CSJM University's AI Agent Compiles Historical Records for Convocation
CSJM University AI Agent Compiles Historical Records for Convocation

Kanpur: The CSJM University Kanpur's indigenous OCR (Optical Character Recognition) and RAG (Retrieval Augmented Generation)-based AI agent, developed by the university's CAM AI team, has successfully completed the task of compiling the university's achievements and historical records.

Objective and Methodology

With the objective of preparing factual and research-based material for the convocation ceremony booklet, an AI agent was tasked with studying and analyzing over 2,100 newspaper reports available on the university's website. The agent read the scanned Hindi news clippings, identified significant achievements, programs, innovations, and institutional activities contained therein, and prepared detailed reports by organizing them thematically.

Processing and Results

According to the media department of CSJM, the system developed by the CAM AI team processed a vast amount of available archival material in a short span of time, transforming it into useful and context-based information. The reports prepared as a result of this work were submitted to the Convocation Committee, thereby providing a robust and authentic basis for the creation of the booklet.

Wide Pickt banner — collaborative shopping lists app for Telegram, phone mockup with grocery list

Capabilities and Recognition

A notable achievement of the project was that the AI agent demonstrated the capability to comprehend scanned news articles available in Hindi and to present meaningful and contextually relevant analyses based on them, the media department stated. Prof. Vinay Kumar Pathak, Vice-Chancellor of CSJM, stated that the indigenous AI systems developed by the university's CAM AI team were not only capable of analyzing complex Hindi documents and records but could also make effective contributions to critical tasks such as institutional documentation, knowledge management, and the scholarly compilation of historical archives.

Pickt after-article banner — collaborative shopping lists app with family illustration