Digital product design studio

We turn complex processes into simple digital experiences.

Every project starts by understanding your work before writing a single line of code.

01The real problem

What slows your business down is not always visible on your website.

Files multiply. The same information is entered several times. Decisions wait. Clients and teams work around tools instead of relying on them.

We first look for invisible work. That is often where the most valuable first transformation lies.

02Transformations

Transformations that reveal how we think.

01

Digital product

JANGAI

A product starts with a clear architecture.

Context

JANGAI is an education platform for Senegal’s Baccalaureate. Its first public phase makes official past papers available through a deliberately simple journey.

01 · Before

Students need reliable resources and a clear path from subject to paper, without confusing a public library, AI and promises of prediction.

02 · Audit

Define a focused first phase — Landing → Home → Subject → Paper → PDF — then build diagnostic and progress tools from real learning data.

03 · Solution

Flutter Web, Supabase Postgres and Storage, PDF.js. The public catalogue isolates published papers and official PDFs; the app then separates papers, diagnostics, weaknesses and study plans.

04 · Impact

The first phase offers 257 papers across 20 subjects. Progress features reveal scores, weaknesses and plans only when sufficient learning evidence exists.

Prototype

Make every interface choice the consequence of an identified need.

Product screens
JANGAI student homeJANGAI lesson viewJANGAI general knowledge question
What we learned

Educational trust comes as much from explicit limits as features: official sources, clearly signalled insufficient data and evidence-based progress.

Technology & system

Flutter Web · Supabase Postgres · Supabase Storage · PDF.js · progress engine

02

Product experience and system

PETAW

Make important decisions easier to take.

Context

PETAW was designed as a “second mind”: an assistant that resumes context, adapts its response and remains calm in its interaction.

01 · Before

A generic assistant forces users to start again and exposes technical complexity that does not help them act.

02 · Audit

Design trust before the technical demonstration: project memory, profile context, clear language, source citations and transparent usage limits.

03 · Solution

A shared API gateway handles model routing, quotas, web search, document ingestion and privacy safeguards; the interface translates that complexity into understandable intentions.

04 · Impact

The product supports French, English and Wolof experiences, sourced answers, continuity memory and readable plan limits. Infrastructure remains invisible to the user.

Prototype

Design a product that helps people act, not merely accumulate information.

PETAW · intention before interface
PËTAW

Your context matters.

A second mind that picks up the thread without making you start again.

◉  Resume
Recent work remains available.
◐  Personalisation
Context refines the response.
Ask PETAW something…
Product screens
PETAW personalised assistant space
What we learned

In an AI product, perceived quality depends as much on explanation, control and continuity as on the underlying model.

Technology & system

TypeScript · multi-provider API · Supabase · web search · OCR · embeddings · hybrid retrieval

03

Shipyard operations

CORE

A daily report became a verifiable decision chain.

Context

On a complex site, the daily report, teams, hours, activities and cables cannot remain separate information or last-minute photographs.

01 · Before

Data had to be entered by the role that knows it, checked by the role that controls it and used by the role that decides — without ambiguity or manual reconstruction.

02 · Audit

Build a clear flow: Capo enters, Ufficio verifies, CORE archives, Direzione reads. Every screen is defined by the role, expected action and evidence produced.

03 · Solution

A React application with Supabase: structured reports and production lines, Capo / Ufficio / Direzione roles, PDF export, archiving and an INCA module for importing and analysing cable routes.

04 · Impact

The system centralises the daily report, preserves operational data and provides role-controlled reading. Import functions connect technical documents to field information.

Prototype

Move data with its context, owner and evidence — from the yard to management.

One path, connected decisionsin motion
CAPOUFFICIOCOREDIREZIONE
What we learned

A field tool becomes useful when it respects the real chain of responsibility, not when it adds a reporting layer above the work.

Technology & system

React · Supabase · PDF.js · XLSX · OCR · PDF generation · role-based control

04

Market research and simulation

Shakur Quant Engine

Read the market without confusing analysis with execution.

