In collaboration with:
This is the story of how Persistent transformed the support to victims of natural catastrophes with a Disaster Housing Assistant app built with OutSystems, artificial intelligence, machine learning, and blockchain.
The Times They Are a-Changing: Defining the Problem
Over the past decade, we’ve been witnessing an increase in the frequency and severity of natural disasters. In 2017, solely in the US, homes were devastated by the violence of Irma, Maria, Harvey, and tornados—to name a few—leaving behind tears of destruction and changing lives forever.
When catastrophes like these happen, state agencies immediately jump in to help the victims get their lives and homes back. But that’s usually a long and complex process. The agencies have to deal with several challenges, such as:
- Identifying intelligent ways to automate the eligibility of the application, determine the benefit, assess the damage, and the construction and closeout processes
- Avoiding the disbursement of duplicate funding for the same need because of the number of different agencies involved. Funds need to be used for their intended purpose. Duplication can cause the accounting of recovery dollars earmarked for repair or reconstruction to become complex or even incomplete.
In summary, it is a laborious and paper-intensive process.
Now, imagine a person who has just had serious personal losses. To get the help of these agencies, that person needs to get in a big line to fill out an application. The agency puts a team together to go analyze the person’s situation—what was damaged and with what severity—then introduce all that information in the system, evaluate the amount of funding that person is entitled to, and only then does the applicant receive the help that he or she desperately needs.
So, at Persistent, we started wondering. How could we simplify this process to minimize the pain that these people and families are going through? There were a few things we knew we needed to do:
- Create the means to make sure the funding disbursement for property repairs were being used appropriately.
- Speed up the delivery of funding to homeowners by building accounting and verification processes.
- Provide assurance to the funding agencies that the funds are spent according to their regulations.
Low-Code Meets Artificial Intelligence: Introducing Persie
Our vision was to accelerate the overall process by automating as many steps, from registration to the delivery of funds, as possible. So, we built an artificial intelligence bot, a friendly robot we named Persie.
Persie was developed to assist and simplify the registration process and claims submission. When a disaster happens, you can have several of these robots in the local area so that people can get immediate service.
They can talk to Persie and show their identification. Persie scans it, and it is used to register a person with the agency automatically.
An account is then generated for the applicant. With these credentials, the applicant can log in the Disaster Housing Assistant app to complete and submit a claim by providing additional information.
To build this app, we used the OutSystems low-code platform because it enables us to give the applicant the flexibility to access the app from a desktop, tablet, smartphone, or email. In addition, thanks to the integration with DocuSign, people can sign the declaration online, which saves them time and eliminates the hassle of going to the agency.
Heavy Machine-Learning Fire
Accelerating registration was only part of the challenge. The rest of the process was still too complex and time-consuming. Lots of property documentation had to be submitted, reviewed, and assessed. Then, the agency needed to identify the right inspector to go evaluate the property damages. And only then could the agency define the amount each person was entitled to.
To overcome this challenge, we decided to apply machine-learning algorithms using IBM Watson to review the information that Persie collected and that the applicant provided to be evaluated. This included confirming that the person is the rightful owner of the property, determining if there’s any document that has expired, establishing legitimacy, or identifying if there’s any information missing.
As a result, the agency collects and evaluates all the information really quickly and speeds up the process of sending the right inspector to the site to assess the situation. Here, we also used technology like geofencing to find the inspectors closer to the property and assign the appropriate activities to them based on their volume of work.
Once the inspector is on the site, he or she can take pictures of the property damages and then submit them to the IBM Watson machine-learning models. And, almost automatically, this system identifies the level of damage to the property, what needs to be replaced, and the estimated cost of the fix.
But there was still one more critical problem we needed to solve.
Adding Some Blockchain to the Mix
The state agencies spend billions of dollars on these catastrophic events to help people rebuild their lives. Now, the challenge is making sure the funds are being used in the right way. In other words, determining that the people who received the funds are actually using them to fix and rebuild their homes, and not to go to the grocery store or restaurants for example.
We knew that the solution had to go through the use of a restrictive currency so that the person could only use it to fix the home damages. In other words, we brought in blockchain.
So, if the person is eligible, he or she gets the funds in the form of cryptocurrency and cannot convert it to any other currency. When any transaction is made using the cryptocurrency, it gets recorded and goes back to the agency so it can review where the funds were used.
The Power of Low-Code and the Art of the Possible
OutSystems played a central role in the orchestration of the entire process. All these different features, from AI and machine-learning to the blockchain integration, are connected in the DHA application built with OutSystems.
This way, we were able to bring various new edge technologies together to solve a complex problem and deliver a new generation of customer experiences, along with pixel-perfect business automation.