IP Author
AI assisted patent platform for practitioners covering drafting, prior art search, office action response preparation, claim drafting, invention management, patent classification and an Evidence of Use assistant, sold in a standard and an IP Author Enterprise configuration. The drafting module generates a complete first draft from an invention disclosure in a stated fifteen minutes, producing claims, specification and figures together and keeping them consistent, including automatically generated and editable flowcharts and block diagrams, with figure references that update across claims, specification and legends when a figure is renamed. Disclosures are accepted as Word documents, PDFs or slide decks and the platform is stated to read the images, tables and diagrams inside them rather than the text alone, generating flowcharts or block diagrams where they are missing. For life sciences and chemistry matters, protein and peptide sequences can be pulled into claims, named and automatically numbered during drafting. Prior art search is integrated into the drafting flow rather than sold separately, running semantic search from the disclosure across worldwide patent literature, non patent literature and technical standards in near real time across more than 100 jurisdictions, with filtering by jurisdiction, CPC and IPC classification, assignee and publication details, and returning similarity scores, AI generated novelty summaries and explanations of overlaps and differences against the invention. Results support highlighted passage review and side by side disclosure comparison, and export as client ready Word reports or Excel sheets including masked versions that withhold filing dates and raw excerpts so findings can be shared with clients or outside counsel without oversharing. Office action support spans 102 and 103 prior art rejections, 101 subject matter eligibility, 112 written description, enablement and definiteness, and double patenting, always producing a traversal strategy, with structured arguments, claim mapping and references generated automatically from the full application history. Workflows are stated to align with jurisdiction specific practice for the USPTO, EPO and other offices. The vendor states that users retain full review, editing and decision making authority at every stage and that attorneys and agents remain in full control of the final work product. The platform is stated to connect its tools, including patent search, claim charting, whitespace analysis and office action response, to agentic frameworks. A 14 day free trial is available directly from the website, and the site publishes a Trust Center, a privacy policy and a pricing page.
Capability grades
All 15 axes, graded from public sources on the date shown. Hover a grade to see what the letter means on that axis.
AI Centrality
How much of the product is actually AI. Whether the machine learning is the mechanism the buyer is paying for or a feature layered onto conventional software, and whether the vendor is specific about which is which.
Every module is generative and there is no underlying platform that would function without them. Drafting produces claims, specification and figures from a disclosure, prior art search runs semantically from the disclosure text rather than from Boolean queries, office action responses are generated with structured arguments and claim mapping from the application history, classification is model driven, and the Evidence of Use assistant is an analysis product. Multimodal ingestion is a real model capability rather than a feature claim: the platform is stated to read images, tables and diagrams inside Word documents, PDFs and slide decks, and to generate flowcharts and block diagrams where a disclosure lacks them. Remove the models and there is no product at all, not even a database, since the corpus is searched rather than owned. Fourth A on this axis in ip-and-patents out of five records, with PatSnap the only exception and that on age.
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.
Grounding is shown to the user at several points and one accuracy claim is published with a comparator but no method. The review surface is the substantive part: similarity scores against the invention, AI generated novelty summaries, explicit explanations of overlaps and differences between the invention and each cited reference, highlighted passages, and side by side disclosure comparison. That lets a practitioner test a retrieval result rather than accept it, and it is more than most records in this category describe. The published accuracy claim is that compared to a human searcher the tool returns similar or the same results 19 out of 20 times. Naming the comparator as a human searcher is meaningful and is credited, and everything that would make it checkable is missing: no sample size, no evaluator, no definition of similar, no date and no methodology. Held at B on that basis. Recorded for the next reader: the vendor also publishes its own comparison content benchmarking itself against named competitors, which is self published competitive material and is not credited as evidence.
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.
An explicit oversight commitment, stated twice in separate surfaces and in plain terms. The product material states that users retain full review, editing and decision making authority at every stage of drafting, and the published FAQ states that the platform keeps attorneys and agents in full control of the final work product. Two independent statements of the same commitment, one of them in a FAQ answering a direct question, is a clearer position than the implied oversight most records in this pull offer. Output is framed as a first draft and first pass response throughout rather than as finished work. Held at B rather than A because nothing is bounded operationally: no statement of what the system does unattended, no confidence threshold, no escalation behaviour, and no description of what happens in the agentic configuration where the platform connects its tools to external orchestration frameworks, which is precisely the setting where a stated commitment to full user control is hardest to maintain and is not addressed.
