New Jersey’s New Year’s resolution to tap AI for better service delivery
Eric Lee/Getty Images
At a time when states are pressed to innovate their public benefit systems, the Garden State is prioritizing expanding its AI tools under a new grant program.
New Jersey is heading into the new year with the goal of further innovating how the state delivers services to residents through artificial intelligence-based tools and training.
The state’s push for innovation comes as the federal government has called for states to implement major changes over the next year to their benefit eligibility and enrollment systems for programs like Medicaid and Supplemental Nutrition Assistance Program by Jan. 1, 2027.
“As more responsibility falls to the states, we need every technology at our disposal,” said Dave Cole, chief innovation officer of New Jersey, in an email to Route Fifty. “That includes AI, which, when used carefully and appropriately, can enable promising new ways to help deliver benefits more efficiently and effectively while not adding burdens for public professionals or those they serve.”
One of the state’s main priorities is to expand the use and capabilities of its AI assistant, which launched in 2024. The generative AI-powered platform has helped state staff draft and edit emails, summarize documents, process government feedback and complete other tasks, Cole said.
Since it launched, about 20% of New Jersey’s workforce has leveraged the AI assistant. The state also offers an AI training program that has amassed more than 13,000 participants and is now available in other states and localities across the U.S., he said.
Now, New Jersey aims to strengthen the tool to further drive the efficiency of and access to state services. Supporting those efforts is a share of an $8.5 million grant program from the nonprofit Center for Civic Futures, which Cole said “will help us scale our AI work across more agencies in New Jersey while also training our public professionals on new ways to use new tools effectively and responsibly.”
The grant program aims to help recipients test and share AI solutions with others in the public benefits ecosystem, state officials said in an announcement last week.
New Jersey will leverage funding, for instance, to “build out capabilities for more advanced use cases across government — such as validating, extracting, or analyzing data,” Cole said.
“We want to scale these applications, providing a platform of cross-cutting AI tools that can lift all boats across our state government and use AI to deliver better for even more New Jerseyans, and hopefully provide impact and value to other jurisdictions as well,” he said.
One potential tool state agencies will work on is an AI solution that immediately identifies and flags errors in benefit applications — such as incorrect or missing information — for applicants to fix, according to Cole. This AI use case could help reduce the time it takes for applications to be reviewed, ultimately alleviating administrative backlogs for agency staff.
The state is also looking to build upon previous work to improve data collaboration across agencies. AI-enabled tools could help “deduplicate and match data” in separate agency databases to identify residents eligible for certain benefits and streamline efforts to enroll them, Cole said.
New Jersey completed similar work earlier this year to deliver food assistance to nearly 100,000 children of low-income families who qualified for but were not enrolled in the benefit program, “and we want to replicate that kind of success,” he said.
To further help residents navigate government resources and services, state officials will explore how AI tools can help expand the language accessibility of web-based content.
Across New Jersey, residents have “varying literacy or … speak another language at home,” which could impact how they are able to comprehend and access resources meant to assist them, Cole said.
“Overall, we’re working across our state and the broader civic tech community to uncover new ways that AI can help us simultaneously address requirements, caseloads and capacity,” Cole said.