Supplier selection has significant impact on entire supply chain and there are various methods being derived to generate supplier scoring for supplier selection. Most of this existing research mainly focus on use of decision support system using fuzzy logic or heuristic and meta-heuristic optimization techniques. Available literature is lagging behind in using more advanced methods like AI, machine learning and deep learning or Generative AI. In this paper, we use the publicly available secondary data about suppliers from social e-commerce websites. Unstructured reviews data is processed using large language models (LLM) to generate supplier sentiment score (SSS) and other extracted data variables are processed to generate supplier reputation scores (SRS) and supplier performance score (SPS) of the supplier. Finally combining these three scores final supplier score is derived which can be useful method for supplier selection. Our findings indicate that, with advanced mechanism like AI and LLM, e-commerce data can be processed to build a supplier scoring mechanism. This framework assists organizations with supplier selection and thereby attain all the benefits.