. And nothing describes what happens in the index between a permission change at source and the next sync.
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
Terms commit to notice where lawfully permitted. No transparency report located.
Notice of compelled disclosure, plus a request for confidential treatment, is a commitment in DeepJudge's terms. Section 8(b) says a party required by law, regulation, court order or legal process to disclose the other's confidential information will give prompt notice. It will also ask that anything disclosed be treated confidentially. Section 8(a) defines Confidential Information to include customer data, so this covers customer material.
The duty is mutual and survives termination. No transparency report was found, so nothing shows how many demands have been received or how they were handled. The website privacy policy allows disclosure to public authorities without a notice commitment, but it covers website visitors, not platform data.
Where does the law in this product come from, and does the vendor have the right to use it?
No located public material identifies the corpus behind the product’s answers.
The index covers only the firm's own documents, searched where they already live. No corpus is identified or claimed, because the product ships no content of its own. It works on the customer's material, pulled from the firm's document management system, email, SharePoint, OneDrive, HighQ, intranets and experience systems, indexed in place and never moved. Section 1(d)(i) of the subscription terms defines Customer Data as data from the customer's databases plus user inputs and generated outputs, and the customer warrants it has the right to share it.
Provenance and licensing questions therefore sit with the firm, not DeepJudge. Nothing was held back. There is simply no corpus to describe.
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.
Subsequent history and good law checks fall outside a search tool built on the firm's own files. No public material addresses subsequent history, treatment flags or good law checking, and none is claimed. The product does not retrieve primary law at all: it searches the firm's own precedents, advice, negotiation history and work product, so a citator has nothing to operate on. The adjacent risk that does exist is different in kind and is not addressed either, namely whether a precedent surfaced from the firm's own archive still reflects current law, which the platform leaves entirely to the lawyer reading it.
What does the product do when the answer is not in the corpus?
No located public material addresses what the product does when it cannot ground an answer.
Uncertainty handling is the open question for both search and workflows. Nothing describes what the platform does when it cannot ground an answer, and no confidence, relevance or grounding indicator is shown to the user. The workflow layer is called fully transparent, and search results are documents the user opens, so checking falls to the user rather than the system declining. Section 4(j) of the subscription terms puts that duty on the customer by contract, requiring human review of accuracy and reliability. That allocates responsibility but does not describe how the system behaves.
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.
Searched the AI Hallucination Cases database maintained by Damien Charlotin, and reporting drawing on it, on 2 September 2026, for both DeepJudge and DeepJudge AG. No court order, opinion or disciplinary record naming the product was found. This describes the public record, not the product. The platform retrieves the firm's own documents rather than citing legal authority, so a fabricated case reference reaching a filing is not the failure it creates. Its related risk is surfacing an outdated internal precedent, which no tracker records.
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
No located public material engages with bar or ethics guidance.
Professional conduct guidance for the lawyers who use the platform sits outside DeepJudge's materials. Nothing engages with guidance from any professional body over the product's users. Neither ABA Formal Opinion 512 nor any state bar, law society, or Swiss or Austrian bar guidance is named, though customers span US, UK, Swiss and Austrian practices. DeepJudge does make compliance with the EU Artificial Intelligence Act a contractual limit on use at section 4(j), excluding judicial and dispute resolution uses. That regulates the technology and its use, not the professional duties of the lawyers using it.
Does the vendor address what happens to the bill when the work takes an hour instead of six?
Public materials claim time savings without addressing billing or disclosure, and the product sits inside a fee relationship between a lawyer and a client where those savings would change the bill.
Time savings come with figures, and nothing is said about the bill. The home page cites 65 hours saved per user per year on searching, a fourfold return in the first year, and 90 percent of users finding results faster than with existing tools. Nothing addresses how AI assisted work is recorded, billed or disclosed to a client, and no per matter record of AI work was found. That matters because nearly all named customers are law firms that bill clients for the time this product saves.
The platform offers configurable audit logs, described as letting a firm monitor its use of DeepJudge. That is a security and oversight tool, not a matter level record of AI work.
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
A current subprocessor or model provider list is published.
DeepJudge publishes a current subprocessor list, but names no model vendor. Annex 3 names five entities with purpose and location. Microsoft Ireland Operations Ltd, Google Cloud EMEA Ltd and AWS EMEA SARL are each described as AI provider and cloud infrastructure, with location chosen by the customer. Intercom provides support from Dublin, and Gong is also listed. Governance of the list is strong. Clause 5 requires written terms at least as strict, due diligence on each subprocessor, liability for their acts, and at least thirty days' notice before a new one starts.
It adds a right to object, with suspension or termination without liability if unresolved. Clause 3.4 bars subprocessors from selling or sharing data, using it for other purposes, keeping or disclosing it outside the direct relationship, or combining it with other data. The gap is the model provider. The three AI providers named are cloud platforms, and no model vendor appears. A firm can tell its client which cloud estates process its content, but not whose model saw it. That partly follows from the design, where the firm picks the model, but the gap in what a firm can forward is real.
If a judge’s standing order requires an AI disclosure, can the product produce one?
No located public material addresses court disclosure or verification certification.
Court disclosure of AI use falls to the firm, with no supporting record built into the platform. Nothing addresses judicial standing orders, disclosure of AI use, or certification that output was verified by a person. Configurable audit logging exists and would record who searched for what, but it is presented as a firm oversight and security facility and nothing describes an exportable per document record capturing which model produced which passage, what was retrieved and who reviewed it.
The product class is relevant: the platform surfaces the firm's own documents rather than generating citations to authority, so the artifact a court would ask about is the work product built afterwards rather than the search itself.
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.