EYA PARTNERS
Back to catalogue
EYA-F05Continuous training programme

AI Agents Governance

Frame AI agents deployment in compliance with the AI Act

Duration
14 hours (2 days)
Quote response lead time
24 working hours
Training start lead time
First start available 7 to 30 days after signing the agreement (subject to scheduling). Agreement signed ≥ 7 working days before D1.
Format
Blended Learning (50% in-person + 50% online)
Group size
12 participants maximum
Location
On-site, in-person Paris or online
Language
French (English on request)

Access conditions

  • ▸Personalised quote within 24 working hours
  • ▸Training agreement signed by both parties
  • ▸Prior positioning test (≥ D-5)
  • ▸Preparatory documents (programme, welcome booklet, internal rules) provided before start

Lead times to training

  • ▸Agreement signed ≥ 7 working days before training start
  • ▸First D1 available 7 to 30 days after signing the agreement (subject to scheduling and funding)
  • ▸Average cumulative lead time: 14 to 30 days between first request and training start

Evaluation methods

  • ▸Entry positioning test (self-assessment + quiz, ≥ D-5)
  • ▸Formative quizzes during session (D1, D2, D3 depending on duration)
  • ▸Individually graded deliverable (validation threshold 65/100)
  • ▸End-of-session evaluation questionnaire (D0)
  • ▸Follow-up evaluation questionnaire (D+90 — field impact)
  • ▸Attendance certificate and realisation certificate provided at end

Overview

A cutting-edge training on generative AI agents: architectures (RAG, multi-agent), governance framework (AI Act, GDPR), supervision and audit.

Learning objectives

  • Understand how AI agents work and their architectures
  • Define a governance framework for AI agents deployment in the enterprise
  • Master the ethical, regulatory (EU AI Act) and security stakes
  • Design a supervision and audit policy for AI agents

Detailed programme

01

AI Agents: fundamentals and typologies

3h

Defining an AI agent: perception, decision, action. Typologies: reactive, deliberative, hybrid, multi-agent. LLM-based agents (AutoGPT, LangChain, CrewAI). Platforms: OpenAI, Azure AI, Mistral, Hugging Face.

02

Enterprise AI agents architectures

4h

RAG architecture. Multi-agent orchestration. IS integration: APIs, connectors, security. Memory and context. Workshop: designing an AI agent architecture.

03

AI agents governance framework

4h

Principles: transparency, accountability, traceability. EU AI Act 2024: obligations, risk levels, compliance. GDPR and AI agents. AUP. Roles: AI Owner, AI Auditor, AI Ethics Officer.

04

Supervision, audit and security of AI agents

3h

Risks: hallucinations, bias, prompt injection. Human-in-the-loop, guardrails. Logging and auditability. Compliance KPIs. AI incident management plan.

Evaluation & follow-up

Initial positioning quiz. Workshop: designing an AI governance framework (D2 delivery). Post-session evaluation questionnaire. Individual attendance certificate.

Pedagogical resources

Live AI agents demos (LangChain, AutoGPT, GPT-4o). Documented real AI incidents. EU AI Act framework provided.

Trainer qualification

AI & IS Architecture Expert, pioneer of enterprise AI adoption. Active tech watch on generative agents and international governance.

Page last updated on : 08/05/2026