Design autonomous AI agents that carry out business tasks, orchestrate complex workflows and integrate with your company tools. From the fundamentals to multi-agent automation, 100% online.
Principles of autonomous decision-making, goal-oriented agents and self-optimisation techniques for your first agentic applications.
Design agents that interact with each other to solve complex problems and improve AI-human interaction.
Develop and optimise multi-agent systems: reinforcement learning, adaptive behaviour, collaboration and competition between agents.
Integrate autonomous agents into your existing business systems and structure a strategy for adopting agentic AI at scale.
Chain tasks, trigger actions and automate complete business processes with AI agents, with little or no code.
An introduction to agentic AI tools (Claude Code, Copilot, Cursor): how they work, how to steer them with prompts, first projects.
Delegate complete tasks to agents, build a customisation stack (Skills, MCP, Agents) and run several agents in parallel.
Understand and implement the MCP protocol to connect your AI agents to tools, databases and internal services.
Use agents to speed up development cycles: code generation, automated review, testing and continuous deployment.
Applying agents to concrete cases: document processing, automated reporting, customer support and supplier relationship management.
Govern the use of autonomous agents in the enterprise: permission control, action traceability and regulatory compliance.
A programme built with you, around your sector, tools and real use cases, from our existing modules or fully dedicated.
An AI agent plans and carries out several steps to reach a goal, rather than answering a single one-off instruction.
Thanks to protocols such as MCP, an agent can read your data, call your internal tools and act directly in your systems.
Several specialised agents can collaborate on the same complex task, each responsible for part of the work.
An AI agent plans and carries out several steps to reach a goal, instead of answering a single question. It can read your data, call your tools and act in your software.
MCP is a standard protocol for connecting an AI agent to tools, databases and internal services. One module of the course is entirely dedicated to it.
The complete pathway lasts 3 to 5 days. It is made up of 1- to 3-day modules (7 to 21 h) that we combine according to your level and goals.
Not for the workflow automation modules, which can be done with little or no code. Agentic development and MCP are aimed at technical profiles; see also the AI for Developers course.
Yes, online in a virtual classroom, on your own tools and real use cases.
RAG gives an LLM access to your documents; an agent acts in your tools. If you mainly need an AI to answer from your knowledge base, start with the RAG & Pipelines course.
Let's talk about your training