HW Barakah RAG System
Completed
AI/ML RAG Automation
Overview
HW Barakah needed a faster way to prepare quotations from project specifications while keeping pricing decisions grounded in internal knowledge and previous work.
The system helps staff upload project details, search internal reference data, and generate structured quotation drafts for review.
Workflow Improvements
- Upload project specifications for structured processing
- Search internal reference data quickly
- Draft quotations based on company knowledge
- Reduce repetitive manual pricing lookup
- Keep staff review in the quotation workflow
Technology Stack
- AI/ML: RAG (Retrieval-Augmented Generation)
- Vector Database: Embedded search capabilities
- Backend: FastAPI / Hono
- Database: PostgreSQL with vector extensions
Workflow
- User uploads project specification
- System searches internal reference data
- Quotation draft is generated from relevant context
- Staff reviews and finalizes the quotation
Outcome
The system improves quotation turnaround by giving staff a faster first draft and better access to internal pricing context.