Founders building artificial intelligence into their products sometimes assume there is nothing to protect because the technology is “just a feature,” “just a workflow,” or “just a wrapper” around an existing model.
Those labels do not answer the legal question. What matters is what the team built, how it works, who created it, and how it differs from what came before.
I hear several versions of the same questions from founders and engineers. Here are the answers in a direct, practical form.
We’re Using an Existing AI Model. Is There Anything to Protect?
Possibly. Using an existing model does not resolve the question either way.
A team may build a particular process for preparing inputs before they reach the model, validating outputs, coordinating several systems, or addressing a technical limitation that the underlying model does not solve on its own. Those components may be evaluated separately from the model.
The model may be a third-party tool while the architecture surrounding it contains the team’s own work. Even then, the fact that the work is useful or detailed does not establish patentability. The proposed claims still must satisfy the requirements for patent eligibility, novelty, and nonobviousness.
The more useful question is not, “Do we own the model?” It is, What did we build around the model, what rights do our third-party agreements allow, and what technical problem does our contribution address?
Does Adding AI to Our Product Make It Patentable?
No. The reverse is not true either: adding artificial intelligence does not make a product unpatentable.
Patent law does not treat the label as the invention. A generic computer implementation of an abstract idea is not enough to create patent-eligible subject matter.Alice Corp. Pty. Ltd. v. CLS Bank Int’l, 573 U.S. 208, 212 (2014). The analysis instead considers the substance of the claimed invention and whether the claim reflects more than an abstract result implemented through conventional technology.
A recommendation function attached to a standard database does not become patentable merely because the company describes it as artificial intelligence. A specific architecture or technical method may deserve evaluation, but only the claims and the relevant prior art can support a reliable conclusion.
Direct your attention to the mechanics, not the label.
We Built an Internal Tool to Speed Up Our Operations. Does That Count?
An internal tool can be evaluated for patent protection even if customers never use it. Internal deployment is not itself a bar to patentability.
At the same time, keeping a tool inside the company does not make it patentable or necessarily make it more defensible. Secrecy may support a trade-secret strategy if the information has the required value from not being generally known and the company handles it accordingly, but secrecy does not establish patent eligibility, novelty, or nonobviousness.
The practical question is whether the tool includes a potentially protectable technical contribution and, if so, whether patent protection, trade-secret treatment, or another strategy better serves the company’s goals.
Our Engineer Who Built This Left the Company. Who Owns It?
The answer depends on the type of intellectual property, the person’s legal relationship with the company, the work performed, and the documents that were in place.
For copyright purposes, an employee’s work may qualify as a work made for hire when created within the scope of employment. Whether someone is an employee or an independent contractor for this purpose is determined under common-law agency principles, not simply by the title used in the agreement.Cmty. for Creative Non-Violence v. Reid, 490 U.S. 730, 740 (1989).
Patent inventorship is a separate issue. Patent law begins from the premise that rights in an invention belong to the inventor unless those rights are transferred.Bd. of Trs. of the Leland Stanford Junior Univ. v. Roche Molecular Sys., 563 U.S. 776, 780 (2011). A company therefore should not assume that ownership of source code necessarily resolves ownership of a patentable invention.
Review the applicable employment or contractor agreement, any invention-assignment terms, and the person’s actual contribution. The timing and wording of the documents can matter, as can the distinction between authorship, inventorship, and ownership.
Please note that a departing engineer may have written code, contributed to an invention, or done both. Those are related but legally distinct questions, and the answer should not be based solely on job title or payment history.
A Competitor Could Build Something Similar in Six Months. Does That Matter?
It can matter, but not only because the competitor may catch up commercially.
Later publications, public uses, sales, and patent filings can affect the available patent landscape. Federal law measures prior art against the effective filing date of the claimed invention and provides only defined exceptions for certain disclosures.35 USCS § 102.
A competitor’s ability to develop similar technology does not by itself determine whether your work is protectable. But waiting may create additional prior-art, disclosure, and filing-priority issues. It may also reduce the practical value of protection if the market moves before the company acts.
The important dates include when the invention was developed, when it was disclosed or offered for sale, and when relevant patent applications were filed. The degree of technical overlap also matters. A competitor’s general announcement is not enough, by itself, to resolve those questions.
Where This Leaves You
None of these answers supplies a simple test for determining whether an artificial-intelligence feature is protectable. The result depends on what was built, how it operates, what was already known, who contributed to it, and what agreements govern the relevant rights.
The useful next step is not automatically a patent application. It is a focused conversation about the technology and the company’s objectives.
If your company has built an artificial-intelligence system within the past year, whether customer-facing or internal, identify the technical problem it addresses, the specific contribution your team made, who participated in that work, and what has already been disclosed. That review is more useful before a competitor, acquirer, investor, or departing engineer forces the issue.
The Garcia-Zamor Law Firm. We’re the general counsel and fractional general counsel for businesses and high-end innovators, protecting both your business operations and your intellectual property. Ruy Garcia-Zamor (founder with 25+ years experience in patents, trademarks, intellectual property, and business strategies, and a registered patent attorney with the U.S. Patent and Trademark Office), Elliott Alderman (40+ years experience in intellectual property and providing guidance to businesses), Claudia Castillo (decades of experience in business law focusing on all employment issues), and Amulya Annasamudram (focuses on patents and intellectual property and is a registered patent attorney with the U.S. Patent and Trademark Office). Contact us at garcia-zamor.com or (410) 531-9853.
