Frequently Asked Questions

The typical approach is to hire lawyers and patent experts which is expensive, time consuming and error prone. Commercial patent tools automate certain tasks but are not geared for value analysis and are expensive. Tools also rely on patent office data and there is simply no information that can be directly applied to determine value since patents are prosecuted irrespective if they have value or not. Pelent is an AI rule-based system using USPTO and US patent infringement data. In essence the system leverages the collective knowledge of thousands of legal, technical and business experts combined with millions of records and real-time data. We provide a systematic approach to efficiently analyze a patent or a portfolio of any size producing results that are data driven and accurate.

Artificial intelligence (AI) refers to a wide variety of computational approaches that emulate human intelligence such as learning, reasoning and problem solving. Pelent’s system is built on subsets of AI called Natural Language Processing (NLP) and Machine Learning (ML). NLP allows computers to understand human language. (ML) is based on models to learn from datasets to perform tasks such as classification or prediction. Our system is designed and optimized to analyze patents and answers the fundamental questions why a patent is special for patent licensing – what’s the invention, is it being used and is it legally enforceable?

Our NLP engine is based on open-source GNU software (developed by the NLP Group at Catalunya Polytechnic University) onto which we layer our proprietary algorithms and models. Our database trains our models for better performance; the system acquires data daily from the USPTO and US patent infringement litigations.  Over a hundred thousand litigations have been analyzed, indexed, and archived.  Since 2016 the database has grown considerably to almost 1 terabyte – equivalent to 200 million books. The net result is that patent analysis takes minutes versus hours, a portfolio typically takes hours versus several days or weeks. Version 2.0, released in 2019 improved performance, reliability and security. The system runs 24/7 using high performance cloud-based servers.

The system has been verified and validated with over  20,000 US patents with known value achieving greater than 93% accuracy. We continuously verify the system with patents and portfolios with known value. More often the system identifies patents incorrectly assessed when done manually or using other tools. Lastly the system is designed for continuous self-improvement – as the database grows system accuracy improves.

Valuable patents are patents that provide economic benefit and can be legally enforced against infringement. Patents are rated according to Business Value and Legal Value, each individually scored and combined to formulate the Patent Rating. Valuable patents therefore have both high business value and high legal value. Checkup analyzes each patent claim and the final rating is determined by the independent claim with the highest Patent Rating.

The system is designed and optimized for US patents. The system supports publications of the United States Patent office including granted patents and published patent applications. The system does not support other foreign jurisdictions or US pharmaceutical patents.

Claims are the most important part of a patent and their analysis is an absolute requirement for determining value. Unidentified claim elements can severely limit the value of a patent or cause hundreds of thousands of dollars in wasted efforts and legal costs. Invariably each and every limitation or claim element must be identified – a doctrine known as the all-elements rule.

No.  Send us your patent list and we will return your Patent Rating results. There is no charge or registration required.