Architectural Patterns for LLM Adoption in Agentic AI
Sumit Ranjan, Divya Chembachere, Lanwin Lobo · Apress eBooks · 2025
As organizations move beyond experimentation with Generative AI, a new design frontier is emerging—one where LLMs are no longer standalone tools but dynamic components of autonomous, goal-oriented systems. Agentic AI systems redefine how LLMs operate: from responding to prompts to initiating actions, making decisions, and adapting over time. To harness this potential, enterprises must rethink the architectural foundations that support these capabilities. This chapter introduces key architectural patterns that enable scalable, reliable, and trustworthy adoption of LLMs within Agentic AI systems.