Artificial Intelligence at EconSight
The EconSight AI concept
State of the art, modular, comprehensive.
Flexibility
Modular structure
All modules can be combined as required or implemented into client environments
Speed
fast results
no extensive pre-training and pre-categorization and selection required
Precision
pre-trained AI models
We use the latest pre-trained LLM models and we are not restricted to any specific model
Completeness
full text based
We have analysed the full patent documentation not just the first page
Artificial and human intelligence combined
We have developed our AI concepts with the client in mind. We have combined state of the art AI with more than 20 years of patent expertise to achieve the best possible combination of human and artificial intelligence. Our systems work on two levels: technology concepts and companies. We have implemented our extensive human curated cutting-edge technology taxonomy and our consolidated patent owners.
Whereas previously most AI systems required a lot of input for a small output, we have turned this around. This leads to significantly more precise results with significantly less intellectual effort up front. The expert can focus on the questions, and we achieve a very good selection of patents or companies as a result.
Patent Categorisation
Precise mapping of large patent sets to existing or newly defined categories based on the full patent text. The categorization is always multilabel with content weighted reliability.
Similarity Analysis
Identification of the most similar text passages based on a patent, a draft or a descriptive concept. Extraction from 140 million patent corpus with maximum precision. Proven in validity, novelty and FTO patent searches.
Freedom to operate analysis
Our proven multi-step approach delivers the relevant patent full texts, claim-specific trained, complete, prioritized and pre-sorted. Ideal also for short-term assessment of a business idea or a start-up concept.
Competition Analysis
Identification of companies with technologically similar activities in recent years. Identification of new competitors or traditional competitors who are leaving the field.
Technology Clusters
Identification of the most interesting clusters within a technology in terms of dynamics, quality, and maturity. Benchmarking company patent portfolios against cluster dynamics to identify relevant patents.
Draft Predictions
Prediction of grant probability to identify the more valuable patent drafts or conceptual ideas before filing – in a full secret in-house setup. A proven multi factor scoring system (uniqueness, citation score prediction, and cluster projection)
Patent/Concept Positioning
Analysis of patent drafts, ideas, and concepts to best position them in relevant, interesting, and growing technology clusters.