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.
Nothing located. No customer or firm is named anywhere in vendor material, no case study, no usage figure, no funding announcement, no independent analyst placement, no award and no third party review were found. The vendor publishes a substantial content library including competitive comparisons and benchmark posts, and all of it is self authored, so the evidence base for this record is entirely the vendor describing its own product. That is the weakest operational position in ip-and-patents by a distance: Patlytics names Am Law 100 penetration and a Quinn Emanuel partner, Solve Intelligence names DLA Piper, Siemens and Finnegan, DeepIP publishes dated funding and traction figures, and PatSnap names NASA, Tesla and MIT in independent trade coverage. Checked the home page, the product pages, the FAQ, the blog library and the site navigation on 29 Aug 2026. Recorded as a documented absence across those surfaces rather than as a judgement about adoption, which cannot be assessed from public material.
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 genuinely thoughtful confidentiality feature, and no professional confidentiality position. The feature is worth naming because no other record in this pull has anything like it: prior art results export in masked versions that withhold filing dates and raw excerpts, stated expressly so findings can be shared with clients or outside counsel without oversharing. That is a product designed around the reality that a prior art report circulated too widely is itself a disclosure risk, and it addresses onward sharing rather than only storage. Against that, nothing located addresses attorney client privilege, work product, or the specific consequence this category carries, being that exposure of an unpublished application can affect patent rights rather than merely creating a data incident. Both Patlytics and Solve Intelligence engage that directly and both reached A. Research limitation stated: a Trust Center and a privacy policy are both published on the site and neither was read in this pass, so this grade is a correction candidate.
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.
Not located. The product drafts complete patent applications and generates office action responses spanning statutory grounds under sections 101, 102, 103 and 112, all of which are filed under a registered practitioner's signature, and it offers a self serve 14 day trial which means the buyer is not necessarily an attorney. The vendor states that attorneys and agents retain full control of the final work product, which is a product positioning statement graded on the Autonomy axis and is not a professional responsibility framework. No reference to USPTO Rules of Professional Conduct, 37 CFR, duty of competence or any bar or patent office guidance was located. Compare Patlytics at B in this category, which names those authorities directly. Checked the home page, the product pages, the FAQ, the blog library and the site navigation on 29 Aug 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.
Nothing published about how the models are governed, evaluated or monitored. No AI policy, no model card, no bias or fairness testing, no evaluation methodology, no accuracy monitoring beyond the single unmethodised search claim, no drift statement, no named governance body, no ISO 42001 and no EU AI Act positioning were located. The vendor publishes commentary on AI industry developments and on intellectual property questions raised by AI systems, which engages the subject matter externally rather than disclosing anything about its own models. This is the weakest governance position in the category alongside DeepIP, against Patlytics holding an ISO 42001 certificate and Solve Intelligence publishing a continuous evaluation process. Checked the home page, the product pages, the FAQ, the blog library and the site navigation on 29 Aug 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.
No stewardship position located in any surface read. Nothing states whether invention disclosures, draft applications, search histories or generated work product are used to train or improve models, no retention period is published, and no deletion right is described. The content at stake is the most sensitive in this category, being unpublished invention disclosures uploaded before filing, and the vendor's own masked export feature demonstrates awareness that disclosure control matters, which makes the silence on its own handling more noticeable rather than less. Research limitation recorded prominently rather than buried: a Trust Center and a privacy policy are both published on this site and neither was read in this pass, and either is the surface where a training or retention statement would ordinarily sit. This grade is flagged as the strongest correction candidate on the record and should not be read as an established absence of policy, only as an absence across the pages actually checked, being the home page, the product pages, the FAQ and the blog library on 29 Aug 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.
