Home › Platform
Platform and competitive landscapeEvery patent tool ever built assumes someone has already decided to file. A firm needs an instruction. A drafting copilot needs a disclosure. Tech transfer software needs a form a faculty member filled in. Patly begins one step earlier, at the research a university has already published, which is where almost all of the value is stranded and where nothing else operates.
American universities spend more on research every year than the market value of most public companies, and the protected output of all that spending passes through a single bottleneck: a form an individual faculty member has to choose to fill in. Budget decides how many filings an office can afford. Attention decides which papers it ever sees. Patly fixes the second and cuts the cost of the first.
Nearly every company in patents sells into one of two segments, and both are efficiency plays: the work already exists, the budget already exists, and the only question is who does it more cheaply. The third segment has no incumbent because the work never starts.
Drafting copilots, search tools and docketing sold by the seat to people who already draft patents for a living.
Filing services and flat-fee providers competing to prepare applications a client has already decided to pursue.
Protection for work that is already public, already valuable, and currently on nobody's list because no one filled in a form about it.
Every filing won there was going to happen anyway. The revenue is a transfer from an incumbent, and the total is capped by decisions made before any vendor is involved. Practitioner tools are the same engine, sold to a market that already buys. The university market is the one nobody serves.
A filing Patly surfaces from a published paper is a filing that did not exist in anyone's forecast. It is not taken from a law firm. It is created out of research that was heading into the public domain unprotected.
In the other two segments the vendor captures a share of a drafting fee. Here the drafting fee is the entry price to an asset with licensing life measured in decades. The fee and the value it unlocks are not the same order of magnitude.
Budget limits how many filings an office can afford, but it is only ever spent on what reaches the office, and a faculty member has to volunteer before anything begins. Read everything and the budget goes to the best of it; cut the cost of drafting and it goes further.
Read left to right, this is the path a paper takes from publication to a signed, filed application and a named licensee, with the university deciding at every gate.
Every paper the university publishes is read and assessed for patentable subject matter, with no form to submit and no meeting to schedule.
Every candidate is ranked against the rest and against the grace window, so the office sees its whole portfolio rather than whatever arrived this month.
Each candidate carries a written assessment, so a decision to file or pass is made on a document rather than an impression.
The work starts from what has already been written. Researchers keep publishing and are asked for nothing until their paper is already ranked.
Applications are drafted to the conventions of the technology area, the inventor's own body of work, and the office's house preferences.
Claim by claim and limitation by limitation, including the dependent scope ladder, rather than a single generated document to take or leave.
The draft is produced through many directed steps that search, compare, revise and re-check, instead of one pass over a disclosure.
The draft is examined against 101, 102, 103 and 112 the way an examiner would, and the rejections are answered before filing rather than twenty months later.
The simulation runs adversarially and repeatedly, hunting for the strongest rejection available rather than confirming the draft is fine.
Likely office actions are forecast and responses drafted against them, so prosecution strategy exists on day one instead of being improvised on arrival.
Each application is routed to a patent attorney whose registered practice matches the technology, rather than to whoever the office happens to retain.
Nothing reaches the USPTO without review and signature by an independent licensed patent attorney. Patly is not a law firm and does not provide legal advice.
Companies whose products read on the claims are identified and ranked, so a granted patent arrives with a list of who would pay for it.
Compared by category rather than by vendor, because the distinction that matters is structural: what part of the path each kind of provider was built to serve. Named examples are given for orientation, and the marks describe the category.
| Capability | Patlyuniversity IP, end to end | Patent law firmstraditional prosecution | AI drafting copilotsSolve, DeepIP, Patlytics, Rowan, PatentPal | IP search and analyticsPatSnap, Clarivate, LexisNexis, Questel, IPRally | Tech transfer softwareWellspring, Inteum, IN-PART | Flat-fee filing servicesvolume filing providers |
|---|---|---|---|---|---|---|
| Before a disclosure exists | ||||||
| Reads every paper the university publishesContinuous, no submission required | ● | – | – | – | – | – |
| Ranks unprotected work for filingOrdering candidates that are not yet applications | ● | – | – | ◖ | – | – |
| Written judge report per candidateA document to decide on, before spending | ● | ◖ | – | ◖ | – | – |
| Tracks the grace window on published workThe clock that starts at publication | ● | ◖ | – | – | ◖ | – |
| Works with near-zero faculty involvementNo disclosure form, no interview to schedule | ● | – | – | – | – | – |
| Drafting | ||||||
| Drafts a complete applicationSpecification, claims, drawings, abstract | ● | ● | ● | – | – | ● |
| Personalized to inventor, field and officeRather than one house template | ● | ◖ | ◖ | – | – | – |
| Fine-grained control over claims and scope ladderLimitation by limitation | ● | ● | ◖ | – | – | – |
| Agentic draftingMany directed passes, not one generation | ● | – | ◖ | – | – | – |
| Examination, before the examiner | ||||||
| Deep examination simulation101, 102, 103 and 112 as an examiner would apply them | ● | ◖ | ◖ | – | – | – |
| Agentic, adversarial examinationHunting the strongest rejection, repeatedly | ● | – | – | – | – | – |
| Prosecution forecast and drafted responsesBefore the first office action arrives | ● | ◖ | ◖ | – | – | – |
| Filing, and what comes after | ||||||
| Automatic matching to expert patent counselBy registered practice area, per application | ● | – | – | – | – | ◖ |
| Licensed counsel reviews and signsIndependent attorney of record | ● | ● | ● | – | – | ● |
| Files with the USPTOEnd to end, not handed back as a draft | ● | ● | – | – | – | ● |
| Automatic licensee discoveryWho would pay for the granted claims | ● | – | – | ◖ | ◖ | – |
| Flat, published price per filingNo hourly exposure | ● | – | – | – | – | ● |
| Buyer is the technology transfer officeNot the attorney, not the inventor | ● | ◖ | – | ◖ | ● | – |
A drafting fee is what a customer pays. It is not what the transaction is worth. Each layer below sits on top of the one before it, and only the first is priced today.
A flat fee to draft and file a non‑provisional, against the $10,000 to $15,000 a firm bills for the same work.
Drafting and filing at outside counsel rates, for the applications an office already planned to file.
Applications that would never have been filed, because the paper was never read and no disclosure was ever submitted.
The licensing life of an asset that was otherwise heading into the public domain, plus the named licensees who would pay to use it.
Drafting is the visible step and the crowded one. Deciding what is worth drafting, out of everything a university publishes, is the step that requires reading all of it and ranking it, and no existing product has a reason to do that work.
A drafting copilot sold to law firms cannot go upstream without competing with the firms that buy it. Tech transfer software cannot go downstream without becoming a filing provider. The gap stays open because closing it breaks somebody's business model.
Examination outcomes feed back into the scoring and into the simulation. The ranking gets better at predicting what survives prosecution the more of it there is, which a competitor starting later cannot shortcut.
Seven US utility applications, filed and pending, cover the path from published research to an attorney-signed filing: detection and scoring, ranking against the grace window, verified drafting and simulated examination, and automated submission.
PATLY.AI