Font Size: 

Empirical evidence on Intelligent Process Automation (IPA) covering the complete administrative workflow of air freight forwarding small and medium enterprises remains scarce; published studies concentrate on warehousing operations or passenger aviation. The present research addresses this gap through a single-case embedded design conducted within a Romanian forwarding company that processes 454 annual shipment files. A time-motion study was performed on 65 sub-processes (34 import, 31 export), each measured across a minimum of ten live operational files. Activity classification based on Lean principles was applied to classify activities into Value-Added, Business-Non-Value-Added and Non-Value-Added categories. Farinha et al.'s 32-criterion framework, structured into Data, Environment, Human Resources, Governance and Structure, was used as the source model for an author-developed three-tier automation triage instrument. A WhatsApp-based AI assistant (DACS AI), in production since March 2026, was benchmarked against the manual baseline. Measured reductions ranged between 90% and 98% across ten functional modules. Aggregate processing time per file decreased from 162 to 49 minutes for imports (‒70%) and from 194 to 22 minutes for exports (‒89%). The asymmetry between import and export automation potential—respectively 77.8% and 97.9% of total processing time—is attributable to ecosystem barriers rather than technological constraints. A five-year cost-benefit model returned a full-investment payback period of three years and nine months, reduced to fourteen months when the dedicated security-hardened server is excluded from the capital outlay. The methodological framework is transferable to intermodal transport coordination and automotive component logistics.