Blab AI
New Jersey Codifies Innovation Authority, Leading Statewide AI Adoption
In January 2026, New Jersey gave its AI ambitions a legal footing by turning the New Jersey Innovation Authority (NJIA) from a grant‑funded office into a permanent state agency. The change followed the office’s growth during the pandemic and the state’s need for a clear, funded vehicle to design resident‑facing digital services, spot AI opportunities, set policy, and drive data‑driven outreach.
The NJIA’s structure goes beyond the usual chief technology officer. Dave Cole, the agency’s chief innovation officer, says it includes a resident‑experience team, a business‑experience team, a communications and engagement squad that applies performance‑based marketing to public benefit programs, and a data‑and‑policy unit. The design is intentional: technology deployment must align with outcomes that matter to residents.
Cole explained that the authority’s creation was driven by a legal gap. The office had outgrown the budget it received during the COVID surge, and the new law now allows the agency to receive state funds and, when appropriate, private grants. This financial footing is meant to sustain the expanding scope of AI work.
AI work began before the law. In 2024 the state launched the NJ AI Assistant, a generative‑AI chat interface available to all state employees via single sign‑on. Hosted on state infrastructure and powered by commercial models accessed through Azure and AWS, the assistant has already seen usage from about 20 % of the workforce and has generated more than 1.2 million prompts. Employees use it for drafting documents, brainstorming, and analyzing spreadsheets, and the platform is paired with a free training program developed with InnovateUS that has been adopted by over 25 other states.
Beyond the assistant, the NJIA has deployed AI in several program‑specific projects. A machine‑learning tool cleans data from state programs to identify students eligible for federally funded summer meals. By cross‑referencing enrollment lists from SNAP, TANF, and Medicaid, the tool eliminates duplicates and unlocks roughly $20 million in additional federal funding. The authority also built a computer‑vision system for the Economic Development Authority that scans PDF applications for small‑business programs, verifies required fields, and flags missing or incorrect information. The result was a reduction in the baseline review cycle from 21 days to near‑instant validation at upload, and the state reports a 20‑day reduction in processing time.
Other wins include a partnership with the Department of Labor to rewrite notice templates in plain language. AI‑generated drafts were reviewed by a designer, and the resulting notices improved response rates by about 30 %. The state has updated more than 100 templates in the unemployment insurance program.
Cole stressed that the NJIA deliberately avoids use cases that do not fit the state’s needs. The authority has not pursued public‑facing chatbots, arguing that poorly designed websites are better addressed through content and information architecture rather than a chatbot overlay. The focus remains on internal tools where staff can recognize and correct AI outputs.
The NJIA’s approach to AI policy is guided by a whole‑of‑government strategy and an AI Task Force that set principles: use generative AI to improve service delivery while safeguarding equity. The authority provides training that covers model bias, the need for human review, and accountability. The policy framework is designed to allow employees to experiment while maintaining public trust.
Cole noted that AI is not viewed as a job‑replacement tool in New Jersey. The state seeks to use AI to reduce repetitive work, improve quality of life for employees, and expand capacity where hiring is difficult. Discussions within the agency include skepticism and cautious optimism, but the overarching stance is that AI should augment, not replace, human work.
The authority’s experience offers lessons for other states. Cole said that a dedicated, well‑funded team that understands both technology and the state’s IT ecosystem is essential before launching AI initiatives. Leadership alignment, as seen with Governor Sherrill’s focus on efficiency and outcomes, also facilitates experimentation.
The NJIA’s work is part of a broader national trend. New York has taken a cautious stance, while cities like Jersey City have begun to deploy public‑interest technology crews to improve digital services. New Jersey’s formalization of its innovation authority positions it as a model for states seeking to integrate AI responsibly into public service.
In summary, New Jersey’s codification of the Innovation Authority and its rollout of the NJ AI Assistant mark a significant step in state AI adoption. The agency’s data‑driven, outcome‑focused approach has produced tangible improvements in program delivery and offers a framework that other states can adapt as they navigate the benefits and risks of generative AI.
The NJIA’s structure goes beyond the usual chief technology officer. Dave Cole, the agency’s chief innovation officer, says it includes a resident‑experience team, a business‑experience team, a communications and engagement squad that applies performance‑based marketing to public benefit programs, and a data‑and‑policy unit. The design is intentional: technology deployment must align with outcomes that matter to residents.
Cole explained that the authority’s creation was driven by a legal gap. The office had outgrown the budget it received during the COVID surge, and the new law now allows the agency to receive state funds and, when appropriate, private grants. This financial footing is meant to sustain the expanding scope of AI work.
AI work began before the law. In 2024 the state launched the NJ AI Assistant, a generative‑AI chat interface available to all state employees via single sign‑on. Hosted on state infrastructure and powered by commercial models accessed through Azure and AWS, the assistant has already seen usage from about 20 % of the workforce and has generated more than 1.2 million prompts. Employees use it for drafting documents, brainstorming, and analyzing spreadsheets, and the platform is paired with a free training program developed with InnovateUS that has been adopted by over 25 other states.
Beyond the assistant, the NJIA has deployed AI in several program‑specific projects. A machine‑learning tool cleans data from state programs to identify students eligible for federally funded summer meals. By cross‑referencing enrollment lists from SNAP, TANF, and Medicaid, the tool eliminates duplicates and unlocks roughly $20 million in additional federal funding. The authority also built a computer‑vision system for the Economic Development Authority that scans PDF applications for small‑business programs, verifies required fields, and flags missing or incorrect information. The result was a reduction in the baseline review cycle from 21 days to near‑instant validation at upload, and the state reports a 20‑day reduction in processing time.
Other wins include a partnership with the Department of Labor to rewrite notice templates in plain language. AI‑generated drafts were reviewed by a designer, and the resulting notices improved response rates by about 30 %. The state has updated more than 100 templates in the unemployment insurance program.
Cole stressed that the NJIA deliberately avoids use cases that do not fit the state’s needs. The authority has not pursued public‑facing chatbots, arguing that poorly designed websites are better addressed through content and information architecture rather than a chatbot overlay. The focus remains on internal tools where staff can recognize and correct AI outputs.
The NJIA’s approach to AI policy is guided by a whole‑of‑government strategy and an AI Task Force that set principles: use generative AI to improve service delivery while safeguarding equity. The authority provides training that covers model bias, the need for human review, and accountability. The policy framework is designed to allow employees to experiment while maintaining public trust.
Cole noted that AI is not viewed as a job‑replacement tool in New Jersey. The state seeks to use AI to reduce repetitive work, improve quality of life for employees, and expand capacity where hiring is difficult. Discussions within the agency include skepticism and cautious optimism, but the overarching stance is that AI should augment, not replace, human work.
The authority’s experience offers lessons for other states. Cole said that a dedicated, well‑funded team that understands both technology and the state’s IT ecosystem is essential before launching AI initiatives. Leadership alignment, as seen with Governor Sherrill’s focus on efficiency and outcomes, also facilitates experimentation.
The NJIA’s work is part of a broader national trend. New York has taken a cautious stance, while cities like Jersey City have begun to deploy public‑interest technology crews to improve digital services. New Jersey’s formalization of its innovation authority positions it as a model for states seeking to integrate AI responsibly into public service.
In summary, New Jersey’s codification of the Innovation Authority and its rollout of the NJ AI Assistant mark a significant step in state AI adoption. The agency’s data‑driven, outcome‑focused approach has produced tangible improvements in program delivery and offers a framework that other states can adapt as they navigate the benefits and risks of generative AI.
