A team member worked at a startup and it completely tanked so we understood the problems of a startup during its initial stages. This bitter experience showed us importance of a taking timely, well informed decisions. This led to the root idea and inspiration for this project.
Introduction In today’s entrepreneurial landscape, students and aspiring founders often struggle to access the mentorship, insights, and tools required to develop and scale their innovative ideas. EDG3 is an AI-powered website that empowers entrepreneurs at George Mason University’s FUSE Digital Innovation Building by offering real-time, tailored business strategies, feasibility assessments, and market insights.
EDG3 uses a combination of Azure Synapse Workspace to filter synthetic data, Azure SQL to manage and store the data, and GPT-3.5 to provide accurate and actionable business advice to students and innovators. This website enables users to confidently navigate the startup process, from idea validation to market expansion.
About the FUSE Building The FUSE Digital Innovation Building at George Mason University is a state-of-the-art facility designed to foster entrepreneurship, collaboration, and cutting-edge research. Located at Mason Square in Arlington, Virginia, FUSE serves as a central hub for students, faculty, and innovators. It provides access to advanced technologies, co-working spaces, and incubation programs aimed at helping students and entrepreneurs turn their ideas into impactful businesses. EDG3 aims to complement FUSE’s mission by providing virtual business guidance through its AI-powered platform.
Problem Statement Entrepreneurs, especially students, often face key challenges, such as:
Lack of mentorship: Limited access to expert guidance on scaling and business development. Data overload: The abundance of online resources makes it difficult to extract actionable insights. Difficulty in assessing feasibility: Many founders struggle to evaluate whether their business ideas can succeed in the market. EDG3 addresses these challenges by training a custom GPT model on a curated and filtered dataset of business strategies, market trends, and feasibility insights. This enables the website to deliver relevant and actionable advice in real-time, helping users turn their ideas into viable businesses.
Solution Overview EDG3 is an AI-powered website that provides entrepreneurs with personalized guidance to help them develop and scale their businesses. It offers:
Business Roadmaps: Step-by-step plans for scaling a startup, tailored to the user’s specific idea and market. Feasibility Analysis: AI-driven assessments of the business idea’s potential success, identifying risks and market opportunities. Scalability Strategies: Detailed strategies to help startups grow, including funding options, partnership opportunities, and go-to-market plans.
Using Azure Synapse Workspace, the website filters and processes synthetic data from various sources to ensure the accuracy and relevance of the insights provided by the GPT-3.5 model.
Technical Implementation Data Collection through Web Scraping and Public Sources:
Synthetic data, including business strategies, market analyses, and funding insights, is collected through web scraping and other publicly available data sources. This dataset is critical for training the GPT model to provide real-world advice. Data Storage in Azure SQL Database:
The synthetic data is stored in an Azure SQL Database, organized into tables that allow for efficient querying and data management. The database acts as the central repository for business knowledge and is updated regularly as new data is gathered. Data Filtering with Azure Synapse Workspace:
Azure Synapse Workspace is used to clean and filter the synthetic data, removing irrelevant or redundant information. This ensures that only high-quality, actionable insights are used to train the GPT model, making the website’s recommendations more accurate and relevant. GPT-3.5 Model Training:
Once the data is filtered and refined, it is used to train a custom GPT-3.5 model (specifically using gpt-35-turbo Version 0301). The model is designed to provide personalized business strategies and advice, tailored to the user’s needs based on the filtered dataset. User Interaction via Website:
Users interact with EDG3 through a user-friendly website interface where they can ask specific questions related to business development and startup strategies. The AI-powered system responds in real-time, offering personalized advice, business roadmaps, and feasibility analyses.
Log in or sign up for Devpost to join the conversation.