inaraX

Agentic AI

For anyone building or overseeing AI agents

Agentic AI covers what separates an agent from a chatbot: the loop of reasoning, acting, and observing, the tools an agent reaches for, and the orchestration that turns one agent into a working system. It ends where production begins — reliability, evaluation, security, and governance.

Created by InaraX
Instruction languages: English, Arabic
Premium

How agents reason, act and remember · MCP, RAG and multi-agent patterns · Evaluation, security and governance

Highlights for this program

6
modules
40
lessons
InaraX

Earn a certificate from InaraX

Complete a level to earn its certificate, which you can share on LinkedIn, social media, or your resume or CV.

What you'll learn

  • ✓What an agent is: the LLM, its tools and its memory
  • ✓How agents reason, plan and call tools, including MCP
  • ✓The context window, long-term memory and context engineering
  • ✓The reason–act–observe loop and structured tool schemas
  • ✓Bringing in outside knowledge with RAG
  • ✓Workflows, orchestration, handoffs and agent-to-agent (A2A) communication
  • ✓Multi-agent patterns: supervisor, pipeline, hierarchical and swarm
  • ✓Evaluating agents, common failure modes, human-in-the-loop and the OWASP Agentic Top 10

Course content

3 levels • 6 modules

  • Module 1: What Is an Agent?Locked
  • Module 2: How Agents Think and ActLocked

Requirements

  • Comfort using AI chat tools day to day.
  • Helpful (not required): having written a prompt you were pleased with, and one that failed.

Description

Agentic AI is a level-based program (Associate → Intermediate → Advanced) for anyone who has used AI tools and now wants to understand the systems being built on top of them. You'll start with what an agent actually is and how one decides what to do next, then build a working agent and orchestrate it into a workflow, and finish with multi-agent systems and the practices that keep them reliable, secure, and accountable.

By the end of this course, you will:

  • Explain what makes something an agent rather than a chatbot.
  • Describe the reason–act–observe loop and where each step goes wrong.
  • Give an agent the right tools, memory and context for a task.
  • Orchestrate agents into workflows, and know when a fixed workflow is the better shape.
  • Choose a multi-agent pattern and understand how multi-agent systems fail.
  • Evaluate, secure, and govern agents well enough to put them in front of real work.