Tradespace
Tradespace is an intellectual property management platform that runs the patent lifecycle from invention capture through to licensing. Its Create module replaces disclosure forms by turning whatever an inventor already has, including slide decks, product specifications, flow charts, manuscripts and audio recordings, into structured disclosure packages, and it connects to the tools engineers work in such as Slack, Gmail and Notion to surface ideas that would otherwise stay buried. Disclosures arrive with prior art review, related portfolio matches, detectability and market assessments already attached, and no-code rules route them through patent committee review. Protect covers drafting and prosecution, with a docketing assistant that captures patent office and outside counsel correspondence, summarises each communication, creates the resulting deadlines and suggests responses. Manage handles annuities and maintenance decisions with portfolio reporting that maps assets to products, revenue and competitive position, and Commercialize identifies likely licensees, builds claim charts and runs licensing campaigns. Alongside the software Tradespace offers flat-fee prosecution through a vetted network of outside patent firms, so a team can use the platform alone or with counsel attached. The architecture is built around the disclosure risk particular to patent work: each customer account is given its own dedicated model instance under a hub-and-spoke design, with no data passing between instances, and retrieval-augmented generation is used to ground responses and supply citations. The platform runs on Microsoft Azure with servers in the US Northern Virginia region and models reached by default through Azure OpenAI Service. Customers include Georgia Tech, the University of Texas at Austin, UNC Charlotte, the FDA, Raytheon, Northrop Grumman, Woodward, Unity and Enveda Biosciences. Tradespace, Inc. is based in San Francisco and acquired the patent drafting startup Paragon in November 2025.
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.
A conventional IP management system sits underneath a substantial and genuinely load-bearing AI layer. The substrate is a system of record: docketing, annuity tracking and maintenance cost modelling, portfolio composition reporting, workflow routing and eSignature-enabled assignments, all of which are what an IP management system has sold for decades and none of which needs a model. What the AI adds is real and is what the marketing leads with: disclosure packages generated from decks, specifications, flow charts and audio, invention reports summarising use cases, prior art and competitive advantage, patent drafting acquired with Paragon, a docketing assistant that summarises office correspondence and suggests responses, and licensing analysis that identifies likely licensees and builds claim charts. The company's own framing is that the AI is built into the architecture rather than added on, and a press announcement describes the product as a secure system of record combined with drafting technology, which is an accurate description of a platform where both halves are real. Verified 2 September 2026.
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 mechanism is named and the question is answered directly rather than deflected. Asked on its own AI page how it prevents hallucinations, Tradespace states that it uses retrieval-augmented generation to limit responses to actual data and provides citations for responses, alongside guardrails intended to keep generated content grounded in fact. The Paragon acquisition announcement describes the drafting technology as offering transparent traceability to source materials so that teams can verify every claim, citation and technical detail. Naming the architecture rather than asserting accuracy is what puts this in the middle band. What is absent is measurement of any kind. No accuracy figure, benchmark, evaluation or test set appears anywhere, and the one number attached to output quality comes from a customer rather than the vendor, with the Director of IP at UT Austin quoted saying the platform gets his team 60 to 70 per cent of the way there, the equivalent of a first draft. That is a useful and candid figure and it is a testimonial rather than a measurement. The description of guardrails as state-of-the-art carries no content. Verified 2 September 2026.
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.
Review is asserted through the service layer and never described as a product control. The strongest oversight signal is commercial rather than technical: prosecution work runs through senior attorneys in a vetted network, and the platform page names attorney review and workflow approvals as stages. Inside the product, the customer can build no-code rules for disclosure routing and patent committee review, which is a configurable approval path a firm sets for itself. Against that, the docketing assistant is described as capturing office and counsel correspondence, summarising it, creating the resulting deadlines and suggesting responses, which is a system generating docket deadlines on a matter where a missed date is malpractice, and nothing published states whether a human confirms a created deadline before it enters the docket. No published material states what any module completes without review, where the checkpoint sits for a customer using the software without the attorney network, or what happens when the system is uncertain. There is no agreement of any kind published, so no review obligation exists in any document a buyer could rely on. Verified 2 September 2026.
