According to a new report from North Carolina’s ChatGPT pilot with Open AI, AI saved workers an average of 30 to 60 minutes per day, with 85 percent of participants reporting a positive experience.
In the video below, state employees who participated in the pilot share their experiences.
INSIDE NORTH CAROLINA’S CHATGPT EXPERIMENT
The three-month pilot was conducted with the Department of State Treasurer, where an enterprise ChatGPT tool was offered to employees in the Unclaimed Property and State and Local Government Finance offices from March to June 2025.
The independent study, conducted by the Institute for Artificial Intelligence and Emerging Research at North Carolina Central University, found that time savings increased as users became more proficient. Initial surveys showed users saving 15 to 30 minutes daily, which rose to 30 to 60-plus minutes by the final survey.
“Adopting this innovative technology has helped us deliver improved results to our constituents and to taxpayers,” Treasurer Brad Briner said in a press release. “This important analysis clearly shows that adding the power and speed of artificial intelligence to the talent, experience and judgment of our state employees is the key to unlocking greater workplace achievements.”
HOW STATE WORKERS USED CHATGPT
Key ways ChatGPT was used included drafting professional communications, reports and memos; translating technical documentation into plain language; brainstorming content for policy documents, training materials and outreach or summarizing legal text; multipage reports; and public submissions.
Workers also used it as a research tool to ask clarifying questions when researching complex or unfamiliar topics.
Participants were given ChatGPT use guidelines and received prompt engineering training led by OpenAI. However, survey responses revealed that some users' confidence could have been improved by role-specific prompts and guided scenarios.
According to the study, the best results came when users collaborated with the tool and refined outputs together.
CHALLENGES AND ROADBLOCKS OF THE PILOT
Notably, the pilot reinforced that ChatGPT couldn’t replace human judgment, as users had to combine their own expertise and judgement for best results.
There were roadblocks with using artificial intelligence. The biggest functionality challenge users noted was that there were occasional inaccuracies, with ChatGPT sometimes generating incorrect statements or citations. Users reported the tool underperformed in tasks requiring coding, legal precision or complex math.
Accuracy concerns, the learning curve to using the tool effectively and questions about privacy and what could safely be entered into the tool where the biggest roadblocks that kept users from using AI more frequently.
OTHER STATES ARE TESTING, TOO
Pennsylvania also concluded a pilot study with Open AI earlier this year. In that case, a yearlong pilot with the state’s Office of Administration, more than 175 employees used ChatGPT Enterprise. The results: 85 percent of participants had a positive experience and reported saving 95 minutes per day with ChatGPT.
