‘Lean first, then automate’: a framework for process improvement in pure service companies. A case study

TLDR

Lean Management, traditionally rooted in manufacturing, has only recently been adopted in services, yet the role of automation and information technology in lean processes remains controversial and lacks robust empirical evidence, especially in pure service contexts such as banking. The study seeks to define a methodology for streamlining and automating processes in pure service‑providing firms. The authors conducted a case study, mapping processes to identify waste, streamlining them first, and then automating only customer‑recognized value‑added activities. The investigation yielded a framework showing that automating unstreamlined processes can slow flow and raise errors, thereby contributing to the debate on transferring manufacturing methodologies to the service sector.

Abstract

Born in manufacturing environment, only recently Lean Management has been implemented in services. In particular, the role of automation and information technology in leaning processes is controversial and literature still lacks a strong empirical evidence to clarify how Lean Management can be applied in a pure service context, such as banking services, where there is an intensive use of automation and Information Systems. This study aims to define a methodology to streamline and automate processes in pure service-providing companies. To achieve the research aim, we conducted a case study. Based on the empirical investigation, a framework was developed. We found out that the automation of a process that is not streamlined can generate problems that can slow down the flow and increase errors. A process should first be mapped to highlight waste. Only when the new process is streamlined, it can be automated. In doing so, the new process will automate only value-added activities recognised by the customers. This research contributes to the debate about the possibility to transfer manufacturing methodologies in the service sector.