Fearn, a startup backed by Andreessen Horowitz, has raised $5.5 million to build artificial intelligence software that lets inventors draft their own patents without hiring expensive lawyers. The company is betting that one of legal technology's biggest opportunities isn't serving law firms but helping companies handle more legal work in-house, reducing their reliance on outside counsel.

Why Is Patent Drafting Ripe for Automation?

Han Kim, Fearn's cofounder and chief executive, came to this idea through personal experience. He spent years prosecuting patents at Morrison Foerster, an elite law firm, and saw firsthand how the process frustrated inventors. Patents are highly structured documents with predictable formats, making them ideal candidates for artificial intelligence automation. Additionally, patent prosecution requires deep technical fluency, not necessarily a law degree, which opens the door for software to handle much of the work.

"Their primary anxiety, in my experience, has been that the patent lawyer does not understand their invention. They're going to mangle it," said Han Kim, cofounder and CEO of Fearn. "But the inventor can't even tell because it's written in legalese."

Han Kim, Cofounder and CEO at Fearn

Kim's frustration with his old job became the spark for Fearn. He realized that patent drafting didn't have to be as painful as it was, and that automation could genuinely improve the process for inventors who know their work better than anyone else.

How Does Fearn's AI Patent Tool Work?

• Document Upload: Inventors upload technical documents describing their invention, providing the raw material the AI needs to understand what they've created.
• Automated Drafting: The software generates a complete patent draft with a single click, cutting the typical timeline from weeks down to minutes.
• Quality Scoring: The system scores the quality and completeness of the material provided, helping inventors understand whether their documentation is sufficient.
• Visual Generation: Fearn automatically creates patent drawings and shows its sources, making the process transparent and reducing back-and-forth revisions.
• Flexible Next Steps: After drafting, inventors can file directly with the patent office or send the application to outside counsel for review and refinement.

The company charges a flat $2,000 per patent draft, a significant savings compared to traditional patent prosecution, which can cost tens of thousands of dollars. Fearn says it is already being used by companies across robotics, pharmaceuticals, energy, and gaming sectors.

What's Driving the Shift Toward In-House Legal AI?

Investment in legal technology has surged dramatically in recent years, growing from roughly $1 billion in 2019 to more than $4 billion last year. However, the legal tech landscape is splitting into two competing visions. Harvey, a legal AI startup valued at $11 billion and backed by Andreessen Horowitz, represents the dominant thesis: artificial intelligence will help lawyers work faster and handle more cases. But Fearn and other startups like Wordsmith and Sandstone are chasing a different opportunity, helping companies bring legal work in-house and reduce spending on outside counsel.

This shift reflects a broader trend in enterprise software, where companies increasingly want to handle specialized work themselves rather than outsourcing it. For patent drafting specifically, the advantage is clear: inventors understand their own innovations better than any outside lawyer, and automating the translation from technical knowledge to legal language could reduce misunderstandings and speed up the entire process.

Who's Backing Fearn and How Did They Get Here?

Fearn's seed round was led by Kindred Ventures, with participation from Designer Fund and Essence Venture Capital, alongside Andreessen Horowitz. The company's founders, Han Kim and Angela Gao, were first-time entrepreneurs who didn't have deep connections in Silicon Valley's venture capital networks. Last year, they applied to Speedrun, Andreessen Horowitz's 12-week startup accelerator program, and were accepted despite their outsider status. That acceptance proved to be a turning point, eventually leading to the firm's investment in their company.

Gao brings strong technical credentials to the partnership. She completed a PhD in computer science at Caltech under Katie Bouman, the scientist known for her work on the first image of a black hole. Kim studied computational neuroscience at Caltech, giving the founding team both deep technical expertise and legal domain knowledge.

What Competition Does Fearn Face?

Fearn is entering one of legal technology's most crowded races. Direct competitors include other patent-focused startups like Patlytics and DeepIP. But the competitive landscape extends far beyond specialized patent tools. Harvey, despite being primarily focused on helping lawyers work faster, could be used for parts of the patent drafting process, though Harvey's leadership has acknowledged that patent drafting is not one of the company's strongest use cases. Additionally, general-purpose AI chatbots are becoming harder to ignore. Last month, Anthropic announced that users of Solve Intelligence, another patent drafting startup, can access Solve inside Claude's app. OpenAI is also pushing deeper into legal technology, recently hiring the founder of Ironclad to lead its legal tech efforts.

Despite this crowded field, Kim believes Fearn has a meaningful advantage: one of its founders actually lived the problem. Many competitors came to patent drafting as outsiders chasing what they saw as a hot business opportunity. Kim, by contrast, spent years frustrated by the inefficiencies of the traditional patent prosecution process and built a tool specifically designed to solve the problems he experienced firsthand.