Advancing API Manufacturing: A Systematic Review of Process Analytical Technology Integrated with Artificial Intelligence and Quality by Digital Design

The pharmaceutical industry is moving toward Pharma 4.0, shifting from traditional batch manufacturing to connected, data-driven systems. Traditional quality checks done after production are slow and often result in material waste. To solve these problems, combining Process Analytical Technology (PAT) and Artificial Intelligence (AI) within a Quality by Digital Design (QbDD) framework offers a major step forward in drug manufacturing. This Systematic Literature Review examines this combination in Active Pharmaceutical Ingredient (API) manufacturing. Following PRISMA 2020 guidelines, a thorough search of ScienceDirect and Taylor & Francis databases found 14 peer-reviewed articles published between 2023 and 2026. The results show that advanced AI models, especially Convolutional Neural Networks (CNN) and YOLOv5, perform much better than traditional statistical methods when handling large and complex data from PAT sensors. These AI models support Real-Time Release Testing (RTRT) by accurately predicting multiple product quality measures at the same time. In addition, QbDD supports eco-friendly “Green-by-Design” approaches by using computer-assisted tools to improve the steps used to make drug compounds. Furthermore, Large Language Models (LLMs) offer Explainable AI (XAI) to turn complex computer predictions into clear, useful insights. Overall, this combination opens the path toward fully automated API manufacturing, though challenges related to computing power and regulatory approval still need to be resolved.