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.
The reference base is broad, named and unusually well spread across sectors that are hard to win. Logos on the home page run to Georgia Tech, the University of Texas at Austin, the FDA, Enveda Biosciences, Unity, UNT Health and Raytheon, and the security page adds Northrop Grumman, the Department of Defense and the Department of Energy, stating that the DoD and DOE were founding customers. Two figures attach to named customers rather than floating free: UNC Charlotte is described as clearing a three-year technology backlog and posting 100 new case submissions in a few months, and Chun Kuo, Director of Intellectual Property at UT Austin, is quoted with both an outcome and a consequence, saying the platform gets his team 60 to 70 per cent of the way there and that without it he would need budget for more staff. Video case studies exist for Woodward and UNC Charlotte. A company announcement adds scale figures of more than 440,000 patents managed for over 80 organisations including 75 per cent of top US research universities. What holds this below the top band is that nothing is dated and no figure carries a method or measurement basis. Verified 2 September 2026.
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.
One limb is met better than anywhere else in this corpus and two are absent entirely. The segregation architecture is described mechanically rather than claimed: a hub-and-spoke design provisions a separate instance of the model for each customer account, and the vendor states that no data is ever transferred from a spoke model to the hub or to another spoke. Tradespace also frames the confidentiality question in the terms that actually matter for patent work, addressing novelty rather than only secrecy, and stating that data uploaded does not constitute a public disclosure and is not discoverable by a third party. Training is addressed as an opt-in, with no training by default and only on explicit customer request. Model handling is disclosed, with Azure OpenAI Service named and Microsoft's data protection addendum linked. What is missing is fundamental. No retention or deletion position exists anywhere: nothing states how long disclosures, drafts or office correspondence are kept, or what happens to them when a customer leaves. Neither privilege nor work product is mentioned, which matters because the platform is sold to general counsel, chief legal officers and law firms as well as to R&D. And there is no published agreement, privacy policy or data processing addendum in which any of these commitments could be held. Verified 2 September 2026.
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.
Nothing published addresses the professional responsibility questions, and this vendor raises more of them than most. Checked the home page, platform overview, Create, Manage, Enterprise AI and Security and Privacy pages, the counsel network page and the full site footer on 2 September 2026. There is no statement that output is not legal advice, nothing on who may use the platform or under what supervision, no reference to verification duties, and no jurisdictional statement. There is also no terms of service and no privacy policy anywhere on the site, so no document exists in which such a statement could sit. The gap is sharper here than on a pure software record because Tradespace markets itself as a full-service patent practice combining software with on-demand patent attorneys drawn from a vetted network of outside firms. That structure raises questions a buyer would expect answered somewhere: who the client is, whether an attorney-client relationship arises with Tradespace or only with the network firm, how conflicts are cleared across a network serving competing portfolios, and who bears the duty of candour on a filing the platform helped draft. None is addressed on any published surface. Verified 2 September 2026.
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.
No governance material was located on any surface. Checked the home page, platform overview, Create, Manage, Enterprise AI and Security and Privacy pages and the full footer on 2 September 2026. There is no responsible AI page, no governance framework or set of principles, no individual or function named as accountable for model behaviour, no account of what is evaluated before a model or module change ships, and no certification such as ISO 42001. The nearest statement is a claim that the AI incorporates state-of-the-art guardrails to keep generated content grounded in fact, which names no guardrail and describes no testing. SOC 2 Type 2 maintained through Vanta is an information security attestation and is treated separately by this axis. Nothing anywhere addresses uneven output, which is a live question on a platform whose disclosure evaluation module scores inventions and whose licensing module ranks assets, since both produce judgements that shape which inventions get filed and which patents get monetised. Verified 2 September 2026.
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.
