screencap for what PolicyWallet is best for
UI Mockups for the iOS App's use cases
Background, Problems Identified
In the United States, consumers spend on average about $998 Billion dollars per year on their lifestyle. ⁱ
The purchasing power of young adults in this market is 34% of that amount. Now more than ever, this segment is regularly using the Internet to shop for new products, home & travel deals. But the challenge we’re facing with these consumers is that they are buying up uninsured products worth thousands of dollars, because they don't trust the monolithic structure of insurance at all.
We're going to change that.
By combining the power of predictive AI, Computer Vision and the advanced APIs Solartis offers, PolicyWallet can change the way consumers ensure their life smartly with the use of mobile product catalogs used to secure insurance policies from various organizations.
What I Solve For Organizations & Consumers
It's like eBay for digital insurance.
PolicyWallet stands out from the competition because as a smart product insurance platform, it is not only servicing insurance organizations but providing a kick-ass solution for their potential customers as well!
Consumers take pictures of their gadgets and appliances which is then cataloged. This product catalog provides unique visual data insights that allow them to filter through dozens of policy providers' offerings in an open marketplace interface to find what works best for their home/lifestyle, save their policies into one centralized wallet.
This will allow Organizations to offer unique policies to individual users based on their item catalogs. This empowers organizations because the platform is connecting young adults directly to policy administrators as well as providing consumers with options to ensure their lifestyles.
How I Built It & Why I Choose My Software Stack
The website is built on CoreML/Vision, Swift & Python scripts for iOS.
It provides the following functionality:
- Wallet dashboard interface for all the policies you hold
- Predictive analytics for image-based item cataloging and value estimation for ratings/quotes
...thanks to libraries and technologies like:
- TensorFlow (https://github.com/tensorflow/tensorflow)
- AlamoFire (https://github.com/Alamofire/Alamofire)
As a developer, I can confirm that while the code is powerful and efficient, the platform has a simple interface that is easy for non-tech users to use recurrently.
Challenges I Ran Into
I ran into a couple of problems, but the most notable was that I didn't have time enough to make both a functional website & native iOS app for PolicyWallet, so I settled for making the iOS app instead. I also decided to make the app consumer-facing but to highlight the future power it held for organizations to make smart decisions based on data.
I couldn't train a fully functioning camera model in time for the hackathon, so I settled for a basic machine learning model that only recognized the objects around my house to generate a boolean value for the protection class in the Renter Program's Rating API JSON request.
Thus, I consider this project more of a concept for what the future Solartis PolicyWallet app could be, how it might look, and how the API could be integrated.
Software Accomplishments That I'm Proud Of
Inevitably, by running into all the aforementioned troubles, I was able to come up with some pretty dope solutions and innovative ways of creating business logic for PW.
Here's some of them:
I successfully trained a machine learning model to recognize my common household objects, tested on iOS CoreML and configured the app to take the data fed from the camera/model to the item generation module for the PolicyWallet catalog. This included its value and model manufacturing date.
I created a slick interface for the consumers that will later benefit policy administration/management that mostly complemented the consumer while collecting data intelligence for the organization offering these ratings. This will allow insurance companies to come onto the platform much quicker and provide an incentive for them to do so.
It was pretty difficult as a junior developer/designer but I think all of this added a sweet touch in the end!
What I Did Differently, and Learned
After receiving the sandbox email & discussion posts from Carol, I realized that I didn't want to just hack together any "one" solution for consumers or businesses. I didn't want to throw buzzwords into the infrastructure for the sake of saying I used them, either.
I decided to create an MVP for a platform that Solartis could build, that would be a true innovation in #InsurTech for both consumers AND businesses.
I learned how to work with the branding of an idea to match the software direction of a project and bring together two opposing problems (organizations, consumers) to one centralized solution for a distributed platform.
What's Next: Future Improvements
I hope to add a number of improvements to PolicyWallet in the near future:
- Opening up the platform for policy organizations to join with new integrations, features, etc.
- Provide better insights and perhaps a physical camera that monitors homes for data intelligence in future claims
- Make it cross-platform for both the Web and Android & further the use of the Solartis API in the policy process
- Make it fully functional with the camera AI and more Developer Sandbox API integrations
The objective for this hackathon project has always been to change the way consumers ensure their belongings/travel, and ultimately their lifestyle, as well as empowering organizations to easily manage existing policies and administer new ones.
I hope that this will help PolicyWallet stand out. Thank you for the opportunity!