Agentic MailBot: Autonomous Multi-Agent Email Response System with Contextual Intelligence

In this paper, Agentic MailBot was introduced as an au- tonomous multi-agent email response framework combining retrieval-augmented generation, SQL reasoning, and modern orchestration frameworks. The system achieved: •        91.2% classification accuracy •        88.4% end-to-end pipeline success •        strong response quality across knowledge and transac- tional queries The evaluation demonstrates that the proposed architecture outperforms conventional rule-based approaches by enabling adaptive contextual reasoning and dynamic information access. The combination of: •        Google Gemini 2.5 Pro •        LangChain •        LangGraph •       FAISS •       ReAct agents •       SQL toolkit integration represents a practical advancement in intelligent customer communication automation. Businesses increasingly require scalable automated commu- nication systems, and Agentic MailBot demonstrates a viable architecture for fully autonomous contextual email processing.

[1]          Rule-based email classification systems, Philosophical Transactions of the Royal Society. [2]          Machine learning approaches to email categorization, IEEE Transactions on Knowledge and Data Engineering, 2003. [3]          J. Devlin, M. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of Deep Bidirectional Transformers for Lan- guage Understanding,” NAACL-HLT, 2019. [4]          T. Brown et al., “Language Models are Few-Shot Learn- ers,” NeurIPS, 2020. [5]          S. Yao et al., “ReAct: Synergizing Reasoning and Acting in Language Models,” arXiv, 2022. [6]          P. Lewis et al., “Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks,” NeurIPS, 2020. [7]          A. Vaswani et al., “Attention Is All You Need,” NeurIPS, 2017. [8]          F. Sebastiani, “Machine Learning in Automated Text Categorization,” ACM Computing Surveys, 2002. [9]          J. Johnson, M. Douze, and H. Jegou, “Billion-Scale Similarity Search with GPUs,” IEEE Transactions on Big Data, 2021. [10]     H. Chase, “LangChain: Building Applications with LLMs Through Composability,” 2022.