In the context of increasing AI integration in technology, deploying AI models on edge devices as part of modern Mobile Device Management (MDM) workflows is gaining significance. While cloud-based AI inference engines have proven successful in concept, they face limitations in scalability due to bandwidth, latency, and data costs, making edge deployment crucial for real-time processing. However, edge AI presents challenges such as securing AI models and maintaining a continuous feedback loop for model updates, requiring effective tools and frameworks. Platforms like Esper facilitate deploying AI workloads at the edge by providing necessary tools for data retrieval and model redeployment, which is essential for operationalizing AI through AI DevOps. This operationalization involves continuously integrating and delivering AI models across device fleets, a process that remains challenging for smaller companies compared to larger ones with bespoke infrastructure. Effective MDM solutions targeting edge AI must address these challenges to enable cost-effective, rapid deployment and evolution of AI models.