No published position located. Nothing was found on liability for AI output, warranty, service levels or remedy. The exposure is the category's own and is concentrated by the speed claim: a complete first draft in fifteen minutes compresses the interval in which errors would ordinarily surface, and claim scope lost at grant is not recoverable. The office action module covering statutory rejection grounds carries the further risk that a mischaracterised traversal argument creates prosecution history estoppel. The vendor's answer to all of this is the stated retention of full attorney control, which allocates responsibility to the user by design rather than by contract, and no contractual position is public. Checked the home page, the product pages, the FAQ and the site navigation on 29 Aug 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.
Internal integration is real and external integration is thin. Within the platform, prior art search is deliberately integrated into the drafting flow rather than sold as a separate product, so references surfaced during search carry forward into claims and specification without re-keying, and results export as Word reports or Excel sheets. Externally, the vendor states that its tools including patent search, claim charting, whitespace analysis and office action response connect to agentic orchestration frameworks, which is a forward looking integration surface and is the only one described. What is absent is everything a firm would ask for: no IP docketing system, no patent management platform, no document management system, no Microsoft Word integration despite drafting being the core product, and no API documentation was located. For a prosecution tool, the docketing gap is the material one, and DeepIP names IP management platform integration in this same category.
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.
Nothing located. No hosting provider is named, no region or data residency commitment is published, and no single tenant, dedicated or on premise option is described. The gap is material in this category because unpublished applications engage foreign filing licence requirements and export controls in several jurisdictions, and a practitioner would need to know where a disclosure is processed before uploading one, particularly given the platform supports prior art search across more than 100 jurisdictions and therefore serves a geographically distributed user base. Compare DeepIP at A for on premise deployment and Solve Intelligence at A for customer selectable jurisdiction, both in this category. Research limitation: the Trust Center was not read in this pass and is where a residency statement would sit. Checked the home page, the product pages, the FAQ and the site navigation on 29 Aug 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.
A Trust Center is published and its contents were not read, which is the whole of what can honestly be said. The site navigation carries a dedicated Trust Center alongside a privacy policy, so the vendor has built a trust surface rather than treating security as a marketing line, and under the three tier test a published trust surface is not an absence. No certification of any kind was located in the pages actually read: no SOC 2 of either type, no ISO 27001, no ISO 42001, no named auditor, no penetration testing partner and no examination date. Held at C rather than D on the same reasoning applied to TrialView earlier in this pull: grading D would assert an absence on the single surface most likely to contradict it, which is the error the trust portal sweep corrected twenty five times, and repeating it after writing the rule would be worse than the original. Flagged as a correction candidate in both directions, since reading the Trust Center could establish certifications that raise this grade or confirm their absence and lower it. Every category peer publishes at least one named certification.
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.
A provider is named, in passing, in a two year old marketing comparison, which is a genuine disclosure of a kind and not a maintained one. In a blog post comparing itself to general purpose tools the vendor states plainly that IP Author also used ChatGPT and that the difference lies in prompt engineering, the patent search engine and user experience. Naming the underlying provider outright is rare on this index and it is credited. What stops this reaching the grade Onspring received for the same act is that the disclosure is incidental rather than deliberate, appears in dated April 2024 competitive content rather than in current product or trust material, is not repeated anywhere current, and gives no model, version or scope, so a buyer cannot tell whether it still holds after two years of the platform being rebuilt around newer capabilities. Separately, the vendor names agentic frameworks its tools connect to, which discloses an orchestration relationship rather than the model layer. No subprocessor list was located.
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.
Two self serve elements exist and no figure was read. A dedicated pricing page is published in the site navigation, and a 14 day free trial is available directly from the website with the FAQ confirming it can be started without going through sales. A published pricing page plus a self serve trial is a materially more open commercial posture than the demo request pattern that governs every other record in this category, where Patlytics, Solve Intelligence and DeepIP all publish nothing and PatSnap publishes only a tier structure through an analyst listing. Held at C rather than higher because the pricing page contents were not read in this pass and no price, range, tier structure or unit of charge was captured, so the openness is established and the figures are not. Flagged as a correction candidate in the upward direction, and a record where the grade may understate the vendor.
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.
