Loading...
10/26/2025
Insurify
code is the future, one innovation at one time
A Project Made By
Submitted for
Overall First Place
Finance Track
Built At
Gator Hack IV
Hosted By
Insurify
Framing the Problem
What is the problem you are trying to solve? Who does it affect?
Purpose: An interactive app that allows users to verify insurance claim using custom AI model, web scrapper, and LLM based github repository
Goal: To automatically analyze property images and contextual housing data (such as location, building details, and condition) using AGI that generates a detailed report. This platform helps users quickly understand the overall housing condition and access key information about maintenance needs, safety concerns, and cost estimates.
Insurify helps when:
Homeowners need quick repair estimates and insurance clarity
Insurance adjusters require fast risk assessment from images
Real estate agents want to impress clients with visual history and analysis
Buyers saves time before they look for houses in person
Idea Explanation
What is your idea? How does it fix the problem?
Github repository: https://github.com/yashnaray/Temp-model.git
Youtube:
Ideation: Extract the data from Google image, score on how the bad image is, run though AGI,
Implementation
How do all the pieces fit together? Does your frontend make requests to your backend? Where does your database fit in?
Challenges
What did you struggle with? How did you overcome it?
Challenge | Overcome | |
Lack of knowledge | At the beginning of the project, our team lacked technical knowledge in areas such as AI image detection, web scraping, and vector databases. | To overcome this, Yash actively researched tutorials, sought online resources, and learned how to use React for building the frontend. Through consistent learning, he gradually gained the skills necessary to implement these features successfully. |
Lack of communication | At first, our group faced communication issues. Some members weren’t fully updated on project progress or responsibilities, which caused confusion and delays. | We solved this by setting up clear communication channels, such as regular team check-ins, shared progress trackers, and open discussions. This helped everyone stay aligned and improved teamwork significantly. |
Lack of motivation | During the middle phase of the project, our team experienced a drop in motivation, especially when progress felt slow or tasks became complex. | We learned we should have set up smaller milestones, celebrated achievements, and reminded ourselves of the project’s purpose. This approach reignited our enthusiasm and helped maintain steady momentum. |
Time organization | Our group struggles with time organization. We rarely have all members present when working on the project, and several people are often missing during meetings or work sessions. As a result, uneven participation affects both our productivity and teamwork. |
Accomplishments
What did you learn? What did you accomplish?
Our team learned:
Webscrapping: Used Openrouter API data to provide up-to-date information like prices, location, and property details for analysis
Vector Database: Designed to store and search vector embeddings to help manage and retrieve similar property listings, documents, or market reports efficiently. It enables semantic search and AI-driven recommendations
LLM integration: Incorporated LLM into the system to interpret, summarize, or generate insights from property data. This allows users to ask natural-language questions such as “What are the top insurance opportunities in Miami?” and receive detailed, data-informed responses.
Custom property evaluation: Building tailored algorithms or models that calculate a property’s value, potential ROI, or investment risk based on user-defined criteria
Historical property analysis: Analyzing differences between housing properties
UX Journey Mapping:
Next Steps
What are the next steps for your project? How can you improve it?
Improve the web scraper
Add more price points
Improve the user interface design for the website