State Medicaid agencies are responsible for health services that touch the lives of every American, making the potential benefits of modern technology tools, such as artificial intelligence (AI) and machine learning (ML), enormous. However, despite the promise of increased effectiveness with the same or fewer resources, modernization, and digital transformation do not come easily – or naturally – to many of these agencies. Real and perceived fears in transitioning away from legacy, familiar tools and processes slow innovation and adoption. This can be particularly true today with AI, given the swirling and oftentimes conflicting rhetoric on its potential to help and harm. State Medicaid agency leaders can overcome these barriers to progress when they understand what thoughtful and selective AI implementation can look like — and what it can do for the people they serve:
Barrier 1: Fear of Over-Automation
One common apprehension among agency decision-makers is the fear of over-automation. How much control should be delegated to AI – and how fast? Handing full control of a system responsible for determining benefits and entitlements to AI could result in a worst-case scenario where resources are withheld from citizens in need. But the truth is that AI tools are not an all-or-nothing proposition—they can be customized to suit the specific needs of each agency. Often, AI systems will handle massive amounts of pre-screening of Medicaid applications, which can move to the top of the pile any cases that require a person to review.
Routine, time-intensive tasks, like verifying eligibility, can be automated, freeing up human workers to focus on more complex decision-making duties. In this way, AI can enhance human work, not replace it, striking a balance that offers both efficiency and the benefits of human judgment in health services. When AI is integrated into a state’s system in this way, it serves as a valuable partner rather than a replacement.
Barrier 2: Fear of Change and Making Mistakes
Change can be daunting, particularly in governmental operations tasked with something as sensitive as Americans’ health. Decision-makers are often worried about the potential fallout from mistakes made during the transition to new systems. And when politics come into play, the fear of a bad headline can sometimes tilt agency leaders toward being practical rather than pioneering. However, the adoption of AI and ML in Medicaid management and administration does not have to be a dramatic shift. With comprehensive training, proper change management strategies, and a culture that embraces data-driven decision-making, new technologies can be integrated smoothly.
In fact, agency leaders hoping to avoid the fallout from costly or embarrassing mistakes have all the more reason to lean on AI tools for support. With the capability to fully envision the big picture across an agency’s data, AI can help reduce the potential for human error, enhancing the accuracy of processes and increasing public trust. While risk aversion is natural, especially in the context of public service, adopting modern technologies like AI and ML can lead to improved processes and service delivery.
Barrier 3: Preference for More Familiar (Inefficient) Methods
Today, the pace of modernization does not afford organizations the luxury of the status quo. Many established operational processes and perspectives have been rendered inefficient. For example, hiring additional staff to manage data-intensive challenges (such as processing a surge in Medicaid applications) can be both expensive and less efficient. As the need grows, additional staff either needs to be trained and hired to accommodate it, or the system incurs frustrating delays.
AI and ML offer different paths. They can efficiently process and analyze large medical or demographic data sets, increasing accuracy and freeing up human resources for tasks that require a personal touch. The shift to AI and ML does not abandon tried-and-true methods but enhances them with powerful tools that can handle heavy data loads, leading to improved efficiency and cost savings.
Business, as usual, may seem like a risk-averse approach, but the hidden cost of ignoring the potential of AI and ML in Medicaid management and falling behind is high. Agencies that do not modernize risk inefficiencies and suboptimal service delivery to citizens. The COVID-19 pandemic, which led to a surge in applications, is a case in point. This kind of influx isn’t something agencies can instantly hire and train thousands of new workers to handle in positions that would be deprecated a year later. A system harnessing AI to process 90% of Medicaid approvals could have helped manage this influx more efficiently, saving time and resources.
While the concerns preventing state agencies from embracing modernization are valid, they should be weighed against the immense potential that AI and ML offer. With careful implementation and management, AI and ML can significantly enhance agency operations, leading to better service delivery. Navigating the challenges of public service in an increasingly digital world requires adaptability and the adoption of new technologies. With AI and ML, agencies can better handle the demands of public service, ultimately benefiting the citizens they serve. That’s what makes the proper involvement of AI and ML in state Medicaid agencies not just an opportunity for advancement—it’s a crucial step towards more efficient and effective public service.
About Victor Sterling
Victor Sterling is a Principal Industry Consultant at SAS. Prior to SAS, he was the Chief Information Officer of Arkansas Medicaid. SAS is the leader in business analytics software and services, and the largest independent vendor in the business intelligence market.
About Tim Taylor
Tim Taylor is a Principal Industry Consultant at global AI and analytics software provider SAS. He previously served as an Assistant Medicaid Director and Medicaid Management Information System Implementation Manager, as well as the Director of the Project Management Office, for Arkansas Medicaid.