AI-Assisted Passenger Goods Inspection at Ngurah Rai International Airport - A Simplified Design Science Research Study

Ngurah Rai International Airport in Bali processes over six million international passengers annually, generating significant demand for seamless customs operations. To address mounting service and risk management challenges, Ngurah Rai Customs has introduced non-intrusive and digitally integrated inspection measures. Manual procedures, however, still add documentation burdens, limiting officer efficiency and service quality. Although AI technologies have improved customs imaging and risk profiling, research on automating customs documentation workflows with AI remains limited, particularly in resource-constrained environments. This study uses the Design Science Research (DSR) approach to develop and evaluate an AI-enabled Telegram bot. The bot automates customs inspection documentation using speech-to-text and rule-based classification. The system leverages OpenAI’s Whisper API for speech recognition and includes a PostgreSQL database backend. Leveraging officers’ existing familiarity with Telegram, the study aims to lower barriers to low-cost automation. Impact was evaluated through semi-structured interviews with 25 customs officers and three expert stakeholders in August 2025. Thematic analysis identified: (1) a 93% reduction in documentation time; (2) platform familiarity benefits; (3) persistent trust and accountability concerns; and (4) strong consensus for innovation despite resource limitations. Results show that AI-assisted workflow automation can expand operational capacity at Ngurah Rai Customs and preserve officer authority. However, effective adoption requires addressing challenges related to integration, data sovereignty, and accountability. This work offers practical guidance for low-cost AI deployment in customs administration. It also advances the theory of AI adoption in developing countries.