Streamlining the E-bidding process through automated compliance check of procurement advertisements

Abstract

The construction industry is an essential contributor to the economy in many countries. However, due to its unique nature, it has been perceived to lag behind in innovation. Bidding is a crucial component of construction projects that involves multiple stakeholders and generates significant paperwork. To address the challenges associated with paper-based bidding, electronic bidding (ebidding) has been proposed as a solution. E-bidding is less costly and time-consuming, making it more efficient and effective. However, compliance checking of bid documents during the ebidding process is currently performed manually, increasing the risk of errors and requiring significant manpower. This research proposes a framework for automated compliance verification of bid documents with predefined standards, specifically the regulations of the Public Procurement Regulatory Authority (PPRA) in Pakistan. By automating compliance verification, the e-bidding process can operate more efficiently, resulting in improved project outcomes and overall performance of the construction industry. The proposed framework also sheds light on the potential of Natural Language Processing (NLP) algorithms in automating compliance verification, which could contribute significantly to the advancement of the construction industry and the field of machine learning for compliance verification. In validating the performance of the automatic model built for this study, we discovered that its precision and recall were both 82% in comparison to manual review. This study is significant because a model capable of conducting PPRA compliance checks has been developed