Generative AI Patent Prosecution Tool with Prompt Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing generative AI programs in patent prosecution lack the ability to retain and organize series of prompts and analysis documents, leading to inefficiencies in document management and analysis support.
Innovation Solution
A system integrating a patent prosecution tool with a generative artificial intelligence program, which maintains a database of language learning models and prompts, allows users to input matters associated with patent documents, select models, retrieve and load documents, and initiate scripts for analysis document production.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple reference documents are loaded into the generative AI program, then the quality and support of analysis documents improve, but the complexity of document management increases
Solution Approach 1:
The patent introduces an intermediary system that acts as a bridge between the generative AI program and multiple reference documents. This intermediary automatically manages document loading, retrieval, and organization, allowing the AI to access comprehensive reference materials without exposing users to the underlying management complexity. The intermediary handles document selection, loading, and coordination transparently.
Solution Approach 2:
The system implements self-service capabilities where the generative AI program automatically retrieves and loads necessary reference documents based on analysis requirements, without requiring manual user intervention. The system autonomously manages document corpora, retrieves relevant materials, and organizes them for AI processing, reducing user burden while maintaining high analysis quality.
2Productivity
If prompts and analysis documents are retained and organized, then future reference and recreation efficiency improve, but the system complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-organizing and storing prompts and analysis documents in a structured repository before they are needed. The system automatically saves generated analysis documents and associated prompts for future retrieval, eliminating the need for users to manually recreate or locate them later. This preliminary organization enables rapid reference and recreation efficiency.
Solution Approach 2:
The system implements copying by creating and storing digital replicas of prompts and analysis documents in a centralized repository. These copied documents can be retrieved and reused without requiring access to original files or manual recreation, enabling efficient reference and maintaining consistency across multiple uses while simplifying user interactions.
3Reliability
If the system automatically loads reference documents, then analysis document support improves, but the automation complexity increases
Solution Approach 1:
The patent introduces an intermediary layer that automatically manages the complex task of loading reference documents. This intermediary handles document retrieval, validation, and loading into the AI program transparently, providing reliable document support while hiding the automation complexity from users. The intermediary acts as an automated document management system that users interact with simply.
Solution Approach 2:
The system implements self-service automation where the generative AI program autonomously identifies, retrieves, and loads necessary reference documents from corpora based on analysis requirements. This automatic document loading process occurs without user intervention, providing comprehensive document support while the system independently manages the complexity of coordination and execution.
Data Source
AI summary
Various embodiments disclosed relate to a method of producing an analysis document using generative artificial intelligence, the method comprising: maintaining a database containing multiple language learning models and a collection of prompts; generating a patent prosecution tool user interface integrated with a generative artificial intelligence program; receiving a user input comprising a matter associated with at least one matter dataset, where the matter dataset includes a plurality of patent documents associated with the matter; selecting a language learning model from the database; retrieving a plurality of patent documents for display on the patent prosecution tool user interface; loading at least one of the plurality of patent documents into the artificial intelligence program; displaying more than one prompt from the collection of prompts; initiating a script in the generative artificial intelligence program based on a user selected prompt and the at least one patent document; analyzing the patent document according to the script to produce a preliminary analysis; producing the analysis document according to the preliminary analysis; saving the analysis document.