.png&w=256&q=75)
2
2
Looking for experience!
.png&w=828&q=75)
FraudLens AI is an AI-powered security platform designed to address two rapidly converging threats in modern digital trading environments: behavioural fraud and expanding technical attack surfaces. As AI-generated phishing and highly personalised scams increase, attackers increasingly exploit user trust rather than purely technical vulnerabilities. The platform combines real-time fraud detection with an AI-driven attack surface pentester. FraudLens AI analyses suspicious emails, chat transcripts, documents, and AI-generated content to detect behavioural manipulation, social engineering, and fraud indicators. At the same time, the AI Pentester module simulates active penetration testing by reasoning about domains and URLs, modelling attacker behaviour, identifying potential attack vectors, and prioritising risks without performing real exploitation. Built using Groq’s high-speed AI inference, the system delivers low-latency analysis, clear risk scoring, and human-readable security reports. FraudLens AI helps organisations reduce fraud, improve security visibility, and scale protection across both human and technical layers in a continuously evolving threat landscape.
7 Feb 2026
As cyberattacks rise, SOC teams face alert fatigue and slow manual triage. Our AI-powered Tier-0 SOC Analyst workflow, built on Opus, automates intake, analysis, risk scoring, and reporting to cut false positives and improve efficiency. We built a simple UI where analysts can upload emails, syslogs, SIEM logs, file metadata, text, and URLs in formats like TXT, PDF, CSV, or direct links. The UI securely sends all inputs to Opus via API, ensuring wide coverage and strong integration. Opus extracts raw content and normalises everything into a unified JSON structure with reliable validation and retry logic. Large-scale IoC extraction identifies IPs, domains, URLs, hashes, and email IDs. Since external services were unreliable, we built a RAG module to classify suspicious or malicious patterns. All IoCs then go through an enrichment stage that adds context, reputation, threat tags, domain age, and confidence, producing a consistent enriched dataset. Two decision nodes handle triage: the first checks whether an IoC is malicious. Clean IoCs go straight to output for automatic report generation, while malicious ones are severity-scored and reviewed by AI. The second node checks if the severity is equal to or greater than 70. Lower scores generate tickets automatically; higher scores trigger human review before finalisation. AI review occurs at key stages—normalised data, enriched IoCs, severity, and final ticket—while human review is reserved for high-risk cases. The workflow ends by generating a report and audit trail, displayed on the UI for full visibility. The system aligns with the UAE and GCC visions in the Middle East by demonstrating a secure, efficient, and scalable AI-driven cybersecurity model. Note: Add our RAG PDF to the RAG Extraction input: https://drive.google.com/file/d/1HYgv4h4W0oWzx2wcMFyGerX1e-4Ba-X5/view Also generate your own PDF API key and add it to the Bearer Auth field in the Report and Audit Artefact Generation node.
19 Nov 2025
.png&w=640&q=75)
.png&w=640&q=75)
.png&w=640&q=75)
