Generative Ai for Document Analysis using Minsky™ Integration with ChatGPT
Background/Problem Statement:
Typically, all organisations are constantly faced with a need to analyse a large number of documents such as Legal Agreements, Contracts, etc. This requires a large amount of manual processing and extracting the salient features of a document which effects the efficiency and productivity of any organization. The objective of the project is to use Artificial Intelligence/ Generative Ai to automatically extract and summarize the salient feature of any documents (PDF, Word etc) based on user defined questions.
Solution Overview:
The solution design was based on using Minsky as the front end for user inputs and database while using the ChatGPT engine for Generative Ai. The Solution was achieved by following these steps at a high level:
- Browse, select and upload the document (that needs to be analysed) from a local machine. This document is also saved in the Minsky DB. This step is accomplished using the Minsky frontend.
- User then enters question for performing document analysis and hits "Submit".
- At this stage, the request (Document & Questions) is sent to the ChatGPT Engine using an appropriate API.
- The output/Results are then returned to the Minsky Environment.
- Results are displayed in MinskyTM and can be save for future reference.
Key Benefits:
- Increase in efficiency and productivity in any organization that required large Document analysis.
- Helps businesses to process more documents at greater speeds and at a fraction of manual labor costs.
- Gives a quick overview of multi pages documents in seconds.
- User-Friendly cloud-based AI platform.
- Scalable across various domains/data.
- Ability to fine tune or optimize the models by trying different algorithms / prediction attributes.
- Easy integration with other third-party Data Warehouses such as Data-Bricks, Snow-Flake, etc.
