S
SixFifty

SixFifty 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. Three things sit on it. A Research module holds federal, state and local US employment law, written and maintained by SixFifty's own in-house attorneys, covering all 50 states and a published list of counties and cities. A document builder generates employee handbooks, offer letters, employment agreements, separation paperwork and policies by running the customer through a guided question flow that the legal team wrote, with an editor and version history afterwards; parallel Privacy Docs and Business Docs lines cover state privacy compliance and routine corporate paperwork.

A legal updates service tracks changes in the law, sends a monthly summary, and offers suggested redlines the customer merges into documents already generated. The AI is layered onto that: Ask SixFifty AI answers plain-language questions by retrieval over the Research database and links each answer back to the entries it came from, and an AI merge pushes the legal team's updates into a customer's own documents. All AI features run on hosted models from OpenAI, named on the estate.

Founded in 2019 with backing from the law firm Wilson Sonsini. Pricing is by number of employees and number of states, with unlimited users and no published figure.

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.

CC on AI CentralityArtificial intelligence is present but peripheral: a feature layer on a product whose value stands without it.

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.

The AI is real and it sits on top of a platform built by lawyers rather than by models. The legal process page sets out how the product is made: documents are sourced from expert lawyers, researched and drafted by SixFifty's own in-house attorneys, and delivered through a guided question flow that asks, in the vendor's words, the same questions a lawyer would; the in-house legal team then monitors federal and state change and writes the updates.

None of that is machine learning. Two features are: Ask SixFifty AI, which answers plain-language questions by retrieving from the vendor's Research database and linking back to the entries it used, and an AI merge that pushes the legal team's written updates into documents a customer has already generated. Remove both and the platform still sells a lawyer-built document generator, a maintained employment law database and a monthly update service.

A buyer should understand what the AI is for here: it makes an existing corpus easier to ask questions of, and it does not write the law or the documents. Verified 20 September 2026.

Source: Vendor Published
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.

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.

The grounding is documented and the accuracy is not measured. Ask SixFifty AI is described as retrieval-augmented generation over SixFifty's own Research database rather than over the open web, and every answer carries links to the Research entries it was drawn from so the reader can open them and cross-reference. The vendor says so directly in its FAQ, and contrasts it with general assistants and search engines on exactly that ground.

The database behind it is written and maintained by named in-house attorneys and covers all 50 states plus a published list of localities, and the vendor states that the assistant always works against the current version of it. Two limits. The links resolve to SixFifty's own summaries of the law rather than to the statute or regulation itself, so verification stops one step short of primary authority. And no accuracy figure, test set or error rate is published anywhere, while the Terms of Use disclaim accuracy entirely. Verified 20 September 2026.

Source: Vendor Published
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.

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 human sits in the middle of every step, and what the system does when it is unsure is not addressed. The review surfaces are real and described: documents are produced by a question flow the customer answers rather than by a model, the result opens in an editor where it can be changed, version history tracks what moved, and legal updates arrive as suggested redlines the customer chooses to merge rather than as silent edits to a live document.

Two written commitments back that up. The Terms of Use state that the AI is not intended or designed to replace or override human decision-making. The Privacy Notice goes further and bars a use outright: the AI features may not be used to make or substantially assist with any legal, financial or employment decision. That is a categorical limit, but it is a restriction placed on the customer rather than a behaviour the product enforces, and nothing published describes a confidence threshold, a point at which the assistant declines, or what happens after an answer turns out to be wrong. Verified 20 September 2026.

Source: Vendor Published
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.

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.

Named customers with real detail, and no measured result. Boise State University is the strongest: a named Associate General Counsel, the only member of her legal team working on employment law, managing documents for more than 4,000 employees across almost 40 states, with the handbook and employment agreements named as what she uses. FloQast is named with its own account of rebuilding an all-US handbook in 2021 and moving to a rip-and-replace approach on the monthly updates, quoted by a named VP of Human Resources.

Hatch appears with a cost comparison for new employee contracts. Customer logos run to a dozen recognisable companies and a dedicated universities page carries the Boise State story again. What is missing is the measurement: no before-and-after figure, no time saved, no method a reader could assess, and the case pages are undated on their face. The home page's claim of more than 500,000 businesses using SixFifty is published with no basis and is recorded rather than counted. Verified 20 September 2026.

Source: Vendor Published
CC on Privilege and Confidentiality PostureConfidentiality is asserted in general terms, or the commitment lives only in a sales conversation and cannot be read in advance.

