What’s New: USPTO Embraces Evidence-Driven § 101 Practice
The U.S. Patent and Trademark Office (USPTO) recently issued two coordinated memoranda explaining how applicants can use Subject Matter Eligibility Declarations (SMEDs) to more effectively overcome § 101 rejections. The first memorandum, “Best Practices for Submission of Rule 132 Subject Matter Eligibility Declarations (SMEDs),” is directed to applicants and practitioners; the second one, “Subject Matter Eligibility Declarations,” is addressed to the Patent Examining Corps. Although the memoranda do not constitute new rulemaking but instead clarify an existing evidentiary mechanism, they provide detailed guidance on what SMEDs should contain and expressly instruct examiners to meaningfully consider them in § 101 determinations.
Practical Impact on AI/ML, Software, and Data Architecture Claims
This guidance builds on the Director’s designation of In re Desjardins as precedential, which held that improvements in machine learning (ML) architectures (e.g., as reduced storage, lowered system complexity, etc.) constitute technological improvements under Alice Step 2. The Appeals Review Panel cautioned that overbroad § 101 rejections risk “jeopardizing America’s leadership” in artificial intelligence (AI).
Together, the memoranda reflect a shift toward evidence-based eligibility analysis by offering applicants a more predictable mechanism for managing § 101 risk across various subject matters, including AI/ML, software platforms, distributed systems, and data architectures. Strategically deploying SMEDs is expected to allow applicants to shift § 101 practice from reactive prosecution to proactive evidentiary planning, with portfolio-level implications for pendency, prosecution costs, and allowance rates.
What a Persuasive SMED Looks Like: Factual Nexus and Technical Proof
A persuasive SMED must demonstrate a factual nexus between the claimed features and technological improvements, such as enhanced computational efficiency, new architectural arrangements, automation of tasks previously requiring human judgment, or improved model-training techniques. Both memoranda reiterate that SMEDs cannot add new disclosures but may explain how a person having ordinary skill in the art would understand the specification and may include objective testing or comparative data.
Notably, the examiner-directed memorandum contains fictional examples of common § 101 rejections (e.g., such as mental-process characterizations, abstract-idea findings, or assertions that elements are conventional) paired with illustrative SMED submissions that would overcome each rejection. These examples serve as practical templates, signaling the types of evidence examiners are now directed to credit.
Standard of Proof and Examiner Obligations
Examiners are instructed to evaluate SMED evidence under a preponderance-of-the-evidence standard and to explain in the next Office action how the SMED impacts the eligibility determination. This requirement increases transparency and reduces examiner discretion.
Thus, the USPTO’s December 2025 memoranda collectively elevate SMEDs into a core strategic tool for overcoming § 101 rejections and improving prosecution outcomes in complex technologies. Accordingly, we recommend that our clients:
- Integrate SMEDs into overall prosecution strategy, particularly for high-value inventions where § 101 risk is considerable
- Revise invention disclosure processes to capture performance data, architectural details, and other technical facts that may later support a persuasive SMED
