GenAI-Powered Contract Intelligence Tool¶
Project Title: GenAI-Powered Contract Intelligence Tool
Client: Dr. Reddy’s Laboratories (via Deloitte)
Timeline: August 2023 - January 2024
Technologies: Python, GenAI, Gemini (PaLM 2), Document AI, Prompt Engineering, Web Scraping, GCP, Vertex AI
Overview¶
Developed and deployed a GenAI-based solution to automate the creation of bids by extraction of critical data from U.S. Government presolicitation and solicitation contracts. The tool enables sales and supply chain teams to proactively identify bidding opportunities and plan inventory and logistics strategies.
Key Contributions¶
- Built a web scraping pipeline to ingest presolicitation/solicitation documents from SAM.gov.
- Used PaLM 2 and Document AI to extract structured data from unstructured PDFs, including drug names, quantities, strengths, packaging, submission deadlines, and contact details.
- Applied RAG framework, prompt engineering and few-shot prompting techniques to fine-tune model responses and reduce hallucinations by controlling input context.
- Mapped extracted data to internal product and customer databases to identify actionable opportunities.
- Enabled early visibility into government demand, improving bid readiness and supply chain alignment.
- Created an initial bid automatically after identifying supply chain, manufacturing, logistics and pricing constraints by collaborating with different teams.
Impact¶
- Accelerated bid creation and opportunity identification for the sales team.
- Improved supply chain planning by forecasting demand 6+ months in advance.
- Enhanced efficiency and accuracy in navigating complex government contracts.