[KiNT Series Vol. 5] Release of 'Order Data Quality Check Function' to Automatically Detect Address Mismatches and Prevent EC Delivery Issues

GoQSystem Inc., which operates the cloud-based order management system 'GoQSystem', has launched the 'Order Data Quality Check Function' as the 5th installment of its AI business efficiency series 'KiNT', which detects address errors in order data before shipping. In EC site orders, address input errors lead to delivery delays and failures, causing significant losses. Previously, this relied on manual visual checks by operators, but this function was developed to overcome the 'limits of visual inspection'. This function implements a hybrid inspection flow that combines rule-based inspection with AI-based contextual completion. It quickly verifies the consistency between postal codes and addresses, and for variations in notation, AI reads and complements them from the context. Additionally, through address normalization, it automatically absorbs variations such as full-width/half-width characters, Chinese numerals, and old fonts. It also supports Japanese conversion of English addresses and provides correction suggestions with confidence levels, supporting rapid responses by operators. Even if an AI service failure occurs, the 'AI fallback design' ensures business continuity by providing standard proposals based on rule-based inspection results. Note that it does not automatically rewrite order data; corrections are executed after the operator confirms the content. The use of generative AI is limited to address checking and is not used for model training.