Prior art search is the foundation of every patent application. A thorough search determines whether an invention is patentable, shapes claim scope, and prevents costly rejections during prosecution. This guide covers the strategies, databases, and tools that patent attorneys use to conduct effective prior art searches.
What Counts as Prior Art?
Under 35 U.S.C. §102, prior art includes any patent, published patent application, or other printed publication available before the effective filing date. It also includes public use, sale, or other availability to the public.
For practical purposes, patent attorneys search across several categories:
- Patent literature: Granted patents and published applications from patent offices worldwide
- Non-patent literature (NPL): Academic papers, conference proceedings, technical standards, product documentation
- Public disclosures: Press releases, product launches, trade show presentations, website publications
Key Patent Databases
A comprehensive prior art search spans multiple databases. No single source covers everything.
USPTO Patent Full-Text and Image Database
The USPTO search tools provide access to all US patents and published applications. PatFT covers granted patents back to 1790; AppFT covers published applications from 2001 onward.
Google Patents
Google Patents aggregates patent documents from over 100 patent offices worldwide. It offers full-text search, CPC code filtering, and machine-translated documents from non-English jurisdictions.
Espacenet (EPO)
The European Patent Office's Espacenet provides access to over 140 million patent documents from around the world, with machine translation from 32 languages.
WIPO PATENTSCOPE
PATENTSCOPE from the World Intellectual Property Organization provides access to international PCT applications and national patent collections.
Search Strategies That Work
1. Start with CPC Classification Codes
The Cooperative Patent Classification system organizes patents by technology area. Starting with the right CPC codes narrows your search to relevant technology fields and reduces noise. Use the CPC scheme to find the right codes for your invention.
2. Decompose the Invention
Don't search for the entire invention at once. Break it down into its core elements and search for each one separately. This catches prior art that covers individual components even if no single reference covers the whole invention.
3. Use Synonyms and Variations
Patent language varies widely. An invention described as a "controller" in one patent might be a "processor," "computing device," or "logic unit" in another. Build search queries that account for these variations.
4. Search Iteratively
The first search is a starting point, not the final answer. References found in early searches often reveal new keywords, CPC codes, and cited references that lead to additional relevant art.
How AI is Transforming Prior Art Search
Traditional prior art search is manual and time-intensive. An attorney might spend 10-20 hours searching databases, reading documents, and mapping the prior art landscape. AI-powered tools compress this to hours.
Modern AI search tools like PatentLawyer automate the decomposition strategy: they break the invention into elements, generate targeted queries for each, search across multiple databases simultaneously, and read full patent documents to assess relevance.
The result isn't just faster — it's more thorough. AI tools can process hundreds of full patent documents in the time it takes an attorney to read ten.
From Search to Application
A prior art search isn't an end in itself. The results directly shape the patent application:
- Claim scope: Prior art defines the boundaries. Claims need to be narrow enough to avoid the art but broad enough to provide meaningful protection.
- Specification detail: The specification should distinguish the invention from the closest prior art references.
- Prosecution strategy: Knowing the art upfront means fewer surprises during examination and more strategic responses to office actions.
The best patent applications are built on thorough prior art searches. Whether you search manually or use AI-powered tools, the depth of your search determines the strength of your application.
PatentLawyer searches 100+ patent offices automatically.