AI & E-Commerce : Assessing the Capability of AI-Driven Intelligent Document Processing in the Inbound B2B Order Processing

Sammanfattning

This study investigates whether AI-driven Intelligent Document Processing (IDP) is perceived by participants as a potential improvement to inbound B2B order processing and the organisational challenges associated with its adoption. The study is an interpretive qualitative case study at Seco Tools, guided by the Technology-Organisation-Environment (TOE) framework. Data were collected through seven semi-structured interview sessions with eight participants from IT and business-side departments. Deductive qualitative content analysis was performed using MAXQDA software. The findings indicate the current OCR system has limited capacity to interpret the meaning of order documents. Participants perceived AI-driven IDP as only conditionally capable of improving semantic accuracy through adaptive learning, contextual field interpretation, and cross-format adaptability. Issues regarding the non-deterministic nature of AI, hallucination risk, and calibration of confidence thresholds were raised. In addition, the study identified employee resistance, uneven skill distribution, weak IT-business collaboration, unclear post-deployment AI ownership, and insufficient lifecycle management as the main organisational challenges for AI adoption. Environmental constraints as external factors for AI adoption include GDPR compliance, corporate AI governance, customer document variability, and next-day delivery service level agreement (SLA) pressure. Thus, the study examines the transition from OCR to AI-driven IDP as a socio-technical transformation, not a simple technology upgrade.