Administrative processes in internal logistics are dominated by repetitive manual work, resulting in delays, fragmented information flow, and increased risks of errors. AI adoption is increasing in supply chain and manufacturing; however, its application to administrative tasks within internal logistics remains limited and is not well understood. This gap is addressed by developing a conceptual framework for the adoption of AI-administrative automation and validating it empirically in the South African manufacturing context. This is materialised by identifying the most suitable administrative tasks, assessing the effects on efficiency, accuracy, and speed, the challenges and risks of integrating AI automation into existing workflows, and the measurement and quantification of success. A mixed-methods design, consisting of a quantitative survey and qualitative semi-structured interviews, was adopted and grounded in a pragmatic approach. The findings indicate the suitability of structured, rules-based administrative tasks to automation. Concerns relating to trust, oversight, and accountability are also identified. The study concludes that AI-driven administrative automation constitutes a socio-organisational transformation shaped by behavioural, organisational, economic, and governance conditions, and proposes a framework to support responsible implementation.

Keywords

Internal logistics, Artificial Intelligence, AI-driven administration automation, AI Adoption Framework, South Africa