Our team deeply appreciates the entrepreneurial spirit and the groundbreaking innovations born from startups. However, we've also seen firsthand the immense friction in the investment process. For investors, the manual due diligence—sifting through countless pitch decks, business plans, and market data—is a monumental task that takes weeks or even months. The process is opaque for startups and often feels like shouting into a void. We realized that this bottleneck wasn't a problem of too much information, but a problem of inefficient processing. This hackathon was our opportunity to apply cutting-edge AI to automate this process, making it faster, fairer, and more accessible for everyone involved.
What it does InvestAI is a dual-sided platform that streamlines the startup investment process using a sophisticated, multi-agent AI system. For investors, it transforms a lengthy due diligence period into a near-instantaneous analysis. You upload a startup's pitch deck and relevant documents, and our nine specialized AI agents get to work, providing a comprehensive report covering market analysis, team expertise, financial projections, and risk assessment. The platform also features a smart RAG (Retrieval-Augmented Generation) chatbot for natural-language Q&A and a simple meeting scheduler to connect with promising startups. For startups, it provides an opportunity to get a professional, unbiased, and free due diligence report on their own venture, giving them a clear snapshot of their investment readiness.
How we built it We built InvestAI as a full-stack web application. The frontend is a modern, responsive single-page application built with React and styled with Tailwind CSS for a clean, professional user experience. On the backend, we used a Python FastAPI server to manage the API calls and orchestrate our AI agents. The core intelligence of the platform is powered by the Google Gemini API and Vertex AI. We utilized the Google ADK (Agent Development Kit) to design and deploy a system of nine specialized AI agents, each focusing on a different aspect of the analysis (e.g., a "Financial Analyst" agent, a "Market Research" agent, a "Risk Assessment" agent). We used the Exa API for real-time web search capabilities to ground our analysis in current market data. All user data, including uploaded files and generated reports, is stored persistently and in real-time using Firebase's Firestore database, ensuring data integrity across sessions.
Challenges we ran into The primary challenge was orchestrating a complex, multi-agent system within a tight hackathon timeline. Getting nine different agents to work together seamlessly—passing information, ensuring consistent formatting, and maintaining context—was a significant hurdle. We also faced the classic hackathon dilemma of managing API rate limits and token usage, which required us to be clever with our prompts and caching strategies. Finally, designing a user interface that could gracefully display a wealth of complex analytical data was a major UX challenge. We had to ensure the reports were digestible and actionable for both experienced investors and new entrepreneurs.
Accomplishments that we're proud of We are incredibly proud to have built a fully functional prototype of a sophisticated AI system in just a weekend. The ability to upload a full pitch deck and receive a detailed, multi-faceted analysis in minutes is a testament to the power of the Google AI platform. We're also proud of our dual-sided approach, creating a product that provides value to both investors and the startups they evaluate. It's not just a tool for one side of the market; it's a solution that fosters a more efficient and transparent ecosystem for everyone.
What we learned This project was a masterclass in full-stack development and AI agent design. We gained invaluable experience working with the Google Gemini API, learning how to prompt an engineer for complex tasks and orchestrate multiple agents. We deepened our knowledge of modern web development and learned the importance of focusing on a robust, scalable backend for AI-powered applications. Furthermore, we reinforced the critical lesson that the best products don't just solve a problem; they fundamentally improve a process for all stakeholders.
What's next for InvestAI This is just the beginning for InvestAI. Our next steps include:
Adding more specialized agents to the platform, such as a "Competitor Analysis" agent that uses the Exa API to find and evaluate competitors.
Building a robust matching algorithm to connect investors with startups that fit their specific criteria.
Expanding our data sources to include social media sentiment analysis and news trends.
Integrating with a full-fledged video conferencing API for a seamless meeting experience, eliminating the need for a separate scheduler.
Log in or sign up for Devpost to join the conversation.