Exploring the Feasibility and Potential Usefulness of AI-Driven Automation in EPCM Procurement and Project Services : A Qualitative Study for COWI, Sweden

Sammanfattning

This study investigates how EPCM procurement and project service professionalsperceive the feasibility and potential usefulness of AI-driven automation in supporting complexworkflow environments. EPCM procurement processes are characterized by highinterdependence across engineering, planning, cost control, document management, andsupplier coordination, yet many activities remain dependent on manual effort despite thepresence of digital procurement systems. This creates coordination inefficiencies, fragmentedinformation flows, and increased administrative workload. The study adopts a qualitative casestudy design within an EPCM organization and draws on semi-structured interviews withprofessionals across procurement and project service roles. Data were analysed using thematicanalysis guided by the Task-Technology Fit framework. The findings show that procurementworkflows are highly iterative and coordination-intensive, with significant reliance on manualcommunication, spreadsheet-based processes, and fragmented systems. Participants perceiveAI-driven automation as most useful for targeted support functions such as documentcompleteness checks, reminders, bid comparison assistance, expediting support, andknowledge retrieval from historical data. However, AI is not viewed as a substitute for humanjudgement, particularly in negotiation, contractual decisions, and stakeholder coordination. Thestudy concludes that the value of AI in EPCM procurement depends on its alignment with taskrequirements, system integration, and preservation of human oversight. It contributes tounderstanding AI feasibility in project-based environments and provides practical insights forincremental, workflow-aligned AI implementation in EPCM organizations.