Agentic architecture is important because AI agents are set to unlock the next era of enterprise productivity. While the term Agent is used in various scenarios, we assume that an agent is an application that iteratively achieves a given task using a phased approach. It plans the job, starts executing the plan using various tools it has at its disposal, and ultimately concludes on the achieved goal. This goes beyond the simple “prompt-response” paradigm. Implementing such agents could be non-trivial, therefore, this becomes an important topic for engineers who are designing and implementing such systems.

Join Max Pavlov, our AI Engineering Director, for an in-depth session on AI agents. We will review a few code examples of how such “agents that take turns” can be built and discuss the expansion and integration landscape for these agents. We will be reviewing an entry-level implementation of the agentic architecture.
How to build multi-turn agents (presentation)
Max Pavlov's Code