Protection and isolation are described well and three of the five elements are absent. What is published is specific: encryption of all data at rest using AES-256, single sign-on through Okta, Microsoft and Google with others available on request, SOC 2 Type 2 compliance maintained through Vanta with a linked trust centre, hosting on Microsoft Azure with all servers stated to be in the US Northern Virginia region, and per-account dedicated model instances under the hub-and-spoke design. Against that, no retention period or deletion practice is stated anywhere for any category of data. No subprocessor list exists; Microsoft is identifiable as the cloud and model route and no other processor is named. And no incident or breach notification practice was located, with nothing stating whether or how quickly a customer would be told. The underlying reason for all three gaps is the same and is worth stating plainly: Tradespace publishes no terms of service, no privacy policy and no data processing addendum, so the documents in which retention, subprocessors and breach notification are normally found do not exist on the site. Verified 2 September 2026.
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.
Nothing published addresses who bears the loss when the output is wrong, because no agreement of any kind is published. Checked the home page, platform overview, Create, Manage, Enterprise AI, Security and Privacy and counsel network pages and the complete site footer on 2 September 2026: the footer carries Solutions, Our Platform, The Tech, Customers, Company and IP Resources, and there is no legal section, no terms of service, no privacy policy, no acceptable use policy and no data processing addendum. Two targeted searches returned nothing from the company either. No liability cap, indemnity, warranty, service level or insurance position exists on any reachable surface. The one commercial commitment published is a satisfaction promise rather than a liability term: on a first filing Tradespace states that if the customer is not happy it will not file and the materials are theirs regardless. For a platform holding unfiled invention disclosures for defence contractors, federal agencies and research universities, the absence of any published agreement is the single most consequential gap in this record. Verified 2 September 2026.
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.
Integration is a genuine strength and it points in an unusual direction. Rather than connecting to the systems lawyers use, Tradespace connects to the systems inventors use, naming Slack, Gmail and Notion as sources it mines to surface ideas, and describing capture from pull requests, call recordings, schematics and design documents. That is the right architecture for a product whose bottleneck is disclosure capture. On the enterprise side the platform states support for integrations with CRM, CLM, ERP and other IP systems through APIs and connectors, with secure bidirectional exchange of data, which names the direction of flow rather than just asserting connectivity, and the Manage module describes integrations with annuity providers and AI-powered integrations with patent offices that capture maintenance events, deadlines and correspondence automatically. Single sign-on runs through Okta, Microsoft and Google. What is missing is documentation an implementer could use: no connector list, no named IP management or docketing system, no API reference and no developer surface was located. Verified 2 September 2026.
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 tenancy answer is better than the residency answer, which is the reverse of most records here. Isolation is described concretely: each customer account receives a dedicated instance of the model under a hub-and-spoke design with no data moving between instances, and the vendor describes private instances as the mechanism that precludes public disclosure, which ties the architecture to the specific legal risk of patent work. Hosting is stated as Microsoft Azure and the location is given precisely, with all servers stated to be in the US Northern Virginia region. What is absent is choice. No alternative region is offered, no European or other option is mentioned, and no on-premises or customer-cloud deployment exists, so a customer with data residency obligations outside the United States has no published path. Nothing distinguishes processing location from storage location. The single-region answer is unusually specific and unusually inflexible at the same time, and for a platform serving research universities and multinationals filing in Europe and Asia the absence of any non-US option is worth a buyer's attention. Verified 2 September 2026.
Security Certifications and Trust Center
Independent attestation a buyer can pull without a sales call: SOC 2, ISO 27001, penetration test summaries, a trust center with current reports and named scope rather than a badge image.
The attestation is real, named and reachable. Tradespace states that it partners with Vanta to maintain SOC 2 Type 2 compliance and links directly to a customer-specific trust centre, which is a route to controls rather than a badge on a page, and the security posture is described in specifics elsewhere with AES-256 at rest, named single sign-on providers and a stated server region. The claim is also contextualised by the customer base, with the company stating it works to these standards with clients including Northrop Grumman, Raytheon, the Department of Defense and the Department of Energy. What the top band asks for is not established. No auditor or certifying firm is named, since Vanta is the compliance automation vendor rather than the auditor. No report date, coverage period or scope is published, no trust services criteria are identified, and the access tier for the underlying report is not stated. Vanta-hosted trust centres render client-side and returned no readable content, which is recorded as a retrieval limit rather than an absence. No penetration testing programme is described. Verified 2 September 2026.
