STL Digital said some of the most significant hurdles in employing genAI include training the technology for specific enterprise use cases, find-tuning algorithms for greater accuracy, ensuring data security, handling multimodal inputs, and improving the quality of the trained data.