Built an AI-assisted stock trading signal system using ChatGPT and UI.Vision automation. The system monitored recent market news and triggered periodic AI checks during trading hours. For each check, it evaluated whether there was a valid short-term trading opportunity, filtered out weak setups, avoided repeated signals, and returned either a trade action or `NOTRADE`. Key features: * Automated UI.Vision trigger running every few minutes * News-based trade screening using recent market events * Signal filtering based on timing, risk, and whether the price move had already happened * Structured trade outputs including action, ticker, confidence, risk, reward, urgency, and forecast * Open-position tracking to avoid duplicate or conflicting recommendations * Tested through a Plus500 demo trading environment This project helped me explore practical AI automation, prompt design, financial decision support, and building rules-based workflows around generative AI.