Abstract: The proliferation of digital commerce has led to an unprecedented rise in corporate domain impersonation and brand abuse, where malicious actors leverage typo squatting, homograph attacks, and sophisticated social engineering to deceive consumers. Traditional detection mechanisms, often reactive and rule-based, struggle to keep pace with the dynamic nature of these threats. This paper proposes a novel Agentic AI-driven framework for the autonomous detection and mitigation of brand-related cyber threats. By employing a multi-agent system (MAS) where specialized agents handle domain scouting, visual brand analysis, and semantic content evaluation, the proposed system achieves high-fidelity detection with minimal human intervention. Our framework integrates advanced machine learning models, including Siamese Neural Networks for logo similarity and Transformer-based models for lexical analysis of URLs. Experimental results demonstrate that the agentic approach significantly reduces the time-to-detection and improves the accuracy of identifying sophisticated impersonation attempts compared to conventional centralized ML models. This research contributes a scalable, autonomous architecture for digital risk protection, offering a proactive defense mechanism for corporate brand integrity.
Keywords: Agentic AI, Machine Learning, Domain Impersonation, Brand Protection, Cybersecurity, Multi-Agent Systems, Phishing Detection.
Title: Machine Learning-Based Detection of Corporate Domain Impersonation and Brand Abuse: An Agentic AI Approach
Author: Muhammad Zaki Ahmad, Asif Ali Khan
International Journal of Computer Science and Information Technology Research
ISSN 2348-1196 (print), ISSN 2348-120X (online)
Vol. 14, Issue 2, April 2026 - June 2026
Page No: 253-256
Research Publish Journals
Website: www.researchpublish.com
Published Date: 16-June-2026