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.

The training position is specific and the confidentiality position does not exist in anything a buyer can read. On training and model use the estate is unusually direct: the model provider is named as OpenAI, user-submitted prompts are stored in a secure internal environment for quality assurance and misuse detection, those stored prompts are not used to train or improve any model and are not shared with the provider for that purpose, personal information will not be used to train generative models without consent, and the vendor states it does not use personal information for automated decision-making or profiling.

Against that, the published agreement is a website terms of use with no confidentiality clause of any kind, no data processing terms, no retention or deletion commitment, and nothing on segregation between accounts, which matters because HR consultants and PEOs run many client companies through the same platform. Attorney-client privilege and work product are not mentioned. The Terms defer to a separate signed agreement that is not published. Verified 20 September 2026.

Source: Vendor Published
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.

UPL and Professional Responsibility Posture

Whether the vendor is clear that it supplies a tool rather than legal advice, who its audience is, and how it addresses unauthorized practice of law, competence and supervision duties, and jurisdiction limits. ABA Formal Opinion 512 is the reference point. Where the advice line is not the duty a product raises, the axis is read through the nearest professional duty it does raise: judicial conduct rules and the reviewing duty for products sold only to courts, and the duty to bill for time actually spent for products that draft time entries.

The advice line is drawn repeatedly, and the marketing spends its day walking up to it. The Terms of Use of 17 November 2025 state in capitals that information provided through the sites does not constitute formal legal advice and that use forms no attorney-client relationship, incorporating a separate legal disclaimer. The AI FAQ says the assistant provides legal information only and is not a substitute for legal advice.

The pricing FAQ says the team does not serve as the customer's personal legal counsel and recommends consulting their own attorney. The Privacy Notice goes furthest, barring use of the AI features to make or substantially assist with any legal, financial or employment decision. Jurisdiction is handled seriously too, with coverage published state by state and locality by locality. What holds this short of the top is the other half of the same estate: manage compliance like an employment lawyer without being one, be your own employment law expert, law-firm-quality documents, and a page arguing the product against hiring a law firm. Nothing addresses the supervision question that raises. Verified 20 September 2026.

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

AI Governance and Bias Disclosure

Published governance over model behaviour: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.

Positions are published; no mechanism behind them is. What exists sits in the Privacy Notice and the Terms of Use: the AI features are described as assistive technologies not designed to replace or override human decision-making, the vendor states it does not use personal information for automated decision-making or profiling, prompts are stored for quality assurance, usage monitoring and misuse detection, and the customer is told they are responsible for validating the accuracy and appropriateness of any output.

That is a and responsibility disclosure rather than governance over model behaviour. Nobody inside the company is named as accountable for it, nothing is published about what is tested before a change to the AI features ships, and there is no statement at all about whether output differs across jurisdictions, employer types or the protected characteristics that employment law is largely about, which is the bias question this particular product raises. The legal content itself is governed, by named attorneys; the model is not. Verified 20 September 2026.

Source: Vendor Published
CC on AI Safety and Data StewardshipA generic privacy policy covers the product without addressing what happens to documents and prompts after processing.

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.

One good disclosure about prompts sits inside an otherwise generic privacy notice. The good part: the Privacy Notice says what happens to what a user types into the AI, namely that prompts may be stored in a secure internal environment for quality assurance, usage monitoring and misuse detection, that they are not used to train or develop models, and that they are not shared with the model provider for that purpose. Everything around it is standard website language.

Security is described as reasonable efforts and administrative, technological and physical safeguards, with no controls named. No retention period is stated for prompts, generated documents or account data, and no deletion commitment appears anywhere. There is no subprocessor list for the service, no incident response practice, no breach notification commitment and no access control description. The named third parties on the estate are analytics and advertising partners, plus OpenAI for the models, rather than a list of who holds customer content. Verified 20 September 2026.

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

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.

The published position is a disclaimer, and it is wider than most. The Terms of Use of 17 November 2025 exclude liability for damages of any kind, under any legal theory, arising out of use of the sites or anything obtained through them, direct and indirect alike, and the exclusion is not backed by a fee-based cap the way most agreements are. The AI clause repeats the point on its own account: all AI is provided as is, with no warranty of any kind, and all liability relating to the AI, its output and the use of that output disclaimed.

