Developing insurance web app

As the first step, our business analyst and the client created a technical scope of work that outlined the inputs and outputs required for the project. Then, we chose core functionality features and prepared the necessary data. The team provided a timeline to ensure that we were on track to meet all of the deliverables.

To fulfill the client's business requirements, the team chose the following technical solutions:

• Parsers. Writing parsers to split a text document into data for databases

• Google Optical Character Recognition. Use Google OCR to convert PDFs and images into a text document for parsers

• Stripe. Implementing Stripe integration to accept payments

After that, our in-house designer prepared UI/UX for the web application. When the design was approved, our team started the MVP development.

Our development team created two parts – SaaS and the data module. The SaaS is for searching and comparing different insurance products in the U.S. insurance market or domain. The data preparation module creates a new database from the inconsistent data provided from CSV files, links database records to particular PDF documents, and populates form numbers (IDs) into databases.

The web application provides two primary services for users:

• searchable database

  The searching is free; however, viewing documents utilizes a pay-per-document approach.

• endorsement lookup

  A user pays per document processed and per document viewed.

During the implementation, we faced the following issue. By default, Google OCR handles PDF files for more than an hour. This timeframe was too long for our client.

Rubyroid Labs developers rewrote part of the logic that allowed for cutting PDFs into several pictures (according to the number of pages) and running recognition in multiple threads. This solution reduced time to around 5 to 10 minutes by showing the operation completion forecast (progress bar).

project phases

• MVP development

  Together with the client, we chose core functionality features and prepared the necessary data.

• ChatGPT integration

  After the release, a client wants to add new features. We integrated chatbot in the system to generates quick and relevant information about insurance products based on the client's proprietary database.

• software launch

  Creation of the UI/UX design for the website (landing), MVP adaptation to SaaS business requirements, monetization, launch.

• Gemini integration

  A Gemini-based tool assists in identifying potential coverage improvements that might be overlooked in the standard insurance coverage review process.

  By leveraging advanced analytics and machine learning, this tool can uncover gaps in coverage that traditional methods may miss, ensuring clients receive comprehensive protection.