Unearthing Silent Data : Back to Basics in Process Design

Abstract

Abstract

We live in technology-driven societies, especially shaped by Artificial Intelligence (AI). The performance of AI models depends heavily on data quality—particularly the issue of data bias. Among the many sources of bias, one critical yet often overlooked type is silent data. This typically originates fromo rganizations with low digital maturity, which face challenges in effectively utilizing their data for operations and decision-making. These difficulties begin at the foundational level—data collection and data quality maintenance—hindering their ability to achieve strategic goals. In this work, we proposed comprehensive protocols for the management of the research projects and related administrative operations with a focus on data collection into two databases and on enhancing data quality. We also developed a strategy of maintaining data quality with limited resources, addressing the challenges posed by diverse and scattered data sources.