Insurify

Your one-stop shop for all things Car Insurance.

Comment

Inspiration

While having car insurance is essential in many of our lives, it is often difficult for us to understand why we pay a certain amount, or why we even have to pay at all. This is especially true for college students, as learning about car insurance is a difficult step in the path to adulthood. We also were aware of the higher number of car accidents involving students, and especially how difficult it might be for a student involved in a car accident to realize what happened, understand what they may have to do next or understand their finances involving insurance going forward. We aimed to tackle these problems by creating a website to simplify and solve many of these problems.

What it does

Insurify is a multi-purpose insurance assistance tool, which provides a plethora of conveniences for its users but is especially targeted towards college students. Insurify allows for simple and seamless damage assessment, utilizing image processing and machine learning to assess car damages from uploaded images. Insurify will detect damage in an image, classify it based on severity, and give you the estimated cost of repair, future premiums, and claim amount. Insurify also has educational content, which allows users to educate themselves and get started on learning more about insurance, how it can benefit them, and how they can understand their finances relating to it. We have also employed a chatbot for answering any insurance-related questions.

How we built it

Insurify's front-end was built with Next.js, React, TypeScript, Tailwind CSS, and DaisyUI. Our authentication was handled by Clerk, with a PostgreSQL database. Our backend was built with Python and Flask, as it was easiest to keep it all the same language. Our ML model leverages the Transformers library, and we used GPT-4 to enhance our model's results. All the ML functionality was built into an API which was leveraged by the frontend.

Challenges we ran into

One of our biggest challenges was finding accurate data on car damage. A lot of the data we needed had to be manually classified, which gave us a bit of trouble due to the time constraint. The scope of the project was also difficult, as we struggled to balance many of the features we wanted to provide while keeping the project manageable for our skill level. We also struggled extensively with Docker and containerizing our Flask app for deployment.

Accomplishments that we're proud of

We're proud of completing a fully functioning product, with all the kinks generally ironed out, and deployed. We successfully created a platform to classify damage to vehicles and provide a quote simply based off that, and a few more parameters. We also derived a system to educate others on car accidents, insurance, etc., and we're all very proud of that as well.

What we learned

  • The importance of proper planning and project management.
  • Effective techniques to divide tasks among a team.
  • Optimization of machine learning models and LLMs for greater results.

What's next for Insurify

We have a ton of ideas going forward! Some goals include full automation of the insurance claim process, real-time accident assistance for customers, and greater accuracy of our image classification model.

Share this project:

Updates