Sustaining optimal task engagement is becoming vital in smart factories, where manufacturing operators' roles are increasingly shifting from hands-on machinery tasks to supervising complex automated systems. However, because monitoring tasks are inherently less engaging than manual operation tasks, operators may have a growing difficulty in keeping the optimal levels of engagement required to detect system errors in highly automated environments. Addressing this issue, we created an adaptive task engagement feedback system designed to enhance manufacturing operators’ engagement while working with highly automated systems. Utilizing real-time acceleration, heart rate, and respiration rate data, our system provides an intuitive visual representation of an operator's engagement level through a color gradient, ensuring operators can stay informed of their engagement levels in real-time and make prompt adjustments if required. This article elaborates on the six-step process that guided the development of this adaptive feedback system. We developed a task engagement index by leveraging the physiological distinctions between more and less engaging manufacturing scenarios and using automation to induce lower engagement. This index demonstrates a prediction accuracy rate of 80.95 % for engagement levels, as demonstrated by a logistic regression model employing leave-one-out cross-validation. The implications of deploying this adaptive system include enhanced operator engagement, higher productivity and improved safety measures.