AI-supported IP search

We combine standard IP search methods (e.g. search with keywords, technology classes, applicant information) with machine learning tools, particularly natural language processing (NLP).

We analyze patent and publication abstracts as well as the associated claims. To this end, a text passage of a patent or a publication becomes a multidimensional vector and based on this we then determine similarities with text passages from other patents or publications.  

The application fields of AI-supported IP searches are diverse and, in addition to prior art searches for patent applications, also include freedom-to-operate/white-spot analyses, opposition searches, technological maturity analyses, identification of “patent trolls” and the search for potential cooperation partners.

Objective, comprehensible, replicable and transparent approach to IP research

Inclusion of a wide range of sources
(e.g. foreign patents, scientific publications, databases of funded research projects)     

Many and a large variety of application fields in IP management and IP monitoring

Natural language processing (NLP) methods allow contextual information to be taken into account

Freedom-to-operate- and white-spot-analysis

Freedom-to-operate analyses help to decide whether patents exist that stand in the way of the development, manufacture and market launch of a technical innovation or product. AI-supported patent search methods can be used to systematically identify patents or patent applications that could restrict freedom-to-operate. AI-supported patent search methods can also be used well in white spot analyses with the aim of identifying and occupying “white spots” in the technology and patent landscape.

Opposition-/nullity search

In an opposition and nullity search, patent and non-patent literature is searched for relevant prior art that was published or at least applied for before the priority date of a patent. The search is carried out without restriction to country and language. A combination of standard search methods with AI-supported search methods achieves good results in this regard.

Identification and circumvention of "patent trolls"

"Patent trolls" (also "non-practicing entities") own IP rights. However, they do not use them in their own processes or products. They are also not active as inventors themselves. They typically operate under the radar and use their claims from IP rights to  reach excessive and overly priced license agreements.  AI-based IP searches can help to identify such “patent trolls” and their IP rights and thus avoid potential problems already at an early stage.