Intellectual Property Review Checklist for Product Features

Conduct an IP review during the design phase, before engineering starts building. The earlier you catch IP risks, the cheaper they are to resolve. A design-around in the planning stage costs hours. A redesign after launch costs sprints. Litigation costs years. At minimum, review IP risks for any feature that involves: new algorithms, AI/ML models, third-party data, open source dependencies, or entry into a new product category.

PMs are not IP attorneys. Your role is to surface risks, not evaluate them. Run this checklist, flag items that seem risky, and bring them to legal. PMs also play a key role in identifying patentable innovations within their product. If your engineering team builds a novel approach to a hard problem, document it and file an invention disclosure. Many companies have inventor incentive programs that reward this. See the [Product Strategy Handbook](/strategy-guide) for how IP fits into competitive positioning.

The most common conflict is using a GPL-licensed library in a proprietary product. GPL requires that derivative works also be open-sourced. If this conflicts with your business model, you have three options: find an alternative library with a permissive license (MIT, Apache 2.0), isolate the GPL code in a separate service (the "GPL firewall" approach), or negotiate a commercial license from the library maintainer. Use the [Open Source Audit Template](/templates/open-source-audit-template) for a thorough dependency review.

This is an evolving legal area. In most jurisdictions, AI-generated content has unclear copyright status. The US Copyright Office has stated that purely AI-generated works cannot be copyrighted. Outputs that involve "sufficient human authorship" (editing, selecting, arranging) may qualify. For products that generate content for customers, your [terms of service](/templates/terms-of-service-review-template) should clearly define who owns the output and what rights are granted.

A patent grants you a 20-year monopoly on an invention in exchange for public disclosure. A trade secret protects confidential information indefinitely but only as long as it remains secret. Choose patents for innovations that competitors could independently discover or reverse-engineer. Choose trade secrets for processes that are hard to detect from outside (algorithms, internal scoring models, data pipelines). Many companies use both: patent the user-facing innovation, keep the backend optimization as a trade secret. ---