The customer is made solely responsible for reliance on AI output, including its accuracy, applicability, reasonableness and bias. Indemnity runs one way, from the customer to the vendor. Claims must be brought within one year, through binding arbitration under consumer rules with a class waiver. No indemnity, warranty or insurance is offered for the loss a buyer would actually face here, which is a document or an answer that turns out not to comply. Verified 20 September 2026.

Source: Vendor Published
DD on Practice Systems Integration DepthNo integration into practice systems located, or the product stands alone and requires work to move to it.

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.

No integration into the systems this work already lives in is named anywhere on the estate. For a product that generates employment documents and tracks the law behind them, the obvious connections would be an HRIS, a payroll system, a document management system or a contract repository, and none appears in the navigation, the product pages or the footer. Sharing and electronic signature are built into the platform rather than connected to anything outside it, and the described route for a finished document is to generate it, edit it in the vendor's own editor and send it for signature from there.

Legal updates arrive inside the platform as suggested redlines against documents the platform itself holds. The Partnerships page is a reseller arrangement for HR consultants and PEOs rather than a technical integration, and no developer documentation or API is published. Checked the full navigation, all product pages and the footer on 20 September 2026. Verified 20 September 2026.

Source: Vendor Published
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.

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.

The region is answered and the tenancy is not. The Privacy Notice states that the services are hosted in the United States and are intended for visitors located within the United States, and tells a user outside the country that using the platform means transferring their information to the US for storage and processing. It then qualifies that in the other direction: the vendor may transfer data from the US to other countries or regions in connection with storage, processing and operating the services, without naming which.

Nothing is published about the tenancy model, so a buyer cannot tell whether accounts share infrastructure or sit apart, which matters most for the HR consultants and PEOs who run many client companies through one subscription. No single-tenant, private or regional option is offered, no cloud provider is named, and the model provider's processing location is not addressed. Verified 20 September 2026.

Source: Vendor Published
DD on Security Certifications and Trust CenterNo independent security attestation located.

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 independent security attestation was located, and the vendor does not claim one. There is no security page and no trust centre in the navigation or the footer, no SOC 2, ISO or penetration test reference in the Terms of Use or the Privacy Notice, and no badge anywhere on the estate. What the Privacy Notice offers instead is a statement of effort: reasonable efforts using administrative, technological and physical safeguards, followed by an acknowledgement that no information system can be completely secure.

A buyer cannot obtain a report, a scope or an audit period, because none is offered to request. The absence is recorded rather than inferred: the navigation and footer were run to the bottom and one search for a SixFifty attestation returned nothing from the vendor's own estate. This is a record of what is published, not a finding about the vendor's actual security. Verified 20 September 2026.

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

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.

The provider is named in plain terms, which most of this index's records do not manage. The Privacy Notice states that all AI features in the SixFifty platform are powered by integrations with hosted third-party models provided by OpenAI, and adds a commitment that runs with it: stored prompts are not shared with the third-party model provider for training, improvement or development. The Terms of Use acknowledge separately that a customer may be engaging with AI provided by SixFifty or by third parties.

What is not published is everything downstream of the name. No model or version is identified, nothing says where the model calls are processed or under what commercial terms, no retention position for the provider is stated beyond the training carve-out, and nothing commits to telling customers if the provider or the model changes, which on a single-provider architecture is the disclosure a buyer would most want. Verified 20 September 2026.

Source: Vendor Published
BB on Commercial TransparencyReal pricing is published for part of the range, with enterprise tiers withheld, or the unit and structure are stated without the figure.

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.

The unit is published and the number is not. The pricing page states the basis directly: pricing is set by how many employees a company has and how many states they are located in, so that a buyer pays for the footprint they actually have. It adds two useful structural facts, that all subscriptions except marketplace purchases allow unlimited users, and that access to document templates varies by subscription plan. Every route to a figure runs through a quote form; no rate, band, tier name or starting price appears anywhere, and no implementation or onboarding charge is stated.

Support is described as included, with a dedicated customer success manager meeting quarterly and in-app chat. The vendor's cost argument is comparative rather than numeric, that the documents cost less than commissioning them from a law firm, and one named customer calls the price immaterial against what the firm gets. A buyer can work out how they will be charged, and not what it will come to. Verified 20 September 2026.

Source: Vendor Published
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.

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.

Who it serves is set out by segment, and the legal coverage is documented to an unusual level of detail. Four buyer groups are named with what each gets: HR teams, consultants and PEOs serving multiple client companies, General Counsel offloading routine drafting and statutory monitoring, and startups. Public sector use is evidenced rather than claimed, through a universities page and a named state university customer.

