Virginia-based Maximus has launched Accuracy Assistant.
The company, which sells tools for case management, cybersecurity and health-care program eligibility, says its newest product focuses on state governments and the Supplemental Nutrition Assistance Program, or SNAP.
A statement from the company describes the new AI-backed offering as using “predictive analytics and intelligent automation to detect data inconsistencies and flag accuracy issues.”
That, in turn, can give states a deeper, real-time look at program files to spot inconsistencies in data that can result in those costly errors.
“From the moment of the first interview to the end of the eligibility process, Accuracy Assistant can proactively identify and address potential issues in SNAP applications, strengthening program integrity and improving accuracy at every step,” said Bruce Caswell, the company’s president and CEO, in the statement.
SNAP has long attracted political controversy, and now states face higher costs to help people buy food in the wake of President Donald Trump’s One Big Beautiful Bill.
New work requirements set to kick in Feb. 1 have left recipients and food program operators both confused and pessimistic about what will happen. They worry how the loss of federal support will impact the work of getting food to people who have a hard time affording groceries.
The federal government spent about $100 billion on SNAP in fiscal year 2024, with monthly benefits averaging about $187 per person, according to federal figures. The program served an average of 41.7 million people each month in fiscal year 2024.
The federal government has already said it will withhold $130 million in funding for Minnesota’s food stamp program because of alleged fraud.
In neighboring Wisconsin, state officials have estimated that Trump’s cuts to SNAP could cost the state $314 million annually and leave 90,000 residents without help from the program.
The Maximus pitch comes down to this: As those cuts make their way through the system, states cannot afford to cover payment errors — that is, according to the statement, the “rate [that] measures how accurately states determine SNAP eligibility and benefit amounts.”
States that have error rates of more than 6 percent as of October will have to pay “a progressively higher share of the SNAP food benefit costs,” Maximus said.
The company’s new tool identifies cases at high risk of pushing up that error rate.
State staff then can work on those cases without slowing down approval of cases that present lower risk of errors. An interactive dashboard helps clients keep track of the process and access metrics.
“States are committed to delivering SNAP benefits accurately and on time, but often face limited technology and resources,” said Caswell.
Some states have already taken their own action to keep SNAP benefits free of error and potential fraud. Missouri, for instance, used a federal grant to build a fraud detection system in-house — a system that has saved the state at least $2 million in fraudulent claims.
The nonprofit Code for America, meanwhile, last year launched resources for states to help them navigate ongoing changes and cuts to SNAP and Medicaid.
