Ask your Transcript: LLM Driven Insights for Academic Advising

Recent advances in Artificial Intelligence have witnessed an unprecedented penetration of AI tools into our day-to-day life like never before. Especially, a generative AI tool such as ChatGPT that is driven by Large Language Model (LLM), has been proven successful and effectively adapted across many domains such as business, health care, education and much more. The ability of LLM models in understating natural language text and generating human-fashioned responses is very astonishing that paved the way for wide usage of such tools among the population. The students in education domain face many hurdles in their academic advising due to the lack of adequate upfront information available to them. In this work, we explore a more viable and feasible solution to alleviate this problem in term of LLM driven tool and provide a student community with concise and informative responses by retrieving the information given in their academic transcripts. The proposed work crafts a possible list of questions that student community seeks answers based on their past academic performance and experience in navigating the transcripts for their information needs. The tool then injects these questions as prompts to the RAG models to generate relevant automated responses in natural language as if generated by human adviser. The proposed tool may compliment both the student and human advisers in term of equipping them with academic equity and comfort.