What Are Agentic AI Systems – And Why Do They Matter?

Agentic AI systems are the next evolution in how we interact with intelligent software. These systems go beyond passive question-answering to actively plan, reason, and execute multi-step tasks — autonomously or with minimal human input. From automating IT resolutions to orchestrating end-to-end business workflows, Agentic AI is becoming essential for building truly intelligent and proactive systems.

In this session, we’ll unpack what Agentic AI actually is, how it differs from traditional LLM pipelines, and why it’s being hailed as a foundational shift in AI system design. We’ll walk through real-world architectural patterns, explore key components (agents, memory, tools, evaluators), and show how these systems are being used across enterprises. Whether you're a developer, data scientist, or aspiring architect, this session will help you understand where the future of applied AI is heading.

Key Takeaways:

• Attendees will gain a clear understanding of what defines Agentic AI and its active, autonomous nature.
• They will grasp the architectural distinctions between Agentic AI and traditional AI systems.
• Attendees will learn about the essential building blocks (agents, memory, tools) for constructing Agentic AI.
• They will identify practical enterprise applications and the future potential of Agentic AI.

Prerequisites:

• Basic understanding of how LLMs work (GPT, ChatGPT, etc. and Python programming language.

Lekha Priya is an AI Architect with deep expertise in designing enterprise-scale GenAI solutions — from retrieval-augmented generation (RAG) systems to agentic workflows for IT Ops and automation. With a background in building modular AI pipelines, I specialize in aligning LLM-based systems with business value, scale, and reliability. I currently lead architecture efforts at Kyndryl, shaping next-gen AI systems that bridge unstructured intelligence and action automation.