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 and the models are not. Asked directly which models it uses, Tradespace states a model-agnostic approach and says that by default it works with Microsoft Azure to access their OpenAI Services, linking to Microsoft's published data privacy terms for that service. That tells a reader whose infrastructure and whose model family sit underneath, and pointing at the applicable data protection terms rather than merely naming a vendor is more useful than most disclosures in this corpus. Where inference runs is stated through the hosting answer, on Azure with servers in the US Northern Virginia region, and the hub-and-spoke design tells a reader that the instance is dedicated rather than shared. What is absent is the specificity the top band requires. No model or version is identified, the model-agnostic claim is not accompanied by any list of what else can be plugged in, and no commitment exists to notify customers when the model or provider changes, which matters precisely because the architecture is designed to make swapping models easy. Verified 2 September 2026.
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 charging model is stated and no number is attached to it. What a buyer learns without contacting anyone is the shape: prosecution is sold on a flat fee, with the vendor stating that the customer knows the cost before the work begins, and the whole commercial pitch is framed against the alternative of hourly outside counsel, with references to the law firm markup and to paying hundreds an hour for correspondence handling. There is a concrete free entry offer, with Tradespace preparing a full first application at no cost and stating that if the customer is not happy it will not file and the materials are theirs regardless. An ROI calculator is published for a buyer to model returns against their own portfolio. What is not published is any figure, tier, seat price, subscription rate or unit for the software itself, and nothing states how the platform licence relates to the flat prosecution fee or whether one can be bought without the other. The only route to a price is a demo request. Verified 2 September 2026.
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.
Segment documentation is among the most thorough in the index. Six buyer roles each have their own page, covering IP leaders, tech transfer leaders, general counsels, chief legal officers, R&D leaders and corporate development, and five industries each have one, covering corporations, startups, universities, government and law firms. Four use cases are separately documented. That is a vendor that has thought carefully about who buys and has published the answer rather than leaving it to be inferred, and the customer roster bears it out across a federal agency, defence primes, research universities and venture-backed startups. Coverage of law firms as a buyer alongside in-house teams is notable given the product is also marketed as an alternative to outside counsel. What is missing is the boundary. No jurisdiction is stated anywhere, so nothing says whether the drafting and docketing modules handle European or international filings or only US practice, which the single US hosting region makes a live question. No portfolio size band is given and nothing states what the platform does not cover; trademarks and copyright are absent throughout without being excluded in terms. Verified 2 September 2026.
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?
Training occurs only where the customer has affirmatively enabled it.
The structure is an opt-in rather than a prohibition, and the vendor is explicit about it. Asked on its own AI page whether it trains on customer or account data, Tradespace answers that by default it does not, and adds the quoted sentence, so training is available and switched on only at the customer's request. The security page states the same commitment in a slightly different frame, that data is never used for training without written consent, which adds a form requirement to the mechanism. A third statement on the AI page goes further than either and sits awkwardly beside them: asked whether using the platform constitutes a public disclosure, the vendor answers that customer data does not train any model, stated flatly and without the consent qualifier. Read together, the position is opt-in with written consent, and the unqualified sentence is best understood as describing the default rather than adding a prohibition. No agreement, privacy policy or data processing addendum is published anywhere on the site, so none of this is a contractual term and there is no document in which the consent mechanism is defined.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
No located public material states how long prompts and outputs are retained.
