Godefroy De Compreignac

👋 The Context

The traditional recruitment equation is broken: endless CVs, complex processes, massive candidate ghosting. A visionary entrepreneur had a wild idea: what if good recruitment was primarily about a good meeting? What if you could apply in 10 minutes from a booth, guided by a benevolent AI that helps you create your CV and connects you to the right opportunities without hassle?

💪 The Challenges

Cibli's ambition was revolutionary: to create an ecosystem of physical booths where anyone could enter, have a natural conversation with an AI in 10 minutes, generate an optimized CV, and discover relevant job offers. A clear philosophy: "good recruitment is above all a good meeting" facilitated by a bit of tech, a lot of human touch... and a dash of audacity. The approach blends technological innovation and inclusivity, where AI serves humans without ever replacing them.

✅ What I Implemented

I completely rethought the candidate experience from scratch. The approach revolves around a natural conversational journey: simplified onboarding with explanatory screens about the concept, AI-guided interview with a progress breadcrumb, automatic generation of an editable CV, and intelligent matching with job offers. The core innovation lies in an AI capable of conducting a genuine discovery interview with kindness. I designed conversational algorithms that adapt to the candidate's profile, integrated text-to-speech and speech-to-text technologies for natural interaction, and developed sophisticated technical prompts that intuitively extract skills and experiences. The AI guides without judging, reveals strengths without forcing, in full respect of the ethics and inclusivity dear to Cibli.

👋 The Context

In a context where digital agencies constantly need to feed their client knowledge to create relevant products, the client agency faced a recurring paradox: the team recognized the strategic value of client interviews but could never find the time to conduct them effectively. How to transform this operational constraint into an innovation opportunity?

💪 The Challenges

The traditional interview process proved time-consuming: scheduling, preparing questionnaires, conducting the interview, taking notes, summarizing, and transforming into usable content. Between project schedule constraints and operational load, this crucial step was systematically postponed or simplified, depriving the team of essential information to fuel strategic and creative thinking.

🎯 The Objectives

With the emergence of conversational AI technologies and their potential to automate repetitive tasks, I saw an opportunity to automate interviews.

✅ What I Implemented

The technical construction was based on a React/TypeScript architecture with GraphQL, integrating advanced AI models to generate contextual questions and adapt the interview flow in real-time. The system can understand responses, follow up with relevant questions, and maintain the conversation's thread like an experienced human interviewer.

The development focused on a system that allows creating customizable interview templates for different content types: blog posts, case studies, LinkedIn posts, client feedback. The AI then analyzes all collected responses to automatically generate ready-to-use content, respecting the style and structure defined in each template.

I am currently leading the development of industrial generative AI solutions for 2 major CAC40 companies. My expertise covers RAG (Retrieval-Augmented Generation) architecture to connect AI to business data, the implementation of multi-agent systems orchestrating several specialized AIs, and MCP (Model Context Protocol) integration to make IT systems and AI agents interoperable.

Beyond the technical aspects, I focus on crucial issues: ensuring the quality and reliability of AI responses in an industrial context, designing intuitive user interfaces that facilitate adoption, and orchestrating large-scale deployment in complex organizations.

My approach combines product vision, technical excellence, and change management to transform promising POCs into robust solutions adopted by thousands of business users.