Assisted Security Investigations Using Cognitive Computing

The purpose of this research is to illustrate the application of cognitive computing and machine learning concepts through the building and training of a chatbot that simulates human conversation for cybersecurity investigation scenarios. The SOC chatbot will offer best-practice advisory dialogue to security analysts as they proceed through security incident investigations, thus simulating technical mentorship. As a security analyst progresses through various investigations, they will become more practiced in the recommended and appropriate workflows, gain investigative tool proficiency, and become more confident in handling standalone investigations. The SOC chatbot will serve as a training tool for less experienced analysts and afford more time to upper-tier analysts to respond to escalated security incidents, as they will no longer need to walk through incidents alongside junior analysts. Security analysts serving in a tier 1 SOC role are ideal end-users of the SOC chatbot. As the first line of defense, their primary function is to address SIEM events. They are familiar with basic security concepts, incident ticketing systems, and hold the appropriate level of access for data gathering and external research.