Coverage is enumerated on the dimension that matters most for prosecution, which is the statutory grounds the tool can actually handle. Office action support is stated across 102 and 103 prior art rejections, 101 subject matter eligibility, 112 written description, enablement and definiteness, and double patenting, with a traversal strategy always produced. No other record in this category enumerates rejection grounds, and it is the enumeration a prosecution practitioner would check first, because a tool that cannot handle a 101 rejection is unusable in software and biotech practice. Search coverage is stated across worldwide patent and non patent literature and technical standards spanning more than 100 jurisdictions, filterable by jurisdiction, CPC and IPC classification and assignee. Drafting is stated to align with USPTO, EPO and other office requirements, and technical domains include life sciences and chemistry with named sequence handling. Held at B rather than A because the drafting jurisdiction list stops at two named offices and others, no corpus date range or update lag is published for the search side, and the technical domain list was not fully enumerated in the material read.
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?
No located term or policy addresses the question either way.
Silent. The quoted phrase is the vendor's clearest statement about the relationship between the user and the system, and it concerns control over the work product rather than use of the data. No statement in either direction was located on whether invention disclosures, draft applications, search histories or generated work product are used to train or improve models. The content at stake is unpublished invention disclosures uploaded before filing, which is the most disclosure sensitive material in this category, and the vendor's own masked export feature shows it understands that disclosure control matters. Recorded as silent, not as a negative commitment. Research limitation stated prominently: a Trust Center and a privacy policy are both published on this site and neither was read in this pass, and either is where such a statement would ordinarily sit, so this value is the strongest correction candidate on the record. Checked the home page, the product pages, the FAQ and the blog library on 29 Aug 2026.
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.
Not addressed in any surface read. No retention period is published for uploaded disclosures, generated drafts, search results or office action responses, and no deletion right is described. Invention Management is offered as a module, which implies durable storage of disclosures as a product function, and nothing states for how long or under what terms. Same research limitation as the training signal applies: the published Trust Center and privacy policy were not read in this pass and either is where a retention statement would sit. Checked the home page, the product pages, the FAQ and the blog library on 29 Aug 2026.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
No located public material addresses walls or matter level segregation.
Not addressed. No permission model, matter level access restriction or tenant segregation description was located. The conflicts scenario in patent practice is concrete, since a firm may prosecute in the same technical field for competing clients, and nothing published describes what separates one matter's disclosure from another inside the platform or what an Invention Management module exposes across a firm. One adjacent feature exists and is credited on the Privilege axis rather than here: masked export withholding filing dates and raw excerpts, which controls what leaves the platform rather than who can reach what inside it. Checked the home page, the product pages, the FAQ and the site navigation on 29 Aug 2026.
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.
Not addressed. No government or law enforcement request clause, no commitment to notify a customer before producing their data, and no transparency report were located. The category stake applies: the platform holds unpublished invention disclosures and pre filing drafts, and disclosure of an unpublished application to a third party can bear on novelty and prior disclosure questions. Same limitation as above, with the privacy policy unread. Checked the home page, the product pages, the FAQ and the site navigation on 29 Aug 2026.
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.
Named by type and jurisdiction count, with no source, licence basis or currency stated. The search corpus is described as worldwide patent literature, non patent literature and technical standards across more than 100 jurisdictions, filterable by jurisdiction, CPC and IPC classification, assignee and publication details. Naming non patent literature and technical standards separately is worth crediting, because prior art is not confined to patents and standards documents in particular are both critical to prior art analysis and encumbered by publisher licensing. That last point is exactly what is unaddressed: no database, publisher or standards body is named, no licensing basis is stated for the non patent literature or standards, no date range is given, and no update lag or refresh frequency is published despite results being described as returning in near real time. A practitioner relying on a novelty summary cannot determine what was searched or how current it was.
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.
Not addressed, and applicable in both forms this category presents. The office action module generates traversal arguments against statutory rejections under sections 101, 102, 103 and 112, which in practice involves citing and distinguishing case law, and nothing states whether cited authority is checked for current treatment before it enters a filing. The patent specific analogue applies to the search side, where the question is legal status: whether a surfaced reference is granted and in force, lapsed, expired, amended in reexamination or invalidated in post grant proceedings. Filtering is offered by publication details, which is bibliographic rather than status aware. Fifth consecutive record in this category to leave this signal unanswered, and the pattern is now unbroken across every ip-and-patents vendor built.
