Healthcare Tech Company:

The Client is a healthcare tech company from the US that offers a digital platform designed to optimize the coordination of healthcare services. Their goal is to bring true transparency to healthcare while enhancing efficiency, reducing administrative burdens, and improving patient care through innovative technological solutions.

The Client solution facilitates communication between healthcare providers, patients, and their families, improving the quality of care and patient satisfaction.

They work with various insurers, each relying on a different platform or software. Because of the Client’s scale, manually executing the process was out of the question.

The goal was to fully automate the crucial part of their Revenue Cycle Management - prior authorization process with various insurance providers. The process has different instances, depending on the providers’ systems and the company’s home state.

We met with the Company’s executives during the HLTH 2023 conference in Las Vegas. After a short talk with their CEO and CTO, they decided they wanted to engage us on the project.

Automation efforts were focused on a patient’s insurance verification workflows, which required connecting with various insurers’ middleware systems, Electronic Health Records, and NPI Registry.

Each automated workflow contains two main steps:

The bot is triggered by adding a new Patient Profile to the system. After that, the automated workflow is launched:

1. Checking the patient data in the NPI Registry and adding all the missing information or updating existing ones
2. Logging into the relevant insurer portal
3. Logging all the patient data, including the codes for the procedures and diagnoses
4. Setting and verifying statuses
5. Sending the submission
6. Printing the application as a digital .pdf file and attaching it to the Customer Profile in CRM and EHR.
7. Updating customer status in CRM and EHR.
8. Sending out alerts for cases that need manual processing

Follow-up bots entail obtaining the decision regarding a submitted prior authorization or medical review. This results in a final status (determined by the health plan), which could be rejected, approved, denied, not required, or partially approved of the prior authorization.

The bot works as follows:

1. Logging into the relevant insurer portal
2. Checking the pre-authorization status
3. Reading and transmitting the pre-authorization result to their systems
4. Sending out alerts for cases that need manual processing

For development, we decided to leverage UiPath, one of our favorite automation platforms. Each workflow has also been integrated with AWS to automate file storage.

From Day One, we have emphasized the importance of sturdy libraries and a cohesive strategy to simplify maintenance and facilitate any future modification requests.

Since the Client connects with various external systems, our goal was also to build an easily scalable and duplicable solution.

To achieve the maximal scalability, we:

1. Restructured the GUI library to make its entire Object Repository (UiPath’s equivalent of website elements library) simplified and organized in a way that minimizes the number of elements and mitigates the risk of duplicates.
2. Organized and developed tasks related to API interactions with AWS, public registries, password managers, and other software.
3. Developed a client-specific library that included activities tailored to the specific business requirements of that company.

After validating our approach with the first bot, with each subsequent order, we continued developing new solutions, extending relevant libraries, and gradually maintaining existing bots to implement the latest features.

Thanks to this strategy, developing each new automation required less effort, as an increasing number of activities and elements are already included in our library package:

• Automation 1 - 145h of work
• Automation 2 - 105h of work
• Automation 3 - 50h of work
• Automation 4 - 45h of work
• Automation 5 - 40h of work
• Automation 6 - 30h of work
• Automation 7 - 20h of work

What’s more, all bots can be easily updated and monitored.

So far, we have developed 13 RPA bots for the Client (ten from scratch and three rebuilt from the ground), automating their prior authorization process with different insurers.

Each of the delivered bots is similarly loaded. On average, a bot takes less than 3 minutes to process a single transaction fully, while manually, one transaction takes, on average, 10 minutes.

Not only was processing time cut down but no human involvement is needed anymore. This gives an estimated savings of 1 FTE/bot, translating into automating the work of 13 full-time employees.

The collaboration continues. As bot development time decreases rapidly, we continue adding new elements to the Client’s RPA infrastructure, increasing their automation ROI and further optimizing operational efficiency.