My PhD research project originated from the following research question: “How can we translate the well-known concept of reconfigurable and autonomous manufacturing, which has been explored in smart manufacturing over the last two decades, into the specific domain of Visual Inspection Systems (VIS)?”. In this context, the ultimate goal was to propose a new technological framework and open new business model avenues, such as configure-to-order, rather than the classical engineer-to-order solutions associated with monolithic and rigid VIS. These traditional approaches often necessitate high capital investments and substantial human intervention from both hardware and software perspectives to reprogram the VIS. The primary objective of my Ph.D career in its initial years was to study the current state-of-the-art of VIS, identify gaps in the existing landscape (e.g., outdated and highly customized yet rigid solutions not adaptable to variable production and more flexible demands), and leverage advancements in Industry 4.0 technology to establish a new wave of flexible VIS capable of handling variable production with minimal setup time and modular architecture, both in hardware and software, ultimately achieving reconfigurable VIS. As a further ambitious step during the final year, I endeavored to define a framework for more advanced VIS, termed autonomous, which retain and inherit the aforementioned features of flexibility and reconfigurability but are also capable of reasoning and making decisions autonomously regarding the selection of the best vision inspection algorithm and parameters, starting from annotated CAD models. This entails integrating CAD model design and inspection concurrently using the model-based definition (MBD) philosophy from the early product development stages. Concurrently, in line with the human-centricity emphasized in the latest Industry 5.0 concepts, I aim to incorporate concepts related to user graphical interface ergonomics and human decision-making or higher-value tasks into the loop of autonomous VIS. This entails relieving humans from low-level coding tasks or manual handling of hardware components, now controlled numerically (e.g., cameras, lights), allowing operators to focus more on process optimization and supervision. During my last year of the Ph.D, I developed a prototypical system to demonstrate feasibility for SMEs and to advance this promising and novel topic in the industrial landscape during my Erasmus+ traineeship period abroad. Currently, I am undertaking a stage with an SME in the local (Tuscany) manufacturing area to implement some of the topics theorized in the latest frameworks, primarily focusing on human interaction.