Purpose – This paper aims to highlight the complex nature of automated guided vehicle (AGV) simulation model building, and especially how system modelling details affect the end results. This is an important issue in all of the transportation simulation systems, since they are service-based by their nature, and additional inefficiencies create unanticipated performance downgrading.

Design/methodology/approach – This paper uses a simulation approach, and simulated systems are based on a real-life case study and on well accepted hypothetical simulation example.

Findings – Simulation system boundaries are often neglected in the model building, and especially interface to inbound (and possibly outbound) material flow should be considered carefully; based on these research results, AGV investments are seen in an entirely different light, as system boundary is enlarged to contain more realistically interacting elements. Similar system boundary issues were found from the case study: interface with overhead gantry did not provide near optimal performance. The case study also revealed that high speed of AGVs is not necessarily worth additional investment; constraints exist in safety, acceleration and ability to turn in corners.

Research limitations/implications – The findings are based on the simulation work and, to see the real implications, real-life implementations on policy level are needed.

Practical implications – Results of this research provide more insights for manufacturing unit investments, and especially in the scope of automated transportation system use. Also changes in manufacturing flow management issues, after investing in, for example, AGV systems, are different from in less-automated manufacturing units.

Originality/value – This research work provides more insights to simulation research work, especially from the perspective of transportation systems. Also implications arising from case study are unique as being compared to previous research in the field.