Many studies from around the world enhance awareness of medication errors in hospitals. A French national study on Adverse Events (AE) revealed that 60 000 to 130 000 AE per year appear to be consequences of medication errors, and 15 000 to 60 000 AE are preventable. Furthermore, hospital spending represents 37% of the French healthcare costs. Enhancing medication system quality and efficiency by reducing medication errors and costs are then considered as important public health issues. Medication management is a complex socio-technical system involving multiple people and numerous steps. Automated dispensing systems have been reported to be key components of strategies to prevent medication errors and create a better workflow. Few hospital pharmacies are implementing automated nominative dispensing robot that increase secure treatment preparation as well as dispensing safety and performance. Automated storage systems and medication delivery devices also reduce hidden costs related to pharmaceutical products stock control and remove professional agent time from non-value-added tasks for a more efficient redeployment tasks. Nevertheless, our results from auditing hospital pharmacies that are investing in automated systems revealed risks and hidden costs related to reception station, dispensing and care unit orders that should be settled by computerization and organizational actions. Our study shows that hospital pharmacy organizations require the implementation of specific actions settling software interface problems and enhancing coordination skills among health care actors in order to secure medication delivery system and prevent from hidden costs related to organizational factors. Then, automation does not seem a self-sustaining option to handle pharmacy’s activity dysfunctions. Many other parameters intervene. This paper presents a pharmacy’s automation project. It takes example at a French Hospital Center in Lens in the north of France (CHL). We present a human centered automation approach that includes an identification of tasks eligible for automation and tasks that need to be supervised, controlled or operated by a human agent.