TAMU Hack - MISSO Discount Farmer

Designed a CRM system to help State Farm build and maintain relationships with customers. During development, we used Google Cloud Document AI, React, node.js, express.js, MongoDB, and javascript.

Comment

Inspiration

We aspire to create an efficient workflow to allow State Farm to better build and maintain customer relationships and reward customer positive home-owner behaviors.

What it does

It is a Web Application that can keep track of tasks given to customers in order to earn discounts from State Farm. Customers can complete tasks by uploading PDF documents as proof of completion. Documents are parsed by our AI back-end algorithm to populate Date, Price, Address, E-Mail and Website URL fields of a table. We hope that presenting all verification requests in a table will improve the workflow of the verifier so that customers may rewarded for their good behavior sooner. Once the verifier accepts the document. Customers will complete the task and earn the points the task was worth. When customers earn enough points, they may level up in tiers to earn discounts to their next bill. We believe that points and tiers will encourage customers to perform positive home-owner behaviors. The Web Application can also allow customers to schedule an annual home inspection with a single click of a button, which results in the completion of a task.

How we built it and how the AI works

Utilized React, JavaScript, CSS, HTML, and material-UI to design a Customer Relationship Management system. We built a back-end function to parse PDF invoices using Google Document AI, Open AI, and Cloud Storage. Documents are parsed into a string by Google Document AI. We feed the string into Open AI to parse the data into Date, Price, Address, E-Mail and Website URL. We then clean and normalize the output from Open AI using regex and post the data into MongoDB database as a JSON object. We connect to the database using Node.js and Express.js

Challenges we ran into

One of our team members pushed to the wrong repository a couple of times. Also, the environment variable was published by accident, causing our database crash and lost productivity. Moreover, another team member was having trouble implementing Date Picker, and it took a significant amount of time.

Accomplishments that we're proud of

Even though there were challenges that we ran into, however, our team was a solid and fantastic team that never gave up before finished. We all know each other better learn different languages, and we were able to put all of our progress together into a complete product to present to all judges.

What we learned

We all learned how to utilize third-party AI libraries better. We learned to minimize merge conflict by pushing the repository multiple times. Additionally, we learned not to set expectations high on third-party AI libraries if we are not familiar with them.

What's next for TAMU Hack - No Name

We will continue to learn and improve on areas that we are not familiar with and grow professionally and socially. Continue to participate in hackathons.

Built With

Share this project:

Updates