Practice coverage is employment law across all 50 states, with federal, state and local levels, and the pricing page publishes the locality coverage explicitly, naming the California, Illinois and Minnesota cities and counties covered for paid sick leave and listing the specific policies covered in San Francisco, Chicago, New York City, Westchester County, Philadelphia, Seattle and Oakland. Two adjacent product lines, privacy documents and routine business documents, are named.

What is absent is any statement of company size boundaries, and a home page figure of more than 500,000 businesses with no basis behind it. Verified 20 September 2026.

Source: Vendor Published
Sources on file

6 public documents

The public pages on file for SixFifty, 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.

Pricing

Not published; quoted by employee count and number of statesUSD, as published, never converted

Unit published, figure withheld. **Charged by how many employees a company has and how many states they are located in**, stated as such on the pricing page and framed as paying only for the footprint the company actually has. **Unlimited users** on all subscriptions, with marketplace purchases expressly excluded from that rule. **Document access varies by plan**: the pricing FAQ says the templates available depend on the subscription, without naming the plans or what each contains.

Three product lines are sold from the same platform, Employment Docs, Privacy Docs and Business Docs, with no published relationship between them and the price. Every route to a number is a quote form or a demo request; no rate, band, tier name, minimum or starting price is published, and no implementation charge is stated. Support, including a quarterly customer success meeting, is described as included.

Implementation: None published as a separate charge. Support is described as included for all customers: a dedicated customer success manager meeting quarterly, email and scheduled meetings on request, and in-app chat. Questions touching regulations can draw in an employment or privacy specialist for coaching, also with no separate fee named, and the vendor states plainly that this is not legal representation. No onboarding, migration, configuration or training charge appears anywhere on the estate, and no minimum term is published.

Confidentiality and data terms: No Business Associate Agreement is offered or referred to anywhere on the estate, and none would ordinarily be expected: the platform holds a company's own employment policies, agreements and handbook content rather than health information. Worth recording for a reader who assumes otherwise: employment documents routinely touch medical leave, accommodation and disability policy, and the platform's HR case management module handles employee complaints, so sensitive employee information can reach it through ordinary use even though nothing here is sold or certified as a health system. The Terms of Use and Privacy Notice make no health-specific commitment, and no data processing addendum is published.

Note: No figure is published anywhere on the estate, so entryPriceUsd is null under R10 rather than zero: there is no free subscription tier, though free interactive maps and tools sit on the marketing site. What puts Commercial Transparency at B rather than C is that the unit is published in plain words on the pricing page itself, employees and states, along with the unlimited-user rule and the fact that template access varies by plan. No tier names, bands or starting rates appear, so the shape is visible and the number is not. One named customer calls the cost immaterial for what the firm gets, and another puts the alternative at several hundred dollars per employment contract from a law firm; both are recorded on the evidence row as customer statements rather than as pricing. A marketplace route exists and is mentioned only as an exception to the unlimited-user rule.

Legal Signals

What each signal means

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.

Confidentiality and Privilege

Client Data in Training

Can material a lawyer puts into this product be used to train a model?

Permitted, in the contract

The published agreement expressly reserves a right to train on customer content, with no opt out located. Any de identification, anonymisation or aggregation qualifier is recorded in the summary.

Training is named in the agreement, expressly permitted, and narrowed everywhere except the agreement. The Terms of Use of 17 November 2025 provide that SixFifty may use information you provide to develop, train and improve its services, business processes and AI, as otherwise permitted by those terms, with your consent, or as authorised by the Company that bought the subscription. That last route matters: the employer can give the consent on the user's behalf.

The Privacy Notice of the same date pulls two things back out. Prompts submitted to the AI features may be stored for quality assurance, usage monitoring and misuse detection, and those stored prompts are not used to train, improve or develop any model and are not shared with the model provider for such purposes. And personal information will not be used to train generative models without consent. Both narrowings are policy statements sitting against a contractual grant, which is the order that decides this value.

Source: Vendor Publishedmay use information you provide to develop, train, and/or improve our services, business processes, AIAs of Sep 20, 2026Evidence

Prompt and Output Retention

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

Disclosed without a period

Retention is acknowledged in public materials with no stated period.

