Open Research and Creative Activities Forum

Business card digitization tools are widely available, yet many solutions remain focused on optical character recognition and contact storage rather than measurable networking outcomes. This paper develops an academic and market-oriented framework for an AI-based multilingual business card intelligence system designed to convert raw card images into structured contacts, prioritized action lists, and AI-assisted follow-up content. The study is motivated by a practical gap between data extraction performance and post-event relationship conversion. Existing literature supports the importance of human-in-the-loop information extraction workflows, privacy-aware data handling, and AI governance, but limited work links these domains to networking operations and event productivity. Using the current prototype design as an empirical design reference, this paper documents the system architecture, key parameters, and a quantitative scoring model for contact prioritization. The system integrates cloud functions, schema validation, multilingual extraction pipelines, confidence-based review thresholds, and event-level batch processing. A weighted networking score model is proposed to rank contacts using role relevance, interest alignment, urgency, strategic value, relationship strength, recency, and profile completeness. The paper also outlines measurable performance indicators, including extraction accuracy, review rate, duplicate reduction, follow-up speed, and conversion uplift. From a commercial perspective, the framework positions system as a workflow intelligence product rather than a scanning utility, with a differentiated value proposition in multilingual precision, event intelligence, and trust-oriented architecture. The contribution is a research-ready foundation for future field experiments and commercialization studies that evaluate whether AI-assisted contact intelligence improves networking effectiveness at scale.