Context

Market reading requires scattered signals to be brought together without turning every chart into a promise of action.

01 · Before

A financial interface can easily create an illusion of certainty or hide the real state of data, models and execution.

02 · Audit

Surface safeguards in the interface: data freshness, simulation mode, volatility regime, signal confidence and a separate analysis layer.

03 · Solution

A private console organises the market view, price data, simulation controls and an agentic layer that analyses without placing real orders.

04 · Impact

The product provides a structured market reading while clearly displaying data status and the absence of real execution.

Product screens
Shakur Quant Engine market interface
What we learned

Trust in a financial tool comes from readable limits as much as accurate charts.

Technology & system

market data · simulation · AI analysis · visualisation · risk controls

05

Driver training

Tekno Autoscuola

The website was not the real problem. The working day needed better orchestration.

Context

The initial request was for a website. It led to DriveOS: an operational space connecting management, the office and instructors around the same working day.

01 · Before

The risk was improving the shop window without reducing the administrative load weighing on the business.

02 · Audit

Turn observation of daily work into a progressive digitalisation roadmap before committing to an interface.

03 · Solution

A clear journey from enrolment through theory, the learner’s permit and the exam, paired with an operational space that circulates useful information without re-entry.

04 · Impact

The website becomes one useful step in a broader project: reducing friction, making information reliable and improving follow-up.

Product screens
DriveOS for Tekno AutoscuolaAutoskuola 4.0 public journeyDriving journey stages
What we learned

A website request can be the signal of a deeper operational problem.

03Before / After

Make change visible, step by step.

01

Beforethe work as it is actually experienced

02

Auditthe established points of friction

03

Prototypechoices put to the test

04

Final productthe tool people adopt

05

Impactwhat genuinely changed

04Our method

Decide before developing.

A method structured enough to make every decision clear and flexible enough to respect the reality of your business.

  1. 01

    Discovery

    Understand the business, people and constraints.

  2. 02

    Audit

    Identify lost time, broken flows and opportunities.

  3. 03

    UX

    Clarify journeys before drawing the interface.

  4. 04

    Architecture

    Organise information, roles and decisions.

  5. 05

    Prototype

    Make choices visible, testable and open to discussion.

  6. 06

    Development

    Build only what solves a real problem.

  7. 07

    Testing

    Validate against real use, not assumptions.

  8. 08

    Delivery

    Put in place a tool the team can actually adopt.

  9. 09

    Evolution

    Measure, learn and improve what matters.

05What we do not do
  • Create a website in one day
  • Use a template and change the colours
  • Develop without understanding the business
  • Multiply unnecessary features
06What we do
  • Business audit
  • UX and product architecture
  • Digitalisation and automation
  • SEO optimisation
  • AI products
07How we work

You take part in the decisions that transform your business.

The client participates. Decisions are documented. Every choice is justified.

We never develop a feature without solving a real problem.

08Build system

An AI-augmented workflow. A decision that remains human.

AI accelerates execution. It does not replace the audit, judgement or responsibility for the decision.

  1. 01

    Understand

    Business audit, flows, constraints and success criteria.

  2. 02

    Design

    Product architecture, journeys, prototypes and documented UX decisions.

  3. 03

    Build

    Coding agents accelerate implementation while architecture remains governed.

  4. 04

    Verify

    Review, testing, quality control and confrontation with the real need.

  5. 05

    Evolve

    Deployment, measurement, documentation and continuous improvement.

Codexassisted exploration, implementation and review
Claude Codecode reasoning and accelerated development
AntigravityAI-assisted work orchestration
Next.js & TypeScriptrobust, maintainable interfaces
Supabase & PostgreSQLdata, authentication and access rules
GitHub & Vercelversioning, controls and deployment
AI, OCR & automationprocess information and remove repetitive work

The tool is secondary. The working system, the quality of decisions and the impact on your business come first.

09Discuss your project

Let us prepare a useful first conversation.

These answers help us understand your context before we meet.

Or write to us on WhatsApp · Telegram.