Case study · AI Labs
Biritu Finance Bot
Your guide to financial wellness in Ethiopia. Empowering Ethiopians with accessible, reliable financial knowledge in local languages.
Project completed · Delivered August 16, 2025
- Timeline
- May 24 – August 16, 2025
- Duration
- 8 weeks
- Format
- Open-source AI Labs project, in partnership with StockMarket.et and Pi Financial
- Technologies
- RAG, LLM, NLP, Multilingual, AI, Chatbot
01 —
Overview
Financial literacy is a cornerstone of economic empowerment and stability for individuals and communities across Ethiopia. However, access to reliable, understandable, and contextually relevant financial information, especially in major local languages, remains a significant challenge. Many Ethiopians struggle to navigate banking services, understand savings and credit options, utilize digital finance tools like Telebirr effectively, or protect themselves from financial misinformation.
This project successfully bridged this gap by developing an AI-powered chatbot using advanced Retrieval-Augmented Generation (RAG) techniques. This chatbot, “Biritu Finance Bot,” provides trustworthy financial literacy guidance through an accessible interface, empowering users to make informed financial decisions.
02 —
Challenge
The lack of accessible financial knowledge contributes to several issues in Ethiopia:
- Limited access to reliable financial information, particularly outside major cities and in local languages.
- Complexity of financial products and services leading to confusion and underutilization.
- Vulnerability to financial scams and misinformation spreading online and offline.
- Difficulty finding Ethiopia-specific guidance on budgeting, saving, debt management, and digital finance.
- Absence of easily accessible, trustworthy support systems for basic financial queries.
These challenges hinder financial inclusion and prevent many Ethiopians from fully participating in the growing economy.
03 —
Solution
The primary goals of Biritu Finance Bot were successfully achieved:
- Developed an intelligent chatbot using Agentic RAG to provide accurate, easy-to-understand answers to financial literacy questions in Ethiopian local languages.
- Curated and structured a high-quality, verified knowledge base containing Ethiopia-specific financial information (banking, mobile money, savings, loans, etc.).
- Designed a user-friendly interface suitable for diverse users, potentially including those with lower digital literacy.
- Empowered Ethiopians with accessible knowledge to improve their financial decision-making and well-being.
04 —
Timeline
Completed in 8 weeks.
Weeks 1–2
Knowledge base curation & foundational setup
Gathering, vetting, and translating information; initial infrastructure setup.
Weeks 3–4
Knowledge base curation & initial RAG pipeline setup
Vector DB implementation and embedding strategy development.
Weeks 5–6
RAG pipeline development & initial agent logic
Retrieval tuning, generation prompting, language detection, and basic dialogue flow implementation.
Week 7
Model testing & refinement; UI development
Evaluating RAG performance, iterating on prompts/retrieval; Streamlit/Gradio interface development.
Week 8
End-to-end testing, final documentation & demo preparation
User acceptance testing, READMEs, and final presentation preparation.
05 —
What participants learned
Technical skills
- Curating and structuring knowledge bases for RAG systems
- Implementing Retrieval-Augmented Generation (RAG) pipelines
- Working with LLMs and embedding models for multilingual NLP
- Developing effective prompt engineering strategies
Development practice
- Building interactive AI applications using frameworks
- Collaborative software development with Git and GitHub
- Project management and documentation best practices
- Testing and quality assurance for AI applications
Ethical considerations
- Understanding ethical considerations in AI for information dissemination
- Addressing cultural and linguistic nuances in AI applications
- Evaluating social impact of financial literacy tools
- Ensuring accessibility and inclusivity in design
06 —
Participants
Requirements
Good communication skills
For team collaboration and accessing resources
Suitable for all levels
AI/Data Science beginners welcome, also valuable for experienced collaborators
Learning mindset
Strong learning mindset and willingness to collaborate actively
Interest in NLP and LLMs
Interest in Natural Language Processing and Large Language Models is a plus
Good sense of humor
Makes collaborative work more enjoyable and productive
Benefits
Real-world impact
Tackle a crucial real-world challenge impacting financial inclusion in Ethiopia
Portfolio enhancement
Build a cool AI chatbot project for your portfolio
In-demand skills
Develop in-demand skills in NLP, LLMs, RAG, and AI application development
Mentorship & networking
Collaborate with talented peers and experienced mentors
Practical experience
Gain practical experience in building and deploying AI solutions
Completion certificate
All active participants who complete the project will receive a completion certificate highlighting their skills and contributions to this important initiative.
07 —
Results
Thank you to all participants, mentors, and partners who made this project a success. Together, we've advanced financial literacy in Ethiopia through innovative AI technology.
- Biritu Finance Bot has been successfully developed — project delivered on August 16, 2025.
- Successfully completed with participants of all skill levels.
- Education-focused with real impact.
- Completed in 8 weeks (May 24 – August 16, 2025).
- Open-source project with community impact.
08 —
In partnership with
StockMarket.et
Pi Financial
Planning something similar?
Tell us about your project and we'll get back to you.