Jason Boehmig and the trouble with paperwork

A corporate lawyer can spend a surprising amount of time behaving like a computer. Find the right document. Check the figures. Compare the versions. Give the resulting file a sensible name. Jason Boehmig became interested in what would happen if an actual computer took on more of this work. It was an inquiry with consequences for his employment.

At Fenwick & West, where he worked with technology companies on corporate transactions, the work put him close to people building new businesses. He was helping them assemble the legal foundations. His own tools left him wanting something better. A profession capable of arranging a venture financing still required an associate to shepherd documents from one place to another.

In one example he later described, a simple script compared documents in two folders and produced PDF redlines. The ambition was modest enough to fit inside a working day. A piece of repetitive work would become less repetitive. The lawyer would have more time for the parts of the job that required a lawyer.

A lawyer goes shopping for better tools

Boehmig did not begin with a lifelong plan to run a software company. His earlier career had already involved a substantial change of direction. A Pittsburgh native, he spent four years at Lehman Brothers before going to law school. He graduated from Notre Dame in 2012 and entered corporate practice. Finance, law, software: the sequence makes more sense when you follow the work through it.

The quantitative traders he had encountered provided a model. They used programming to improve their trading, combining knowledge of a field with tools that helped them operate in it. He imagined a similar arrangement for his legal practice. He wanted to use technology to become a more capable attorney. First he looked for something he could buy.

The search did not produce what he wanted. Learning to make small applications became a practical response. He hired a coding tutor and worked at it over time. There is something pleasingly unceremonious about that detail. An industry’s future can begin with someone paying for a lesson because a task on their desk has become intolerably fiddly.

“My intent was never to build the technology.”

Jason Boehmig, recalling his search for legal tools

The irritation had a deeper source than awkward file handling. In corporate legal work, important judgments depended on information buried in agreements and company records. Getting to that information could consume time that an attorney might otherwise spend advising a client. Boehmig began to see an opportunity in organizing the information itself. By 2014, he had left Fenwick to work on the idea that became Ironclad.

An assistant with a human inside

His partnership with engineer Cai GoGwilt had an appropriately accidental beginning. Boehmig recalled attending a small lecture at Stanford where people introduced what they were working on. He was a lawyer learning to program for legal automation. Across the room was a software engineer learning about law for much the same purpose. The introductions did a useful portion of the matchmaking.

An early Ironclad service worked through email. A customer copied the assistant into a transaction; information was collected and the agreement moved toward signature. Behind some of that apparently automated correspondence was Boehmig himself. The founder was doing the work the product was supposed to make easier. It was a particularly direct form of customer research, with an inbox attached.

By the summer of 2015, Ironclad was in Y Combinator. Its early proposition addressed the disconnected steps of business paperwork: preparing a document, getting it to the right people, arranging signatures, and storing the result. Each step might have its own tool. The effort of moving between them belonged to the customer.

That made the company’s initial pitch easy to understand. Standard documents and administrative handling absorbed the attention of expensive professionals. Ironclad would coordinate the process so those professionals could devote more of their day to advice. The first opportunity was close to the ground: agreements businesses already needed, handled with less interruption.

The information inside the agreement

Contracts have a curious second life. After the negotiation, the corrections, and the signatures, they contain a business’s commitments. Somewhere inside are the answers to questions people will ask later. What did we agree to? When does it end? Which terms survived negotiation? A signed document can be settled while the business questions around it remain very much alive.

Boehmig’s public description of his work focuses on turning static documents into workflows and usable data. The appeal is easy to see. If information is captured as an agreement is made, people should be able to use it again. The legal document becomes part of how the business understands itself. Filing and remembering cease to be entirely separate chores.

The founders had imagined an AI legal assistant early on. The available technology required patience. They stepped back from presenting Ironclad as an AI platform and built the underlying system while waiting for the capabilities to improve. For a business selling dependable processes, a compelling demonstration would only get them so far. The work still had to hold together for a customer.

The eventual arrival of stronger models changed what could sit on top of that groundwork. Ironclad’s product history traces the progression: Smart Import for bringing contracts into a searchable system, AI Assist for negotiation work in 2023, and Jurist for tasks including drafting, editing, and explaining agreements in 2024. The product surface expanded as more of the original idea became technically practical.

There is a lesson in the order of events. A founder may recognize a useful destination before the tools can reach it. That leaves years of decisions about what to build now, what to defer, and what customers can reliably use. Boehmig and his colleagues spent those years working on the process surrounding the document. When AI improved, there was an existing business for it to help.

A company big enough to hand over

The financing announcements gradually gave this argument with paperwork a different scale. In January 2022, Ironclad raised $150 million at a reported valuation of $3.2 billion. Franklin Templeton led the round, with existing investors participating. Boehmig, who had once helped startups prepare their financing documents, was now signing them for a company of his own.

In April 2025, the company announced that Boehmig would become executive chairman and Dan Springer would take over as chief executive. Springer brought experience running Docusign and Responsys. Boehmig had led the search for his successor. The transition arrived as Ironclad passed $150 million in annual recurring revenue and reached 2,000 customers.

A founder’s biography often turns a company into an extension of one person. This moment complicates that picture usefully. The business had grown large enough to need its own next chapter. Choosing someone else to lead it became part of Boehmig’s work. The role he had created could be handed to another executive.

His current personal profile describes him as an Ironclad founder and board member. That is a different relationship from running the daily operation. It also makes room for another version of the question that started his company: what tools should people working in law have available to them now?

Back to the classroom, forward to the builders

Boehmig’s connection to legal education had continued alongside the business. At Notre Dame, he co-taught Law 2.0 with Ron Dolin, introducing students to the technologies and organizational changes affecting their profession. The course connected Silicon Valley and the campus. Students preparing for practice were being asked to think about the tools they might encounter and the systems they might help create.

That work gives his career another point of continuity. Building software changes the tasks a profession asks its newest members to do. Teaching offers a place to consider how they will learn instead. In later conversations, Boehmig has discussed legal engineers and other roles that combine technical skills with an understanding of legal work. His own route gives those combinations a practical precedent.

In June 2026, he joined OpenAI to lead its legal products. By September, he was participating in the launch of Astra for Law. The New York event included a visit from a Harvard Law School seminar, with David Wilkins giving a keynote on the profession’s future. The gathering brought the classroom and the product launch unusually close together.

“I’m kind of a builder at heart.”

Jason Boehmig, AI Sidebar, October 2026

In October, discussing the new role, Boehmig described an industry in which law firms themselves were increasingly building technology. His focus was on supplying the materials for that work, whether the builder sat inside a firm, a legal department, or a software company. His customers’ ambitions now extended beyond selecting a finished application. They wanted to assemble tools around their own expertise.

The setting has changed considerably since his days writing small scripts. The thread is still recognizable. A professional understands a task, notices the awkward parts, and wonders whether the available tools could be improved. Boehmig once answered that question by finding a tutor. He now works with an industry full of people asking versions of it. There is still plenty of paperwork to discuss.