Since the Supreme Court's 2014 Alice Corp. v. CLS Bank decision, software patents have to clear a much higher bar than a decade ago — the claims need to describe a specific technical improvement, not just an abstract idea implemented on a computer. Craige Thompson, an EE-licensed patent attorney, drafts software claims built to survive that scrutiny, from AI/ML models to SaaS platforms and backend algorithms.
A software patent protects an invention where the innovative step is implemented in code — not the idea alone, but a specific technical way of solving a problem. Examples we regularly handle:
What doesn't qualify: an abstract business method or mathematical concept with no specific technical implementation. Part of our job during your free assessment is identifying exactly where the line falls for your invention, and drafting claims that land clearly on the patentable side of it.
Most patent attorneys who handle software work are trained as lawyers first. Craige Thompson is a licensed electrical engineer (PE) and USPTO-registered patent attorney (Reg. No. 47,990) — he can read your architecture diagrams and codebase documentation directly and draft claims that accurately capture the technical innovation, rather than translating a simplified explanation secondhand. That precision is what makes a software patent durable enough to hold up in licensing negotiations or litigation.
Tell us about your invention in a free patent assessment. We'll tell you honestly whether it's patentable, what it would take to protect it, and what it would cost.
Get a Free Patent AssessmentYes, if the app or software includes a specific technical improvement, such as a novel way of processing data, improving system performance, or solving a technical problem. A generic idea for an app (e.g., "an app that lets users order food") is not patentable on its own; the specific technical implementation can be.
It's the two-step framework courts use, from the 2014 Alice Corp. v. CLS Bank Supreme Court decision, to decide whether a software claim is patent-eligible. Step one asks whether the claim is directed to an abstract idea; if so, step two asks whether the claim adds an "inventive concept" beyond that abstract idea. Claims that recite a specific technical solution, rather than just applying a generic computer to an old idea, are far more likely to survive.
No. You need to be able to describe the invention in enough technical detail that someone skilled in the field could implement it — architecture, data flow, and the specific technical steps. Working code can help illustrate this but is not itself required.
Typically 18 months to 3+ years from filing, depending on the technology area and USPTO backlog, with software/AI applications often falling in art units with longer average pendency. Track One expedited examination can shorten this substantially for an added fee.
Yes — both software that uses AI/ML techniques as part of its solution, and in some cases specific improvements to AI/ML methods themselves, can be patentable, provided the claims describe a specific technical improvement rather than the general concept of "using AI" to solve a problem.
