The answers to questions about the state’s Medicaid program, MO HealthNet, could change soon as the state prepares to roll out an artificial intelligence program to accommodate changes to the program.
As part of President Donald Trump’s budget proposal, known as the “One Big Beautiful Bill Act,” Medicaid enrollees between the ages of 19 and 64 would be required to certify 80 hours of work, school or caregiving each month, with exceptions for parents of children under a certain age and parents with disabilities.
The budget bill would force the biggest changes to Medicaid in years, and states are rushing to implement new technology to keep up with the overhaul.
takeout
- Missouri is one of six states planning to leverage artificial intelligence to manage new Medicaid work requirements.
- State officials say AI could help answer registration questions, process paperwork and automate parts of the MO HealthNet system.
- Experts say the technology could reduce paperwork confusion, but warn that testing and monitoring will be important to ensure people don’t lose insurance unnecessarily.
Missouri was one of just six states in a recent survey that reported plans to leverage AI to handle the operational requirements of their state’s Medicaid program.
It can manifest itself in different ways. Missouri said it plans to use AI to interact with customers, process documents, collate data, provide additional staff support and increase program automation.
But most of Missouri’s 1.26 million Medicaid recipients are already working or eligible for a waiver. A recent analysis of 2024 U.S. Census data found that 67% of working-age MO HealthNet patients are employed, while most others are disabled, in caregiving, or in school.
The work requirements are expected to take effect in December. It’s unclear exactly how the state plans to implement AI to handle the additional workload, but as part of the new law, Missouri lawmakers approved spending nearly $50 million in state and federal funds for the state’s software updates and other new mandates.
While questions remain about how the technology will be used, experts say states face significant challenges as they prepare for the change, especially work requirements.
Why is Missouri relying on AI to qualify for Medicaid?
For years, Missouri has had difficulty administering Medicaid effectively.
The state has been plagued by staff turnover and outdated technology, which has increased wait times for everything from time spent answering phone questions to backlogs of registration applications.
One of the main problems is that many of the state’s computer programs can’t communicate with each other, so it takes time to verify someone’s data, such as income or hours worked.
“Missouri is kind of behind the curve in most technologies,” said Elias Tsapelas, director of state budget and fiscal policy at the libertarian Show-Me Institute.
He said the federal government appears to be pushing states to use electronic verification wherever possible.
“The government wants states to make these decisions as electronically as possible, essentially using the data they already have access to,” Tsapelas said. “So this is not a huge (operation) where the state is sending out a bunch of letters and asking for a bunch of new documents.”
The Missouri Department of Social Services, which oversees Medicaid, has requested $6.2 million in state and federal funds to build or contract a new program to handle all the information.
The proposed federal budget would require Medicaid patients to go through the redetermination process twice a year instead of once. Ideally, Missouri would be able to cross-reference SNAP records, income databases and employment information with the goal of reducing the number of people who have to manually file forms, said Brian Colby, vice president of public policy for the left-leaning Missouri Budget Project.
“We want the transaction to go as smoothly as possible,” Colby said.
DSS also requested more than 200 additional full-time employees to keep up with the changes set for Missouri in the coming years.
“Their workload has essentially doubled,” Colby said.
DSS emphasized that the AI tool does not make enrollment or eligibility decisions.
“No AI tool or algorithm will make an adverse eligibility determination or coverage termination decision,” DSS spokeswoman Bailey Watts wrote in an email to the Beacon. “The system’s role is strictly limited to data aggregation, such as identifying missing documents and summarizing uploaded files.”
“If automated data matching confirms that an individual meets established work requirements, the system can process that positive confirmation to ensure uninterrupted continuity of care,” Watts said, adding that if someone’s status cannot be automatically verified, state officials will handle the case.
Watts said if the AI determines someone is ineligible, state officials would also need to review that decision.
“State officials always review complete files, evaluate system-generated summaries, and make final decisions before any adverse action is taken,” she wrote.
But studies of human behavior using AI models show that humans are quick to follow AI suggestions, said Elizabeth Edwards, senior attorney at the National Health Law Program.
“What the research shows is that computers are right most of the time, so eventually humans will start relying on them too,” Edwards said. “They’re going to stop looking at themselves so closely. … The idea that humans are in the loop and solving all kinds of AI surveillance problems is not really good enough.”
How will AI be used?
Missouri has listed several plans to use AI. But experts hope new technology will bring much-needed improvements to the state’s Medicaid system.
According to the budget request, the state is already working to improve its chatbot software, which allows people to instantly get answers to basic questions, including details about applications and cases, without having to speak to a DSS official. The department says Missourians will be able to report changes to their households and fill out applications and updates without having to pick up the phone.
“A lot of times, the reasons why people call the state (with questions) are not that difficult,” Tsapelas said. “It’s things like: ‘How’s my application going? What do you need from me?’ These are the kinds of things where AI could be very helpful.”
The state said it plans to leverage AI to increase back-end automation, which Edwards said could be something like easily writing policy or federal Medicaid rules to help human staff tackle more complex cases.
“When you have a 300-page manual, it’s difficult for workers to find out what the policy is,” Edwards said. “Training an AI on a world of such documents may still introduce errors, but it can be trained, tested, and monitored so workers can find the right policy.”
But Edwards is also concerned about the potential for mistakes. That could mean the difference between keeping Medicaid insurance or not.
“What we’re seeing with a lot of AI products, at least some of them, is that not all of them fully understand Medicaid requirements,” Edwards said. “What’s really consistently important is that we’re always concerned about how it’s being monitored and tested.”
The countries expressed their concerns in a survey conducted by KFF earlier this year.
Many states shared concerns about the short timeline and inability to test new software before work requirements begin. One state said solutions offered by vendors planning to use AI would likely not be able to verify the eligibility of gig workers, medically fragile enrollees, enrollees who may be homeless, etc.
Missouri has already partnered with an AI-powered program called SteadyIQ, which allows enrollees to verify their income by linking their bank accounts to gig work programs such as rideshares.
Beyond SteadyIQ, DSS told the Beacon that Missouri plans to finalize contracts with new vendors and integrate them into the state’s existing eligibility system.
Countries are turning their attention to technology while competing with each other moment by moment.
Many states reported that they had to begin making decisions about how to address work requirements without additional guidance from the federal government because of the tight timelines Congress set in their budget proposals.
Some critics point to the development of the Affordable Care Act market and the need for states to meet the requirements, even though it took more than three years to prepare. Officials in states such as Oregon warned at the time that the software was not ready for market deployment and that residents were initially unable to successfully register online.
“If you look at history, you see that timelines really matter,” Edwards said. “And even with a longer schedule, problems can still occur.”
Tracking work requirements has also proven difficult for states to master. In Arkansas, more than 18,000 adults lost Medicaid coverage during the brief period that work requirements were in place before a court halted the program.
Researchers later found that the policy did little to increase employment, and the study found that coverage losses disproportionately affected Arkansans without home internet access. In 2025, Missouri ranked 30th in the nation for residents’ broadband access.
While AI programs have the potential to significantly improve the user experience for both enrollees and state employees, Tsapelas cautioned that the costs of new technology can quickly add up.
“I think AI has the potential to improve things for enrollees,” Tsapelas said. “As with everything, we have to pay attention to Medicaid, because when Medicaid makes mistakes, it’s very costly.”
For now, Missouri is one of the few states planning how to leverage AI to address the changes that come with the new federal law.
It remains to be seen how exactly the tool will be used and whether it will facilitate the process.
“Improved data matching would significantly reduce the workload,” Edwards said. “But they have to build it all, test it, and make sure it works.”
