What Are SMEDs, and Why Should Inventors Care?

The USPTO has introduced a new tool called a “Subject Matter Eligibility Declaration,” or SMED, to help applicants respond to patent‑eligibility rejections under 35 U.S.C. § 101. On April 30, 2026, the Office published updated “Best Practices” for SMEDs, based on early feedback from applicants and practitioners.

A Modern Substitute for “Working Models”

In the early days of the patent system, some inventions had to be filed with a physical “working model.” That requirement disappeared long ago, but it left a gap for technologies that are hard to demonstrate physically—such as software, AI, quantum computing, and diagnostics. SMEDs are meant to fill that gap: they are written declarations you can voluntarily submit that explain how your invention is actually used in the real world and how it improves technology.

Think of a SMED as a constructive model: instead of a physical prototype, you provide a clear, factual explanation—through someone skilled in the field—of why your invention is better, faster, cheaper, or more efficient than what came before. That explanation is then used by the examiner when deciding whether your claims are directed to patent‑eligible subject matter or just an “abstract idea.”

What Goes Into a Strong SMED?

According to the USPTO memo, a good SMED does three main things:

• Connects the evidence directly to the claims, so there is a clear link (“nexus”) between what is being said and what is being claimed.

• Describes how a typical expert in your field would read your patent’s description and see a real technological improvement.

• Stays within what is already written in the application; it explains, but does not add new subject matter.

In practical terms, your patent team may ask you for detailed, concrete examples: performance improvements, error‑rate reductions, resource savings, or other measurable benefits that show the technology is more than a generic implementation.

Why SMEDs Are Filed Separately

The memo strongly encourages applicants to file SMEDs as standalone documents focused solely on eligibility, rather than mixing them with other issues, such as obviousness or written description. Although the rules technically allow a single declaration to address several types of rejections, combining everything can make the evidentiary record messy and harder for the examiner to interpret. By contrast, a separate SMED keeps the story about eligibility clean and makes it easier for the examiner—and later, the PTAB or a court—to see exactly what evidence supports § 101.

The USPTO notes that its internal training materials and Federal Circuit case law all point in the same direction: evidence works best when it is clearly targeted to the specific legal question it is meant to answer.

What This Means for Software and AI Inventors

If you are working in software, AI, or other data‑driven technologies, you are more likely to encounter § 101 rejections and to benefit from a SMED. The Office explicitly notes that modern frontiers such as software, AI, quantum, and diagnostics do not lend themselves to traditional “models,” so a well-crafted SMED can be a key part of your response. Expect your counsel to focus your input on how your system changes how computers operate, how it improves accuracy or efficiency, or how it solves a technical problem in a specific way—those are the kinds of facts that belong in a SMED.

Bottom Line for Inventors

Based on the April 30, 2026, memo, the USPTO’s current best practice is to use SMEDs as focused, separate declarations that explain the real-world technological impact of your invention. Done correctly, a SMED can make it easier for examiners to understand your technology, strengthen your case for eligibility, and reduce the risk of confusion in the record. If you receive a § 101 rejection, ask your patent professional whether a SMED might help tell the story of your invention more clearly to the Office.