ARCHIVES

The rapid expansion of global online shopping platforms has fundamentally shifted consumer behavior, leading to a critical need for secure and automated delivery verification mechanisms to handle increased parcel volumes. Conventional authentication approaches, such as manual One-Time Password (OTP) sharing between customers and delivery personnel, often suffer from significant operational inefficiencies and security vulnerabilities, including human error and unauthorized access. To address these limitations, this paper introduces Bot2Door, an intelligent delivery authentication system powered by artificial intelligence that integrates hybrid intent recognition with context- aware OTP validation. The proposed framework utilizes a sophisticated dual-layered approach, combining rule-based filtering with transformer-based natural language processing (NLP) techniques to accurately interpret and process delivery-related conversations in real time. Furthermore, a context-driven OTP retrieval mechanism is designed to handle multiple incoming messages by applying temporal constraints and semantic similarity analysis, ensuring the correct selection of credentials for specific delivery events. A comprehensive multi-factor authentication strategy is incorporated into the system to evaluate delivery legitimacy based on several parameters, including intent confidence scores and OTP matching accuracy. Experimental evaluations demonstrate that Bot2Door significantly enhances system reliability, minimizes the risk of incorrect OTP usage, and improves overall delivery security. The resulting solution is highly scalable and optimized for deployment in real-world edge-based environments where low-latency and robustness are essential.