We propose a disaggregated representation of production through an agent-based fund-flow model (NGR-ADAPT) within which inefficiencies, such as factor idleness and production instability, emerge from endogenous frictions. The model incorporates productivity dynamics (learning and depreciation) and is extended with time-saving process innovations. Specifically, we assume that workers possess inherent creativity that flourishes during idle periods. The firm, rather than laying off idle workers, is assumed to exploit this potential by involving them in the innovation process. Results show that a firm's organizational and managerial decisions, the temporal structure of the production system, the speed at which workers learn and forget, and the pace of innovation are critical factors influencing production efficiency in both the short and long run. The co-evolution of production and innovation processes emerges in our model through the two-sided effects of idleness: whereas it drives skill decay it is also a condition for creative thinking that can be leveraged for innovation. In doing so, we question the utilization of labour as an adjustment variable in a productive organisation. The paper concludes by discussing potential solutions to this issue and suggesting avenues for future research.