Joe Sandbox AI is a GenAI-based detection tool designed to identify phishing attempts and malicious content across a range of digital inputs. It analyzes URLs, webpages, emails (EML and MSG formats), JavaScript, DOM trees, screenshots, OCR data, and documents including PDFs, DOCX, XLSX, and SVG files. Rather than relying on static reference lists, brand databases, or predefined signature patterns, the product uses multiple GenAI models — each specialized for different aspects of phishing analysis — to assess the intent and authenticity of content. This approach allows it to detect previously unseen or emerging threats that may bypass conventional methods. Key capabilities include: - Deep contextual analysis of email content to detect phishing, Business Email Compromise (BEC), and social engineering attacks by evaluating intent, tone, and linguistic cues. - Automated navigation and interaction with phishing webpages, including bypassing CAPTCHAs and QR codes, and submitting simulated user data to landing page forms. - AI-generated summaries covering behavioral insights, threat indicators, and contextual findings to support incident response. - Detection of malicious SVG files with embedded scripts or redirects. - Verification of suspicious PE Authenticode signatures to uncover forged or tampered certificates. The product integrates with Joe Sandbox Cloud and is available via trial upon request from Joe Security.