Prompt retention is addressed head-on and never given a length. The Privacy Notice says what most estates leave out: prompts entered into the AI features may be stored, that storage is in a secure internal environment, and the reasons are quality assurance, usage monitoring and detecting misuse. It then adds the two commitments that go with it, that the stored prompts are not used to train, improve or develop any model and are not shared with the third-party model provider for those purposes.

What is missing is the period. No retention window is published for prompts, for the answers returned, or for the documents a customer generates, no deletion practice is described, and no setting is offered to a customer who wants a shorter window or none. The general retention language elsewhere in the notice covers personal information rather than product content.

Source: Vendor PublishedSixFifty may store user-submitted prompts ... in a secure internal environmentAs of Sep 20, 2026Evidence

Ethical Walls and Matter Segregation

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

Not addressed

No located public material addresses walls or matter level segregation.

Nothing published describes how one account's work is kept apart from another's. The Terms of Use cover account credentials and tell a user not to share their account, which is authentication rather than segregation, and the Privacy Notice describes security only as reasonable administrative, technological and physical safeguards. No tenancy model, no role-based access description, no statement that one customer's documents and answers are isolated from another's. The gap has a specific edge on this product, because the vendor sells to HR consultants and professional employer organisations expressly so they can create documents and policies for many client companies, and nothing says how those client workspaces are separated from each other inside one subscription. Checked the terms, the privacy notice, the consultants page and the product pages on 20 September 2026.

Source: Operator VerifiedAs of Sep 20, 2026Evidence

Third Party Request and Subpoena Notice

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

Disclosure addressed, notice absent

Published terms or policy address disclosure to authorities or in response to legal process, and no commitment or reservation regarding customer notice is located anywhere. The vendor has told the customer that data can leave and has said nothing about whether the customer hears of it.

Disclosure is provided for and notice is not. The Privacy Notice says SixFifty may access, preserve and disclose information where it believes doing so is required or appropriate to comply with law enforcement requests and legal process such as a court order or subpoena, to respond to requests, or to protect rights, property or safety. Nothing anywhere commits to telling the customer that such a demand has arrived, to waiting before complying, to narrowing what is handed over, or to helping the customer seek a protective order.

The Terms of Use carry no confidentiality clause at all, so the compelled-disclosure provision that usually supplies the notice commitment is absent, and the published agreement defers to a separate signed agreement that is not available to read. No transparency report is published.

Source: Vendor Publishedcomply with law enforcement requests and legal process, such as a court order or subpoenaAs of Sep 20, 2026Evidence
Accuracy and Authority

Primary Law Corpus Provenance

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

Sources named, basis unstated

Sources are identified without stating the licence or rights basis.

The corpus is the vendor's own and it says how it is made, without saying what it rests on. The legal process page describes four steps: best-in-class legal documents sourced from expert lawyers as the foundation, extensive research through precedent documents so the options are considered, jurisdictional and circumstantial analysis by the in-house legal team to build the question flow, and continuous monitoring of federal and state change to keep the documents current.

The Research database that the AI answers from is described the same way, as up-to-date US employment law maintained by SixFifty's in-house legal experts across all 50 states and a published list of localities. The legal team is named, with biographies. What is not stated is the licensing basis for the sourced precedent documents, or which primary sources the research entries are built from, so the chain from a statute to an answer is described in process terms rather than in citations.

Source: Vendor PublishedAs of Sep 20, 2026Evidence

Good Law Verification

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

Own treatment signal

The vendor computes and surfaces subsequent history itself, with the method described.

There is no case-law citator here, and there is a real currency mechanism doing the equivalent job for statutes and regulations. The vendor's in-house legal team monitors federal and state law changes continuously, writes the updates itself, and the platform tells a customer when a change affects a document they have already generated, offering the revised language as a redline they merge. A monthly Legal Update summary goes out by email and the changes are also recorded in the Research module, so a customer can see what moved and when.

The assistant is stated to work against the current version of the database at all times. Two limits worth naming: nothing published describes how quickly a change is reflected after it is enacted, and nothing flags a document as out of date if the customer declines the update, so currency depends on the merge being accepted.

Source: Vendor PublishedAs of Sep 20, 2026Evidence

Refusal and Uncertainty Behaviour

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

Not addressed

No located public material addresses what the product does when it cannot ground an answer.

Nothing published says what the assistant does when it does not know. The AI pages describe what it answers, how fast, and where the answer came from, and the FAQ addresses coverage by topic and jurisdiction, but there is no description of behaviour at the edge of the database: no statement that it says so when a state or locality is not covered, no confidence signal, no abstention, and nothing about a question that falls outside employment law altogether.

