An Advanced Patent Search Engine Call us at 919-599-5512


I have heard of semantic search engines before. What makes your technology “smarter” than the rest?

Artificial Intelligence Patents Fiber Optic Example

Our algorithm largely depends on patent examination documents that are written on a daily basis from patent examiners at the USPTO and EPO. Our engine has learned from the “prior art” documents that patent examiners have cited to reject pending patent applications claims. As new technical terms are adopted over time, e.g. Intel’s “hypervisor”, our engine dynamically learns about those other terms that are mostly closely associated with the introduced words (i.e. virtual machine manager may represent words that are identified to be technically associated with hypervisor).

An extremely simplified, but illustrative example: patent examiners, in their daily reviews of inventions, have recognized that a “mobile communications terminal” is equivalent to a “portable telecommunications unit” which may be deemed equivalent to a “cellular phone”, and so on. By incorporating these examiner-determined associations into a user’s search, a more complete result set is achieved.






I am used to traditional Boolean and proximity-based searching. Is this dramatically different?

Yes, it’s much easier to use as our engine allows you to type or copy and paste into the search box any sequence of words. In performing a search focused on patent invalidation, you may just copy and paste all of claim 1 into the search box, coupled with the filing date limitation.

If you wanted to review an invention disclosure from one of your scientists, you may simply copy and paste his or her invention summary within the search engine. Often, simply copying and pasting the inventive concept into the search box yields the highest quality of results.

At the same time, we do offer Boolean functions including the traditional AND and NOT functions. In addition to the free-form textual input box, the searcher can mandate a certain word or synonym associated with those words. Frequently, this approach works well as you are including the inventive concept’s “must have words” along with the greater context of the invention. Overall, however, there is no absolute need for Boolean/operators, stemming symbols, and parenthesis matching when using our search engine.


What is the difference between your main search window (i.e. the one that you can copy and paste into without regard to syntax) and the smaller search input box?

Yes, our user interface has two main search input windows.  One is a free-form text box where inventive concepts are written.  By searching using this input, the searcher relinquishes all the power to the engine itself to internally supplement the search process with its own synonyms.  The second box is a Boolean text input, where one can mandate a certain word or words to be included in the results or not included (whereas the free form text box does not mandate that each word in the search query to be found in any of the result set documents).


Can the user search for patents that are related to a specific product?

Definitely – this is one of the big advantages this tool has over other patents search engines. This is a feature that we are constantly building out as this will incredibly help marketers, product developers, lawyers, patent aggregators, and so forth.  For example, a product manager may simply want an overall landscape of the published product information as this will help her shape the positioning, R&D, and pricing related to her own product.  An IP Licensing manager may want to get an idea if anyone is infringing on his company’s patents; as a result, he may simply copy and paste the independent claims of his company’s patent assets into the search box and then check for infringers.