Checked the home page, platform overview, Create, Manage, Enterprise AI and Security and Privacy pages and the complete site footer on 2 September 2026. Nothing addresses retention of any kind. No period is stated for disclosures, generated drafts, prompts, office correspondence or portfolio records, nothing describes what happens to data when a customer stops using the platform, and no deletion route or export right is published. No retention setting is offered. The reason is structural rather than an oversight in one document: Tradespace publishes no terms of service, no privacy policy and no data processing addendum, so the documents in which a retention position normally lives do not exist on the site, and two targeted searches returned nothing from the company. What is published nearby concerns isolation rather than duration, with each account given a dedicated model instance and no data moving between instances, which addresses who can reach the data rather than how long it is kept.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
The product maintains its own permission model, documented, requiring the firm to keep it aligned.
This is the first record in the pull to reach this value, and the mechanism is described rather than asserted. Asked how it ensures account data sent to AI models is not co-mingled, Tradespace describes a hub-and-spoke approach in which it provisions a separate instance of its hub model for each customer account, giving the quoted result, and states that no data is ever transferred from a spoke model to the hub model or to any other spoke model. The security page frames the same architecture against the risk that actually matters in patent work, stating that a private model per customer is what ensures uploaded data does not constitute a public disclosure. That is a documented separate-model design with the direction of data flow stated in both directions. Two limits belong on the record. The separation described is between customer accounts, and nothing addresses separation between matters or between clients inside a single account, which is the question a law firm using the platform would ask. And none of it appears in any agreement, since none is published.
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.
Checked the home page, platform overview, Create, Manage, Enterprise AI and Security and Privacy pages and the complete site footer on 2 September 2026, and ran two targeted searches for company legal documents. Nothing addresses disclosure to authorities or in response to legal process, and nothing addresses notice to the customer if a demand arrives. The question is not reached rather than answered adversely, and the reason is that no terms of service, privacy policy or data processing addendum is published at all. The nearest published statement concerns third parties in a different sense, with the vendor stating that customer data is not publicly disclosed or otherwise discoverable by a third party, which is a claim about the platform's architecture and confidentiality rather than about compelled disclosure. The gap is worth naming on this record because the customer base includes the Department of Defense, the Department of Energy, the FDA and defence contractors, where government demands for records are a foreseeable event rather than a hypothetical.
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 licence or rights basis.
The training corpus is described by category and no source or rights basis is given for any of it. Tradespace states that its AI tools were trained on knowledge gathered from millions of publications, patents, product specs, licenses, and search reports. Two of those categories are public records in most jurisdictions, and three are not: product specifications, licence agreements and search reports are ordinarily confidential or commercially licensed material, and nothing published states where they came from, on what basis they were obtained, or whether any of it originated with customers. That question sits directly against the vendor's separate statement that it does not train on customer or account data without explicit request, and a reader is left to reconcile the two without help. No database, publisher, patent office or data supplier is named, no jurisdictions are listed, no volume figure is given for any category, and no update cadence is stated. The platform separately consumes live patent office correspondence and prior art in operation, and nothing describes the provenance of that either.
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.
Checked the home page, platform overview, Create, Manage, Enterprise AI and Security and Privacy pages on 2 September 2026. Nothing addresses whether material surfaced by the platform remains current or good. The question takes a particular form in patent work and is unaddressed in each of its forms: nothing states whether prior art returned in a disclosure evaluation is checked for currency, whether patent status and legal events are validated against office records when the Manage module reports on a portfolio, or whether the licensing module confirms that an asset it recommends for monetisation is in force and unencumbered. The docketing assistant does capture patent office communications and create deadlines from them, which is currency in the procedural sense, but nothing describes verification of what it captured. No treatment, status or validity signal is described anywhere.
Refusal and Uncertainty Behaviour
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.
Checked the home page, platform overview, Create, Manage, Enterprise AI and Security and Privacy pages on 2 September 2026. Nothing describes what the platform does when it cannot ground an answer, and no confidence, coverage or grounding indicator is described as shown to the user. What is published is a constraint on generation rather than an account of behaviour at its limits: the vendor states that retrieval-augmented generation is used to limit responses to actual data and that citations are provided, and that guardrails keep generated content grounded in fact. One product behaviour comes closer than the rest and is recorded here, since the Create module is described as flagging missing information in an inventor's submission, which tells a user the input is incomplete rather than telling them the model is uncertain about an output.
