Inspiration

Starting from this alarming observation: “With 60% of the African population under the age of 25, the continent is home to the most dynamic youth in the world... but also the most vulnerable to unemployment (23% among young people, World Bank 2023).” Yet there is a paradox of stifled ideas. For example, one of the members of my project wanted to create a clothing brand but did not know the procedure, where to find suppliers, how to set competitive prices, and above all, what risks to anticipate.

Observing these challenges, we realized that:

Traditional mentoring is a luxury. In fact, only 18% of African entrepreneurs have access to expert advice (GSMA, 2022).

Resources exist, but they are invisible:

  • “In Kinshasa, a handmade soap manufacturer is unaware that a glycerin supplier is located 5 km from his workshop.”

Digital illiteracy is not inevitable. So our solution is to democratize expertise through:

 - Voice interfaces (to overcome illiteracy)  
 - Visual guides (infographics, video tutorials)  
 - Predictive algorithms (budget/risk simulations)  

Our vision is to “replace missing years of experience with a smart, community-based toolbox, where each user becomes a potential mentor.”

What it does

Our platform breaks down barriers: whatever your level of education, budget, or industry (agriculture, crafts, tech, etc.), Assistant Business guides you step by step. Finally, anyone can bring any project to life—from a small street business to a local recycling plant.

Assistant Business is an all-in-one entrepreneurial coach for Africa that:

  • Transforms ideas into action plans: From an idea to a list of necessary machinery with suppliers and partners according to budget
  • Creates synergies: Connects manufacturers or sellers with exporters

How we built it

Frontend (Next.js)

We developed a highly optimized interface that adapts to African realities:

  • Lightweight pages (less than 100KB) that load even with an unstable 2G connection
  • An integrated voice dictation system allows illiterate users to submit their business ideas by simply speaking into their phone
  • Interactive budget tables update in real time when the user adjusts their settings, with clear visualizations even on small screens

Backend (FastAPI)

Our server architecture combines power and simplicity:

  • Python allows us to quickly process complex requests while remaining accessible to our African team of developers
  • Integration with Google AI Studio is done via optimized API calls to minimize data consumption
  • Our lightweight API can even run on local mini-servers in neighborhood internet cafes

Artificial Intelligence

The brain of our solution uses Gemini to provide relevant advice:

  • Project analysis takes into account the local context (regional prices, specific regulations)
  • Step-by-step generation transforms a vague idea into an executable checklist with realistic deadlines
  • Our supplier search engine cross-references official data with verified contributions from the community

Text-to-Speech (TTS)

We make information accessible to everyone thanks to:

  • A voice in French that reads the advice with the right intonations and local expressions

Challenges we ran into

Finding the ideal prompt for Gemini

(Most fundamental problem—without this, nothing works)

Displaying AI results in simple tables

(Critical user experience)

  • Challenge: Transforming technical paragraphs into clear visual steps
  • Innovation: Self-adaptive system that converts text into color-coded cards

Localized speech synthesis module

(Major accessibility barrier)

  • Unsuitable robotic voices
  • Weight of audio files in rural areas

“Each obstacle we overcame brought us closer to our mission: to make expertise as accessible as a WhatsApp message.”

Accomplishments that we're proud of

Functional React interface with budget filtering page. Dynamic idea generation based on contextualized prompts. Use of Google AI Studio for the generative back-end (serverless). Clear output system (tables, speech synthesis, smooth navigation).

What we learned

Functional React interface with budget filtering page. Dynamic idea generation based on contextualized prompts. Use of Google AI Studio for the generative back-end (serverless). Clear output system (tables, speech synthesis, smooth navigation).

What's next for BUSINESS INNOV

Short Term (1–3 months)

Creation of a library of tested and validated prompts Redesign of AI responses in table format (budget, steps, suppliers) Improvement of voice integration (more stable, multi-language)

Medium Term (3–6 months)

Native offline mode: the tool works without a network User feedback module on AI results Customization of the type of assistance: ideation, budget, suppliers, steps

Long Term

Contextual AI assistant: understands whether the user is a beginner or advanced Integrated training in self-employment (mini audio lessons + local examples) Direct connection with suppliers and partners via a marketplace

Built With

  • fastapi
  • gemini
  • google-ai-studio
  • next.js
  • tts
Share this project:

Updates