The solution
DealGPT is an AI-powered deal intelligence platform I built for investors, executives, and deal teams. It streamlines the full deal lifecycle: sourcing opportunities, running due diligence, structuring terms, negotiating, and tracking execution after close.
The platform combines configurable AI agents, a conversational assistant, collaborative workspaces, and a global deal marketplace in one product.
Technical approach
The front end is a React and Vite application with a responsive chat interface, collapsible navigation, and dark/light mode. Users configure agents with uploaded knowledge bases (PDF, DOCX, CSV, and more), pick LLM providers, and enable capabilities like long-term memory and responsible AI guardrails.
Deal progress is tracked through five structured stages: concept validation, detailed due diligence, deal structuring, negotiation, and post-close monitoring. The marketplace lets teams filter by industry, region, type, and size, express interest, and manage saved deals.
The stack includes Supabase for auth and data, AWS for ML and storage, and containerized deployment for scale.
Outcome
A production platform that makes advanced deal workflows and AI assistance accessible to business users without deep technical expertise, with enterprise-grade security across sensitive deal information.