Introducing the PrecisionAI Patent Explorer

The patent search and analysis revolution

«An overwhelming experience. Finally, patents can be analyzed the way they were always meant to be.»

Corporate client

«I can find answers to questions I didn’t know I had.»

Corporate client and participant in the development phase

«An overwhelming experience. Finally, large amounts of patents can be analyzed the way they were always meant to be.»

Corporate client

«I can find answers to questions I didn’t know I had.»

Corporate client and participant in the development phase

Next-generation patent search and analysis based on the latest AI-models enhanced with retrieval-augmented generation (RAG)

Traditional method

Currently, the search and retrieval of potentially relevant patents to the topic is carried out using patent classes and keywords. The analysis of the retrieved patents is performed by patent experts based on expert knowledge and experience.

Deficits of traditional method

Soft definitions such as “bio-” or “generative” are difficult to search and classes for cutting-edge topics are not yet available. The procedure is time consuming and difficult even for patent experts, and it is not possible for non-experts.

New EconSight method

A simple but well defined natural language text concept is used as input for the search and retrieval of relevant patents. The analysis of the retrieved patents is done using the latest LLMs in a chatbot style Q&A. The EconSight PrecisionAI Patent Explorer is based on the latest generative AI-models embedded in a retrieval-augmented generation (RAG) system for highest accuracy, conceptual relevancy and exact source citation.

Benefits of EconSight method

Search and retrieval is possible for soft-terms and future class concepts. The analysis is fast, flexible, and precise – usable for experts and non-experts. The whole approach is open to many more analytical options including non-patent data.

We have developed a multi-stage process. The first step is the precise selection of a patent data set based on a cutting-edge AI analysis. In the second step, a ChatGPT-style interface is used to analyze and interrogate the dataset. All other currently available approaches use a ChatGPT LLM on a vague and undefined data set. The results are imprecise and the hallucination problem typical of AI often arises. However, it is not enough for such results to be “pretty good”, “mostly relevant” or “relatively precise”. As long as they are not clear, precise and reliable, they are useless for a well-founded patent analysis. This is precisely where we come in and achieve “revolutionary” results with our new PrecisionAI Patent Explorer.

We currently offer our Patent Explorer as a consulting service. Get in touch with us

No, our approach goes well beyond ChatGPT. Think of ChatGPT as the interface between you and the underlying information. Our added value is primarily in the AI-based precise selection of the patents to be analysed.

The Patent Explorer is currently used primarily in the corporate environment for highly specific questions regarding new markets and new applications. The Patent Explorer is also used by experts for exhaustive overviews of methods and concepts used for specific use cases. Searching for concepts and summarizing the most important methods in a particular field provides many new insights and can be realized in the shortest possible time. Speed and precision have been mutually exclusive in the past, but with the Patent Explorer this is possible.

EconSight specializes in highly complex patent analyses for patent experts and also wants to make patent analysis accessible for non-experts. Clients come to us because they can’t get satisfactory answers to their questions elsewhere. Demand has increased significantly for information about cutting-edge technologies which cannot be defined with traditional classes and keyword searches. Over the last three years, we have invested heavily in AI to take conceptual (not classes and keywords) patent search and analysis to the next level for experts and non-experts, alike. The result is the PrecisionAI Patent Explorer.

Working with patents is traditionally about researching first and then analysing the content in detail. The second part is very time-consuming and requires a lot of expertise. Trends or content that has not been scrutinised are usually overlooked, if not completely ignored. The content information level is only accessible through intensive reading and analysing. We address this content level directly and produce contextually relevant knowledge without intensive reading or expert knowledge.

The entire analysis is based on a precisely defined set of patents. In the first step of our approach, these are identified by using EconSight cutting-edge AI retrieval tools. The document set for Chatbot analysis can be enriched with or replaced by other text documents or custom selected patent sets. Subsequently, exactly this set can be analyzed and queried with the help of the most advanced LLMs.

The analysis presented here is based on the technology of biopolymers. The selected set consists of the patent publications in this technology in the last two years. A total of 2375 patents were identified and used for the chatbot analysis. Of course, the definition of the patent set can be freely determined. For example, it is possible to select a specific sub-technology such as Polylactides in the case of Biopolymers, a completely different technology, different time period, or a regional restriction. It is similarly possible to restrict the patent set to a selected competitor portfolio only.

Multiple use cases – from patent searchers to business developers

The entire showcase is based on the example of a biopolymer analysis. Of course, analyses of all patents and all technologies are possible.

Research Development
This example shows the real answer of our system to the question posed.

For Patent Searchers – Find anything.

Try this:

“I search for new biopolymers. List the most often described biopolymers in the selected patents and add a short description. Ignore common petrochemical polymers such as Polyethylene, Polypropylene, PET, Polyamide.”

Business Developers – Find markets.

Try this:

“I look for biopolymer use cases described in patents. Where are these materials used, or used for or applied, like implants, cosmetics, packaging etc.“

Business Development
This example shows the real answer of our system to the question posed.
Asset Manager
This example shows the real answer of our system to the question posed.

For Analysts – Ask anything.

Try this:

“I look for patents describing biopolymers in automotive applications. Cite patents, show patent number and the company behind these.”

For Patent Validity Searchers – Find specifics.

Try this:

“I search for text elements describing features of a fermentation reactor. Cite the text element and patent number.
”

patent validity
This example shows the real answer of our system to the question posed.
patent attorney
This example shows the real answer of our system to the question posed.

For Patent Attorneys – Draft anything.

Try this:

“I need to draft patent applications. Generate an explanative definition of Polyhydroxyalkanoate by only using the selected patents as context. Cite text examples and PNs from the patents.
”

For Patent Searchers – Find anything.

Try this:

“I search for new biopolymers. List the most often described biopolymers in the selected patents and add a short description. Ignore common petrochemical polymers such as Polyethylene, Polypropylene, PET, Polyamide.”

Research Development
This example shows the real answer of our system to the question posed.

Business Developers – Find markets.

Try this:

“I look for biopolymer use cases described in patents. Where are these materials used, or used for or applied, like implants, cosmetics, packaging etc.“

Business Development
This example shows the real answer of our system to the question posed.

For Analysts – Ask anything.

Try this:

“I look for patents describing biopolymers in automotive applications. Cite patents, show patent number and the company behind these.”

Asset Manager
This example shows the real answer of our system to the question posed.

For Patent Validity Searchers – Find specifics.

Try this:

“I search for text elements describing features of a fermentation reactor. Cite the text element and patent number.
”

patent validity
This example shows the real answer of our system to the question posed.

For Patent Attorneys – Draft anything.

Try this:

“I need to draft patent applications. Generate an explanative definition of Polyhydroxyalkanoate by only using the selected patents as context. Cite text examples and PNs from the patents.
”

patent attorney
This example shows the real answer of our system to the question posed.