• Understand the foundations of agentic AI and its specificities
• Identify The most used architectures to create autonomous agents
• Evaluate The relevance of a use case for an AI agent
• devising an agent functional diagram integrating the key components
• Handle or test existing agents using open source or no-code tools

Morning: understanding and modeling AI agents‍1. Introduction to agentic AI (45 min)‍

• Definition of an autonomous agent and differences with classical AI
• types of agents: reactive, proactive, mono-agent, multi-agent
• concrete use cases in business

‍2. Technical architecture of an AI agent (1h)‍

• key components: LLM, tools, memory, orchestrator
• overview of frameworks: LangChain, AutoGen, CrewAI, Dust
• reading and analyzing a functional diagram

‍3. Demonstration of an autonomous agent (45 min)‍

• live demonstration
• perception—action loop, behavior analysis
• Discussion of observed limits

‍4. Use cases and limitations (1h)‍

• areas suitable for agency
• risks: hallucinations, supervision, RGPD
• criteria for validating an agent project

Afternoon: understanding and experimenting with an AI agent

‍‍5. Application and integration of AI tools into a workflow (1h30)‍

• definition of a business need and objectives
• construction of a functional diagram (inputs, outputs, tools)
• presentation and group discussion

‍6. Handling or testing an AI agent (1h15)‍

• testing an agent via a no-code tool or dedicated interface
• Analysis of behavior under realistic conditions
• identification of optimization levers

‍7. Integration into a business workflow (30 min)‍

• Reflection on concrete uses
• Steps for launching a pilot project
• exchanges and collective synthesis