GroveGuard

The current uploaded home insurance claim
Claimant's home telematics data and previous claims
Grover's claim analysis

Inspiration

53% of insurance fraud comes from home insurance fraud, whether it is from filing multiple claims under the same incident, or claiming an incident happened when it did not. This program helps both State Farm and its many clients save time and money by preventing home insurance fraud.

What it does

Temperature, humidity, and motion sensors are placed around the home, and telematics is used to collect and send the data to a user interface website that insurance agents have access to. The website displays the data from the sensors, and when a claim is submitted by a client, it uses generative AI to compare the data from the sensors with the claim. It also checks if any similar claims were previously submitted by the same account. Then, it determines whether insurance fraud was attempted, and provides insights on its reasoning.

How we built it

The sensors were built and coded using the Arduino IDE, the website's front end was built using HTML and CSS, and the generative AI model was built using Python, LangChain, and an Open AI API.

Challenges we ran into

It was challenging to create a PDF dropbox in the website's front end, but we were able to get it done by researching more about HTML. We also ran into issues implementing Flask and LangChain.

Accomplishments that we're proud of

We have managed to finish the generative AI algorithms, and create a device that can collect and analyze data.

What we learned

We learned how to implement Flask, LangChain, Open AI, and React.js.

What's next for GroveGuard

We will next make the AI model reach out to the client automatedly if insurance fraud is suspected, and put them in contact with an agent to discuss the next steps.