Insider agents pose a significant threat to consumer security, continuously evolving to bypass detection techniques. This paper addresses the escalating challenges by proposing adaptive solutions that integrate AI capabilities, particularly leveraging natural language processing (NLP). While conventional parameters associated with email attacks can be easily altered, the inherent sentiment of emails often remains consistent. The research focuses on the synergy of NLP techniques and established AI algorithms to introduce a resilient detection method for insider threat emails. Recognizing that the AI classifiers performance depends on meticulous selection of hyperparameters, this research delves into various contemporary optimizers. Moreover, a tailored solution is proposed, specifically designed to meet the unique requirements of insider threat detection. Therefore, the aim is to enhance the effectiveness of email security measures in the face of constantly evolving tactics.