Fabricated Citation Record
Does a public court record exist involving output from this product?
No court order, opinion or disciplinary record naming this product has been located as of the date shown. This is a statement about the public record, not a finding about the product.
Searched the AI Hallucination Cases database maintained by Damien Charlotin, and reporting drawing on it, on 2 September 2026 on the product and corporate name Tradespace and Tradespace, Inc., and on the acquired drafting brand Paragon. No court order, opinion or disciplinary record naming the product was located. This is a statement about the public record rather than a finding about the product. The tracker records fabricated citations in court filings, and this platform's outputs are patent applications, office action responses and docket entries filed at patent offices, so the forum that would surface an equivalent failure is the USPTO rather than a court, and its responses take the form of examiner objections, prosecution history consequences or an inequitable conduct allegation raised later in litigation rather than a published sanctions order. None of those is indexed anywhere comparable.
Bar Guidance Alignment
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.
Checked the home page, platform overview, Create, Manage, Enterprise AI, Security and Privacy and counsel network pages and the complete site footer on 2 September 2026. No public material engages guidance from any professional body governing the platform's users. Neither the USPTO Rules of Professional Conduct nor its guidance on the use of artificial intelligence in filings is named, no state bar opinion is cited, and nothing addresses the duty of candour that attaches to a patent filing or the competence and supervision duties of a practitioner relying on generated drafts and machine-created docket deadlines. The absence carries more weight here than on a pure software record, because Tradespace markets a full-service patent practice delivered through a vetted network of outside firms and does not publish how professional obligations are allocated between itself, the network firm and the customer.
Billing and Fee Posture
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.
Cost claims are central to the pitch and none of them is a disclosure record. Tradespace sells explicitly against the billable hour, describing itself as an alternative to the law firm markup, telling buyers to stop paying hundreds an hour for outside counsel to summarise office correspondence, offering flat-fee prosecution so that cost is known before work begins, and publishing an ROI calculator. The Manage module tracks outside counsel spend as a reporting feature. Nothing addresses how AI-assisted work is recorded or disclosed, and no per matter record of AI-assisted work was located. One structural note belongs on the record. This signal assumes a vendor selling to a firm that bills a client, and Tradespace mostly inverts that, selling to the in-house or university team that pays the bill, so its flat-fee model addresses the cost question at the commercial level rather than through disclosure. That inversion does not hold completely, because law firms are a named buyer segment, and for those customers the question the signal asks is live and unanswered.
Outside Counsel Guideline Readiness
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.
The model route is named and nothing else is. Asked which models it uses, Tradespace states a model-agnostic approach and that by default it works with Microsoft Azure to access their OpenAI Services, and it links to Microsoft's published data privacy terms for that service rather than merely naming the vendor. A firm can therefore tell a client which provider processes its content and point at the applicable data protection terms, which is why the bottom value is not made out. It stops well short of the top value. No subprocessor list exists in any form, no other processor is named anywhere, and the model-agnostic claim means the named default may not describe what a given customer is actually running. There is no data processing addendum, no privacy policy and no terms of service, so no forwardable client-facing artifact exists at all; the SOC 2 Type 2 report behind the Vanta trust centre is the only obtainable document and its access tier is not stated.
Court Disclosure Support
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.
Checked the home page, platform overview, Create, Manage, Enterprise AI and Security and Privacy pages on 2 September 2026. Nothing addresses disclosure of AI use to a patent office or tribunal, and no exportable per document record of which model produced which passage, what was retrieved and who reviewed it is described. The platform holds much of the raw material: citations are stated to accompany generated responses, the Paragon drafting technology is described as offering traceability to source materials, the docketing assistant captures all office and counsel correspondence for a case, and audit logs are named among the security controls. None is presented as a record of model use. The gap is worth naming because the duty of candour in patent prosecution attaches to what is submitted, and a platform that drafts applications, suggests office action responses and creates docket deadlines is generating exactly the material a later inequitable conduct challenge would probe.