The Challenge

Our client had built a substantial library of legal documents — covering workplace harassment, discrimination law, and pay gap legislation — but had no way to make that content accessible to the thousands of people who needed it. Users seeking legal guidance either waited days for a staff response or left without answers. Staff time was consumed by repeat queries already answered in existing documents.

Specific problems the team was dealing with:

• Staff answering the same legal questions repeatedly via email and phone
• Users unable to navigate dense legal and policy documents independently
• No capacity to scale guidance across 25 NGO institutions without hiring more staff

Our Solution

We built a RAG-powered legal assistance chatbot deployed on WhatsApp — converting the client's entire legal document library into an always-available, conversational resource accessible on any mobile phone.

• Document vectorization: All legal documents were processed, chunked, and stored in a vector database, enabling semantic search across the full library rather than simple keyword matching.
• Contextual answer generation: When a user sends a query, the system retrieves the most relevant passages and generates a clear, plain-language answer — without the user needing to know which document to look in.
• WhatsApp deployment: The chatbot runs natively in WhatsApp, meeting users on a platform they already use and trust, with no app download or account creation required.

Technologies Used

• RAG (Retrieval-Augmented Generation)
• Vector Database / Vector Embeddings
• Python / JavaScript
• WhatsApp Business API

Results

• 20–50 queries per day handled automatically, with no staff involvement for routine legal questions
• Deployed across 15 NGOs — 15 institutions served from one system
• Instant access to guidance for users who previously had no realistic path to professional legal help for quick queries

‍