By Michael Drapkin and Nathan Mutter

Patents can help startups establish competitive advantage, attract investors and protect innovations. But the traditional patent process is expensive, slow and difficult to navigate, particularly for companies with limited resources.

AI is changing that equation. General-purpose tools can research existing patents, document inventions and generate drafts that look like patent applications in minutes. Handling the process themselves can seem like an easy way for founders to save money.

But AI’s capabilities are jagged. AI can perform some tasks remarkably well while missing legal and strategic issues that determine a patent’s commercial value. An application can look sophisticated while providing little meaningful protection.

For example, a patent’s claims define its legal boundaries and what you can stop competitors from doing. Mistakes involving claim scope may remain hidden until an investor, acquirer or competitor scrutinizes the company’s intellectual property.

The question isn’t whether startups should use AI. It’s how much founders can safely do themselves and when legal expertise becomes essential.

Where AI can help startups save time and money

Founders can use AI to organize technical documentation, summarize existing patents and identify potentially relevant prior art. Turning engineering notes into an organized invention disclosure, for example, can reduce preparation costs and help counsel focus on strategy.

But the opportunity extends beyond DIY. The right counsel can orchestrate AI across the workflow, coordinating invention discovery, drafting, drawings and quality checks. AI can help counsel explore alternative implementations with the inventors, identify disclosure gaps and check support for claims. Attorneys can direct that work, evaluate the results and connect each step to best protect business needs.

That combination can produce more comprehensive applications with substantially greater efficiency. It is fundamentally different from generating an application with a chatbot and asking a lawyer to review it afterward.

When DIY patent filing becomes a liability

• Incomplete provisional applications. A provisional’s early filing date is only as useful as its disclosure. Later claims receive that date only if the provisional adequately describes and enables the claimed invention. A DIY draft may omit technical details, alternatives and variations essential to future claims. Adding those details later cannot repair the original disclosure retroactively.
• Premature disclosure. Pitching investors, demonstrating products or announcing innovations before filing can jeopardize patent rights, particularly internationally. Public disclosure can also destroy trade secret protection. Sharing confidential invention details with AI tools without evaluating their data practices creates additional risks.
• Unclear ownership. Paying an employee or contractor does not substitute for securing appropriate invention assignments. Secure those agreements when people join, rather than waiting for diligence to expose gaps. Founders should also assign relevant work to the company and verify obligations to prior employers or universities. Ownership gaps can derail fundraising or acquisitions.

The cheapest filing option is not necessarily the most economical. A poorly prepared application can create false confidence while leaving valuable innovations exposed.

When legal counsel matters

Counsel can help decide what to patent versus maintain as a trade secret, prioritize filings within the company’s budget and assess deadlines before disclosure or filing limits the available options. An internal process that competitors cannot observe or reverse engineer, for example, may be better kept as a trade secret.

Claim scope requires particular judgment. A fraud detection claim tied to one algorithm may leave competitors free to deliver the same functionality using another. Overly broad claims may overlap with existing technology. Valuable claims target what competitors must do to compete. Ideally, a single identifiable competitor can infringe them, and that infringement can be detected and proven.

The specification must describe alternatives and variations that support broader claims later without unnecessarily limiting their scope. AI can assist, but strategic judgment must guide the work.

Why investors look beyond ‘patent pending’

Filing volume says little about commercial value. Investors and acquirers will look beyond the number of pending applications to ask:

• Do you own it? They want signed assignments from relevant founders, employees and contractors, with no competing claim from a prior employer or university.
• Does it protect your moat? The claims should cover what the business depends on, not a feature it will abandon next year, and the application must contain enough detail to support them.
• Can you enforce it? The claims should be difficult to design around, and infringement should be something you can realistically detect and prove.

Making patent protection more accessible

AI’s greatest value is not simply generating more applications at lower costs. It is making meaningful protection more accessible without sacrificing legal judgment.

Founders should ask how a firm coordinates its AI tools and legal expertise, not simply whether it uses AI. The answer should explain how that process improves both efficiency and substance.

Ultimately, the most expensive patent mistake may be discovering too late that a DIY application never adequately protected the innovation on which the business was built.

Related Crunchbase query:

Michael Drapkin is a founding partner of Landfall IP, an AI-native patent law firm based in Boulder, Colorado. He previously founded one of the largest electronics and computer patent prosecution practices in the U.S. at an Am Law 100 firm and was named U.S. Patent Practitioner of the Year by Managing IP in 2022.

Nathan Mutter is managing partner and founder of Landfall IP, an AI-Native law firm. He has served as top outside counsel at Oura, Coinbase, Rubrik, Salesforce 1 and many others. Mutter pioneered the creation of automation systems and tools that enhance efficiency across all aspects of legal services. His clients include both startups and large public companies.

Photo by Conny Schneider on Unsplash