Researchers developed Amylo-Detect, a machine learning model that analyzes electronic health records to identify patients at risk for cardiac amyloidosis, a progressive heart disease that's frequently diagnosed too late. Trained on 11,616 patients across multiple hospitals, the system achieved 91% accuracy in detecting high-grade cardiac amyloid uptake and identified 29% of cases that clinicians had missed during routine evaluation. The model uses 50 standard health parameters and demonstrated consistent performance even when key predictors were unavailable. This represents a significant advance in addressing one of cardiology's most challenging diagnostic problems. Cardiac amyloidosis affects thousands annually but remains underdiagnosed due to nonspecific early symptoms, delaying life-saving treatments like tafamidis. While promising, this preprint awaits peer review and the real-world clinical impact remains to be proven. The findings suggest AI could substantially improve early detection rates, potentially transforming outcomes for patients with this historically fatal condition. However, implementation challenges include integration with existing workflows and ensuring the model performs equally well across diverse patient populations and healthcare systems.