Our proprietary AI search is now integrated into the global patent database and available to all. This advanced technology, powered by a specialized AI model, is revolutionizing the way we search and analyze patents.
Specialized AI model for patents
Our new AI model has been specially developed for analyzing patent texts. The model currently processes texts in English. At the same time, we are developing an AI-based translator for patents that will enable input in different languages in the future in order to use our global database even more effectively.
Integration into the existing search functionality
The AI search has been seamlessly integrated into our existing platform. This means that all familiar functions such as combining with classes, applicants and performing automatic searches are still available. Users can also use the familiar filter functions to further refine their search.
Future improvements and optimizations
We are continuously working on expanding the functionality of our AI search. Currently, titles and summaries are fully integrated. Claims are available in an initial version, but will continue to be improved. In early 2025, we plan to release comprehensive optimizations, including AI-based search in descriptive texts.
Search modes and application
Our AI search offers two main modes:
- Enter any text describing the problem.
- Search from a specific patent to find similar patents.
We will first focus on the similarity of the abstracts to illustrate how the AI search works.
Example: An example image shows a patent that describes a bicycle helmet with a rear brake light. We use this patent to demonstrate the search function.
Another example screen shows how up to three patents can be selected from a list to find similar patents
These most similar patents are then displayed, and users have the option of marking exactly those patents that match the requirements.
Further processing with relevant results
Users can continue working with the relevant results list by applying filters such as classes, applicants or countries or by using the results as a basis for further searches. In the example, we have applied a class filter that shows exactly the right classes by analyzing the hits.
The AI search replaces our previous semantic search and offers intuitive text input for the description of technical features, which currently has to be in English.
As with a normal text search, the AI text search block can be combined with other search blocks.
Insight into the future
This represents a further step in our vision of increasing efficiency and precision in patent searches through the use of artificial intelligence. The use of AI in the IP7 Compass is now an integral part of further developments.