IA & Data

IA générative - Avancé - EN

2 jour(s) • 14h

Description

Training Objective : Use generative AI in a professional, high-performance, and secure way to automate workflows, build custom business assistants, and integrate generative AI into operational systems with appropriate governance.

Objectifs pédagogiques

Design dynamic, modular, and high-precision prompts,
Build a specialized conversational AI agent contextualized to business needs,
Develop micro AI tools using low-code connectors (Make, Zapier, Notion AI, etc.),
Manage the quality, traceability, risks, and accountability of AI-generated content,
Structure sustainable and ethical governance of AI tools across teams and departments.

Public concerné

Business or project managers working on data/AI implementation,

Managers from HR, Marketing, Communications, Legal, Innovation or IT departments,

Transformation or automation leads in charge of process improvement,

Product Owners, consultants, or “AI Champions” driving adoption within their teams.

Prérequis

Solid understanding of generative AI principles (LLM, prompting, use cases),

Regular experience using ChatGPT, Copilot, or similar AI tools in a professional context,

Familiarity with at least one workflow or automation tool (Notion, Zapier, Google/Microsoft Suite).

Déroulé du programme

Advanced Prompt Engineering (2h)

Modular structure: conditional segments, adaptive templates,
Role/format/style logic and output control techniques,
Model comparison and response optimization.
Practical activity : Design a prompt engine that dynamically generates a professional document (email, memo, HR note…) based on context.

Build a Generative Micro-Service (3h)

Chaining prompts with tools: e.g., ChatGPT → Google Sheets → Email,
Creating workflow automations using Make / Zapier,
Adding conditional logic and post-treatment layers.
Practical activity: Build a workflow-based AI assistant that generates a deliverable based on form inputs (HR report, meeting summary, content draft...).

Designing a Specialized Conversational AI Agent (2h)

Using memory/context features (e.g. GPT-4, Claude 3),
Designing structured expert personas with clear scopes and limits,
Managing consistency and fallbacks.
Practical activity: Create a legal or HR expert chatbot from a business corpus, and simulate a dialogue for accuracy testing.

Building AI Governance Structures (2h)

Human-in-the-loop, documentation, output auditing,
Input/output control: prompts, templates, post-processing,
Risk and responsibility distribution.
Practical activity: Build a governance grid for an AI use case: versioning, validation, access, risk levels, and human checkpoints.

Scaling and Industrializing Generative AI (2h)

Identifying scalable use cases,
Cost/benefit/risk modeling of AI assistants,
Moving from POC to organizational deployment.
Practical activity: Define a roadmap to industrialize an AI assistant in a real business context: milestones, risks, indicators, and teams.

Final Capstone Workshop (3h)

Design and simulate the deployment of a generative AI tool integrated in a workflow,
Evaluate performance, governance, and user engagement,
Pitch the result and receive peer feedback.