Along with AI agents, the new generation of large language models, vision-language models and other large multimodal models are enabling powerful new capabilities that promise to transform industries. In this talk, we explore the requirements and architectures of agentic applications, including AI and non-AI requirements, and explore two main approaches to agent-based application architecture: integrating separate models and multimodal approaches. Through detailed examples, we demonstrate the pros and cons of each approach and discuss the challenges and opportunities of building practical agent-based applications on edge devices, including challenges associated with implementing large models at the edge.