On April 30, 2025, the Federal Circuit issued a precedential decision in Fintiv, Inc. v. PayPal Holdings, Inc., No. 23-2312, with far-reaching implications for the validity of patents in financial technology (fintech) and artificial intelligence (AI). The court found several claims in Fintiv’s patents invalid for indefiniteness under 35 U.S.C. § 112 ¶ 6. This decision underscores a recurring vulnerability in software-based patents: reliance on functional language unsupported by sufficient structural disclosure.
The Patent Bargain and Its Breakdown
The U.S. patent system is built on a quid pro quo: in exchange for a limited monopoly, inventors must publicly disclose how their inventions work. This is codified in the “written description” and “enablement” requirements of 35 U.S.C. § 112. To be valid, a patent must describe the invention in enough detail to enable a person skilled in the art to make and use it.
While most inventors do not intentionally withhold information, many rely on abstract, functional terms—such as “cloud-based processing system” or “payment handler”—without supplying the necessary technical specifics. When this happens, courts may declare the patent claims indefinite, meaning they fail to inform the public what the invention actually is or how to replicate it.
The Fintiv Decision: A Wake-Up Call
In Fintiv, the patents claimed a cloud-based financial transaction system using components described in purely functional terms. Although the claims avoided the term “means”—which typically triggers § 112 ¶ 6—they nonetheless fell into its scope. The court held that terms like “payment handler” lacked any definite structure and operated as generic black boxes.
Importantly, the court emphasized that:
- Simply swapping the word “means” for a nonce term like “handler” does not avoid § 112 ¶ 6.
- If a claim recites a function without providing the structure to perform it, the claim may be deemed indefinite.
- Even if a claim term lacks structural meaning, the specification may rescue it—but only if it discloses an algorithm or other concrete structural detail clearly linked to the function.
In Fintiv, the court found no such saving disclosure. The patent described what the “payment handler” was expected to do, but offered no description of how it did it. That omission was fatal.
Why This Matters for AI Patents
AI patents are especially susceptible to the Fintiv problem. AI innovations are often described in terms of outcomes—predicting, classifying, generating—without disclosing the inner workings of the models that produce those results. Courts are increasingly skeptical of patents that invoke terms like “classifier,” “scoring engine,” or “intelligence module” without accompanying detail.
Moreover, many AI models (e.g., neural networks, LLMs) are treated as black boxes even by their inventors. But when it comes to patents, opacity is a liability. Without structural or algorithmic support, claims that rely on these black-box components risk invalidation for indefiniteness.
Drafting Stronger AI Patents Post-Fintiv
To avoid the pitfalls highlighted in Fintiv, AI and fintech patent drafters should observe these best practices:
1. Avoid Purely Functional Language
Do not rely solely on high-level descriptions of what a component does. Specify how it operates or interacts with other components in structural terms.
2. Include Detailed Algorithms
If a function is implemented by software, disclose the algorithm—via flowcharts, pseudocode, or step-by-step logic. This is especially critical when the function is performed by a general-purpose processor.
3. Define and Support Technical Terms
Clearly define coined terms like “prediction unit” or “decision module,” and explain their architecture, dependencies, and processes. Avoid relying on catch-all language or labels.
4. Incorporate Tangible Components
Tie claim elements to physical components where appropriate—e.g., processors, memory, data structures. Make explicit how each structure enables the claimed functionality.
5. Diversify Claim Types
Use a variety of claim forms: method, system, apparatus, and computer-readable medium claims. If one form is invalidated, others may survive.
Final Thoughts
Fintiv is a cautionary tale for any entity drafting or enforcing patents in fintech or AI. Functional claiming, though convenient, cannot substitute for technical disclosure. Courts now demand more than abstract ideas and vague system descriptions—they want to see the mechanics.
Robust AI and fintech patents will be those grounded in specificity, backed by algorithms, and crafted with judicial scrutiny in mind. As the Federal Circuit made clear: if the court can’t see how your invention works, your patent will not survive.
Need help future-proofing your AI patent portfolio?
At Brett J. Trout, PC, we guide innovators through the complexities of patent law in rapidly evolving fields. Contact us today to ensure your patents are enforceable, defensible, and ready for the future.
Recent Comments