The patent system is dysfunctional in large part because it consists of too much text for humans to reliably comprehend. Recent developments in artificial intelligence – specifically the transformer model – create the possibility for a patent system that is more legible and highly automated. To this end, we have spent the past two years building technology to search for and reason about patents across all categories of patents. Our first implementation of this technology is a high end prior art search service that is more accurate and complete than anything else that exists on the market currently. We offer this service at an irrefusable price point so that we may choose our customers tactically. This service is already gaining market share inside several of the largest patent litigation firms in the world. We have chosen this beachhead market because the reputation conferred by serving these firms should facilitate (a) the procurement of substantial government contracts that automate the patent examination process, and (b) adoption of our technology inside companies that deal in invention. For these high tech companies, our core insight is that the same technology we have built to search and reason in the context of prior art search can be repurposed to augment and partially automate all of the other functions of an in-house IP department – i.e. harvesting, prosecution, freedom to operate, monitoring, litigation, licensing, etc. While these functions are currently fragmented between a multitude of outdated software products and commoditized services, we believe in consolidation and verticalization: a singular software product carefully designed to delight composed of a series of synergistic products and services constituting a modern operating system for invention.