IP Author wants to give patent lawyers their two days back.

Here is a fact about patent law that is easy to miss if you have never billed for it: a great deal of what a patent attorney does is not lawyering. It is typing. It is turning an engineer's half-formed disclosure into claims, then a specification, then an abstract, then diagrams - each in a rigid, self-referential dialect that the patent office insists upon and that no human enjoys producing. The judgment part is scarce and valuable. The typing part is neither. IP Author, a company founded in 2023 in Fremont, California, is a bet that generative AI can absorb the typing and leave the judgment alone.

The pitch is disarmingly concrete. A drafting task that used to take two or three days, the company says, can now be done in hours. A prior art search that would eat four to six hours of an associate's afternoon returns comparable results in about two minutes. An office action - the letter in which a patent examiner explains, at length, why your claims are no good - gets summarized automatically, with counterarguments drafted claim by claim. None of this is presented as magic. It is presented as time, given back.

The company was founded by Samir Raiyani, who runs it as CEO and has spent a career at the seam between technology and business. His framing of the product is worth pausing on, because it explains a design choice. Patent professionals, he argues, do not want ten tools. They want one. Invention disclosure, drafting, prior art search, office action responses - each of these had, until recently, its own vendor, its own login, its own way of not talking to the others. IP Author's proposition is that they belong in a single workspace, and that the workspace should be generative AI from the ground up rather than a legacy database with a chatbot bolted on.

That sounds obvious the way good ideas sound obvious in retrospect. It is not free. Unifying a workflow means owning every step of it, which is harder than owning one step well. But it is also, plausibly, a moat. A tool that only drafts claims competes on claim-drafting. A tool that carries an invention from an inventor's rough notes all the way to a filed response is competing on something stickier: the shape of a patent department's day.

The customers suggest the pitch is landing where it counts. IP Author names Qualcomm, ResMed, Chevron, and Elmos Semiconductor among the companies using it - which is to say, sophisticated in-house patent groups that are structurally suspicious of anything new. These are not organizations that adopt AI because it is fashionable. They adopt it when someone has done the unglamorous work of making it trustworthy.

The hard part of selling AI into a corporate patent department is not the AI. It is the department. These are teams whose entire job is finding the flaw, the prior art, the unsupported claim - trained skeptics, professionally allergic to hand-waving. So IP Author leans on the two things that reassure them: proof and process. SOC 2 Type II compliance and a zero-data-retention option address the fear that your inventions leak. A customer advisory board of senior IP leaders - from ResMed, Innate Pharma, Elmos, and top firms - keeps the product honest to how the work actually gets done.

In November 2025, IP Author raised a seed round led by eLab Ventures and Illuminate Ventures, the latter bringing experience scaling enterprise software. The plan is unsurprising and sensible: expand research and engineering, deepen the draft-optimization engine and prior-art integration, and push further into the U.S. and Europe. Reported total funding sits around $3.12 million - modest next to rivals like DeepIP, Patlytics, and Solve Intelligence, which have raised in the tens of millions.

Being the smaller, later-funded player in a hot category is usually described as a disadvantage. IP Author's implicit answer is that closeness to the customer beats a bigger war chest - that a product built shoulder-to-shoulder with practicing attorneys wins the accounts that matter. Whether that holds as the field consolidates is the open question. For now, the customer list suggests it is at least a defensible bet.