Neural networks have grown in a crescendo to become a cornerstone of current artificial intelligence (AI) systems. They are used in diverse real-world applications in nearly every domain, including medicine, arts, humanities, and social sciences. However, they have never been applied to Byzantine sigillography. This article addresses the specificities of Byzantine seals' images and how neural network algorithms must be adapted to help with their analysis. We also discuss the challenges, opportunities, and potential of combining different artificial intelligence methods to analyze Byzantine seals.