ENFR

Personal project | Android

Eclair

Éclair is an app designed to help foreign residents in France pass their civic exam for multi-year and permanent residency cards.

  • AI agent design
  • Prompt engineering
  • Multi-agent orchestration
  • Product management
Eclair app cover: "Designing an intuitive, adaptive study app for effortless preparation for the French civic exam". Three phone screens show the exam date picker, the completed daily session, and the 5 exam themes.

✨ Overview

Role

Product Manager & Automation Specialist

Discovery & Research, UX Architecture & Strategy, Rapid Prototyping & Testing, Brand Identity & UI Design

Platform

Android

Timeline

2026 – Current

Status

In Development

Deliverables

Design

User Flow, Wireframes, High-Fidelity Prototype, Design System, Branding, Visual Design, Logo

Overview

Éclair: an AI-orchestrated product team. I designed and ran a team of seven specialist AI agents (product, UX/UI, architecture, backend, frontend, QA, DevOps) to take a civic-exam prep app from problem to build-ready plan.

Tools

  • Claude
  • Mobbin
  • Figma

✨ Context

Since January 2026, anyone in France applying for their first multi-year residence card (Carte de Séjour Pluriannuelle – CSP) or resident card (Carte de Résident – CR) must pass a mandatory 40-question civic exam. Failing the exam means retaking it and facing direct delays in securing legal residency status.

Balancing high-stakes exam preparation with everyday work, administrative duties, and family life creates immense stress. Candidates need an everyday app that gives quick access to the exact content they need, so they can study on the go: on the metro, or during any small pocket of free time.

Preparing for a critical exam shouldn't feel like a heavy, time-consuming burden on top of an already busy life. By providing direct, friction-free access to study materials on mobile devices, candidates can seamlessly integrate exam preparation into their daily routine.

✨ The Challenge

Problem Framing

Most existing resources for the French civic exam fail to provide a centralized and efficient learning experience. Applicants are given access to long lists of around 200 potential exam questions, but are left on their own to research, verify, and compile accurate answers.

Additionally, passively reading through hundreds of questions creates severe cognitive overload, making memorization difficult. There is a clear gap in the market for an intelligent, on-the-go tool that combines verified content with adaptive learning technology, turning studying into a structured, low-stress daily routine.

Pinpointing Issues

  1. Missing & Fragmented Answers: Official sources provide a large bank of 200 questions, but candidates have to spend valuable time hunting for verified, accurate answers across multiple platforms.
  2. Passive & Inefficient Memorization: Rote-reading massive lists of text is ineffective, leading to poor retention and high exam anxiety.
  3. One-Size-Fits-All Learning: Traditional methods force candidates to review questions they already know, wasting time rather than focusing on their personal weak spots.
  4. Lack of Progress Visibility: Candidates have no clear way to measure their readiness or know when they are truly prepared to score at least 32/40 on the official test.
  5. Rigid Study Demands: Existing materials require long, uninterrupted study sessions, making it hard for busy people to study during quick daily downtime.

✨ The Solution

Goals

The solution is an on-the-go study companion that combines verified exam material with intelligent micro-learning. It transforms the overwhelming 200-question pool into manageable, effortless daily prep where users can:

Solution Grid

  • Verified Answers

    Access all ~200 official questions paired with clear, verified answers in one place.

  • Spaced Repetition

    Master key facts for the long term using proven review intervals.

  • Adaptive Focus

    Automatically concentrate on challenging questions based on individual performance.

  • On-the-Go Preparation

    Fit bite-sized study sessions into daily commutes, metro rides, or quick breaks.

  • Readiness Tracking

    Monitor real-time progress toward achieving the required 32/40 passing score.

  • Stress-Free Learning

    Build true exam confidence without adding dedicated study hours to a busy schedule.

✨ Design Process

Approach: a product team made of AI agents

Instead of one general-purpose AI chat, I split the product lifecycle into phases and gave each phase a specialist agent:

  1. Product ManagerProblem analysis, competitor research, the full product spec.
  2. UX/UI DesignerThe design system and screen-by-screen specs.
  3. System ArchitectTech stack, data model, security and delivery order.
  4. Backend & Frontend EngineersReady to build from the architecture’s interface specs.
  5. QA & Test AutomationA test plan focused on the failures that would cost user trust.
  6. DevOps & DeploymentBuilds, app store releases and runbooks.

Plus a Field Researcher agent that collected real candidates’ experiences of the exam’s situation questions.

Every handoff is a document. The designer works from the product spec, the architect works from both, and the engineers work from all three. Because each agent builds on the previous agent’s work, every decision can be traced back to where it came from.