Author : Ramesh Bishukarma

Cloud computing (CC), particularly through the Software as a Service (SaaS) model, has revolutionized a way organization manage their IT infrastructure. Nonetheless, SaaS has numerous benefits including cost, scalability, and accessibility, and for the same reason, SaaS creates a number of risks. Another important area is anomaly detection on SaaS platforms, so that the identification of abnormal activity may suggest a violation of security, operational deviance, or flawed performance. This paper proposes a novel AI-enabled adaptive anomaly detection method that facilitates the protection of SaaS solutions. Mirroring the ML techniques such as supervised, unsupervised and semi-supervised learning, it provides real-time surveillance of an organization’s networks for data breaches, insider threats, malware and denial of service (DoS) attacks. Incorporation of adaptive AI methods enables the framework to learn as well as adapt in the detection of new forms of security threats and hence keep SaaS environments secure. This paper also presents various types of anomalies in SaaS environments, and AI approaches to adaptive SaaS anomaly detection, with focuses on possibilities of these techniques in preserving cloud-based services.

Keywords: SaaS Security, Cloud Computing, Anomaly Detection, AI-Driven Security, Machine Learning, Deep Learning, Cybersecurity, Adaptive Detection.