Welcome to Module 8 — Agentic AI.

In this lesson, you will move beyond AI systems that simply respond to prompts and explore AI systems that can reason, plan, use tools, access information, take actions, and complete multi-step tasks. You will learn the architecture behind AI agents and understand how modern agentic systems combine LLMs, tools, memory, planning, reasoning, and feedback.

Topics Covered:

1. What is Agentic AI?
   Understand the concept of Agentic AI and how AI agents differ from traditional AI applications and simple chatbots.
2. AI vs AI Agents
   Learn the difference between systems that generate responses and systems that can make decisions, perform actions, and work toward defined goals.

3. Agent Architecture
   Understand the core components of an AI agent:

   Goal → Reasoning → Planning → Tool Use → Action → Observation → Feedback

4. LLM as the Agent Brain
   Understand how Large Language Models can provide reasoning and decision-making capabilities within an AI agent.
5. Tools and Function Calling
   Learn how agents can interact with external tools and functions to perform tasks beyond generating text.
6. Types of Agent Tools
   Explore tools that allow agents to interact with:
   • Web Services
   • APIs
   • Databases
   • Code Execution
   • Browsers
   • Files and Documents
   • External Applications

7. Memory in AI Agents
   Understand how agents can maintain information during and across interactions.

   Explore:

   • Short-Term Memory
   • Long-Term Memory
   • Conversation Memory
   • External Memory
8. Planning and Task Decomposition
   Learn how agents can break complex goals into smaller tasks and determine the sequence of actions required to complete them.
9. Reasoning and Decision-Making
   Understand how agents evaluate information, determine next steps, select tools, and adapt their actions based on results.
10. ReAct Pattern
    Learn how reasoning and actions can be combined in an iterative process where an agent reasons, takes an action, observes the result, and continues toward the goal.
11. Planner-Executor Architecture
    Understand how one component can create a plan while another component executes individual tasks.
12. Reflection and Self-Correction
    Learn how agents can review their own results, identify problems, and improve their actions or outputs.
13. Autonomous Agents
    Understand how agents can perform multiple steps with limited human intervention while working toward a defined objective.
14. Human-in-the-Loop AI
    Learn how human approval and intervention can be incorporated into agent workflows, especially for sensitive or high-impact actions.
15. Multi-Agent Systems
    Understand how multiple specialized AI agents can collaborate, communicate, and divide responsibilities to solve complex problems.
16. Agentic RAG
    Learn how AI agents can dynamically retrieve information from knowledge sources and decide when additional information is required.
17. MCP and AI Tool Connectivity
    Understand the concept of standardized connections between AI applications and external tools, data sources, and services.
18. Agentic AI Frameworks
    Get introduced to popular frameworks used for developing agentic applications, including:
    • LangChain
    • LangGraph
    • CrewAI
    • AutoGen
    • LlamaIndex
    • Semantic Kernel
19. Agent Security and Reliability
    Understand important challenges such as unauthorized actions, prompt injection, tool misuse, unreliable reasoning, data access, permissions, and the need for human oversight.
20. Agent Evaluation and Monitoring
    Learn how agent performance can be evaluated based on task completion, tool usage, accuracy, reliability, cost, latency, and safety.
21. Real-World Agentic AI Applications
    Explore applications such as:
    • AI Research Assistants
    • Business Automation Agents
    • Customer Support Agents
    • Coding Agents
    • Data Analysis Agents
    • Sales and Marketing Agents
    • Enterprise Workflow Agents
    • Multi-Agent Business Systems

Major Project:
Build a Multi-Agent AI Research & Business Assistant where specialized AI agents collaborate to research information, analyze data, use external tools, generate reports, and complete a defined business task with human approval where required.

Learning Outcome:
By the end of this lesson, you will understand how AI agents work, how LLMs, tools, memory, planning, reasoning, and actions are combined, and how single-agent and multi-agent systems can solve complex multi-step tasks. You will also understand the foundations required to design and build production-oriented Agentic AI applications.

Lesson Resources:
Video Lesson + Presentation PDF + Quize