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.
Not addressed as a behaviour, with a partial answer on the search side that stops short. Similarity scores are surfaced against each cited reference alongside novelty summaries explaining overlaps and differences, which gives a user a graded signal about retrieval relevance and is genuinely useful. It is a relevance score for a returned reference, not a confidence signal about the system's own output: nothing states whether a low confidence generated claim is flagged, whether the drafting module will decline where a disclosure is too thin to support a claim, or what happens when the office action module cannot construct a traversal, despite the vendor stating a traversal strategy is always produced, which is a commitment to always answer and the opposite of an abstention path. Checked the home page, the product pages, the FAQ and the blog library on 29 Aug 2026.
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.
None located, with the instrument named. General web searches combining the vendor and product names with court, order, sanction, fabricated citation and patent terms returned nothing on 29 Aug 2026, and no named docket database, patent office record system or court record tracker was searched. Recorded as a statement about what this search found, not as a clearance. The exposure shape is a mischaracterised prior art reference or a defective case citation inside an office action response filed at a patent office, which would surface in a prosecution file wrapper rather than in a published opinion. Worth noting that no customer of this vendor is publicly identifiable, so even a proper docket instrument would have little to anchor a targeted search against.
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.
Not addressed. No named ethics opinion, no USPTO Rules of Professional Conduct reference, no 37 CFR citation, no duty of competence discussion and no bar guidance was located. The vendor publishes an extensive practitioner facing content library covering drafting workflows, tool comparisons and AI industry commentary, and that material addresses capability and efficiency without reaching the professional rules governing the practitioner who signs the filing. Fourth of five records in this category at this value, with Patlytics the only exception. Checked the blog library, the home page, the product pages, the FAQ and the site navigation on 29 Aug 2026.
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.
Savings claims only, denominated in time to first draft rather than as a percentage. The headline figure is a complete patent draft in fifteen minutes, against a manual baseline the vendor describes as weeks, and prior art results returning in near real time against days of manual database work. Those are concrete and testable framings and they carry no methodology, sample or definition of what a complete draft comprises. Nothing appears on the client's side of the equation: no position on how AI assisted drafting time should be recorded on an invoice, and no exportable record showing what portion of an application or office action response was machine generated, which matters for a product sold to firms billing prosecution hourly. Checked the home page, the product pages, the blog library and the FAQ on 29 Aug 2026.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
The material exists behind a sales conversation or an executed agreement.
On request, on the strength of a published Trust Center whose contents were not verified. A dedicated Trust Center and a privacy policy are both published in the site navigation, which gives a firm a defined destination to point a client to, and that is the distinction this value turns on rather than the strength of what sits behind it. Nothing was located outside it: no named certification, no subprocessor list, no auditor, no data processing agreement and no downloadable summary. Held at on request rather than not addressed because a published trust surface exists and was not entered, and held no higher because nothing behind it is established. Note the contrast within this category: PatSnap and Solve Intelligence reach this value through trust centres whose named contents were verified, and this record reaches it on the existence of the surface alone. Correction candidate in both directions.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Some elements of the record are available, short of a document level export.
Partial record, with an unusually strong internal consistency mechanism and the familiar two gaps. The source limb is answered on the search side: similarity scores, novelty summaries, highlighted passages and side by side comparison mean a practitioner can show what a novelty position rests on and where in a reference it appears, and exports carry to Word or Excel. The drafting side adds something no other record in this category describes, being automatic propagation of figure references so that renaming a figure updates every reference across claims, specification and legends, which addresses a real and common source of procedural rejection and is a traceability feature in substance. The gaps are the standard two: nothing indicates that output records which model produced it, and no human verification record is captured, so a practitioner who reviewed and adopted a generated claim set or traversal argument cannot evidence that they did. The relevant forum here is a patent office file wrapper rather than a court.