The published material points the other way instead, telling the customer that accuracy is not guaranteed and that they are responsible for validating any output. That places the whole burden of detecting an uncertain answer on the reader, on a product whose coverage is explicitly uneven by locality. Checked the AI page and its FAQ, the research product page, the terms of use and the privacy notice on 20 September 2026.

Source: Operator VerifiedAs of Sep 20, 2026Evidence

Fabricated Citation Record

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

None located

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 record was located of this product's output being found fabricated or inaccurate in a proceeding, a regulatory action or a published account. Searches on 20 September 2026 across the vendor's estate, press and directory profiles returned nothing of the kind. The exposure is a different shape from a litigation drafting tool: the assistant answers questions about statutes and regulations rather than citing cases, and the documents are generated from templates written by attorneys, so the failure a buyer should watch for is a policy that is out of date or wrong for a jurisdiction rather than an invented authority.

Source: Operator VerifiedAs of Sep 20, 2026
Professional Responsibility

Bar Guidance Alignment

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

Generic reference

Public materials refer to professional responsibility in general terms without naming guidance.

Professional responsibility is engaged in general terms, four times over, and no rule or opinion is named. The Terms of Use of 17 November 2025 state in capitals that information provided through the sites does not constitute formal legal advice and that use of the sites forms no attorney-client relationship with SixFifty, incorporating a separate legal disclaimer by reference. The AI FAQ says the assistant is not a substitute for legal advice and provides legal information only, while noting it may reduce routine calls to the customer's attorney.

The pricing FAQ says the team does not serve as the customer's personal legal counsel and recommends consulting their own attorney for advice on their situation. Nothing names a rule of professional conduct, an ethics opinion or any bar guidance, and nothing engages the guidance on generative AI, which is a live question for a platform selling legal document generation to non-lawyers.

Source: Vendor PublishedAs of Sep 20, 2026Evidence

Billing and Fee Posture

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

Outside the fee relationship

The product does not touch a fee between a lawyer and a client. It operates before an engagement exists, or it is bought by a team that bills no client for the work. Savings claims aimed at the buyer’s own cost are recorded in the summary and do not make the row a savings claim, because no client bill is in the loop.

There is no lawyer's fee in this product's path. The buyer is the company itself, or an HR consultant or professional employer organisation acting for its own client companies, and the subscription is an operating cost set by headcount and state footprint rather than anything billed on to a client of a law firm. The vendor's commercial argument is a substitution one, that generating a document on the platform costs less than commissioning the same document from a law firm, and one named customer puts the saving at hundreds of dollars per employment contract.

That is a claim about replacing legal spend rather than about disclosing how a fee was earned. Nothing here bears on what a client is told about machine-assisted work, because no client bill is involved.

Source: Vendor PublishedAs of Sep 20, 2026Evidence

Outside Counsel Guideline Readiness

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

Not addressed

No located public material supports a client side disclosure obligation.

A buyer working through a diligence checklist would find one answer and no pack. The model provider is named, which is the one item many estates omit: all AI features run on hosted models from OpenAI. Everything else is missing. No subprocessor list for the service exists; the named third parties in the Privacy Notice are analytics and advertising partners, Google, Meta, LinkedIn, Microsoft, Salesforce, Default and Pendo, rather than the parties holding customer content.

There is no security certification, no penetration test reference, no data processing addendum, no confidentiality clause in the published agreement, no retention or deletion schedule, and nothing offered on request. Checked the terms of use, the privacy notice, the full navigation and the footer on 20 September 2026.

Source: Operator VerifiedAs of Sep 20, 2026Evidence

Court Disclosure Support

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

Not addressed

No located public material addresses court disclosure or verification certification.

Nothing published addresses disclosing the machine's involvement to anyone outside the company. The product does keep a record of a kind, in that documents carry version history and legal updates are applied as tracked redlines a customer accepts, so a firm could show how a handbook policy reached its current wording. But that record is about document changes rather than AI participation: nothing distinguishes text the assistant produced from text the attorney-written template produced, no export or format is described, and nothing addresses a tribunal, an agency or an opposing party asking how a policy was assembled.

The question is less remote than it looks, because these documents are the ones produced in employment litigation and agency charges. Checked the product pages, the terms of use and the privacy notice on 20 September 2026.

Source: Operator VerifiedAs of Sep 20, 2026Evidence
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 303 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
September 20, 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