Article’s

The increasing sophistication of social engineering attacks and phone-based fraud schemes has created significant security challenges within modern call center operations. Despite routine training programs, many organizations remain vulnerable due to static learning modules that fail to replicate the evolving tactics used by real-world scammers. The problem arises from limited scenario variability, lack of real-time adaptability, and absence of measurable performance assessment, which often leaves agents underprepared to detect complex scam attempts involving impersonation, urgency tactics, and psychological manipulation. This paper presents the design and implementation of a Real-Time Scam Script Generator, an intelligent simulation-based training framework developed to enhance defensive preparedness. The proposed system employs scam pattern databases, behavioral modeling techniques, and rule-based conversational logic to dynamically generate realistic fraud scenarios. It supports multiple scam categories including phishing, impersonation fraud, and financial exploitation schemes. A dedicated performance evaluation module measures metrics such as detection accuracy, response time, protocol compliance, and false positive rate to assess agent effectiveness. Experimental results indicate improved threat recognition capability and faster decision-making compared to traditional training methods. The proposed solution provides a scalable and automated approach to strengthen cybersecurity resilience and reduce fraud susceptibility in call center environments.