Lessons from Pennsylvania's Generative AI Pilot with Chat GPT

The Commonwealth launched a ChatGPT Enterprise pilot to explore how generative AI could support and enhance public sector work. Over the course of the pilot, more than 175 employees gained access to ChatGPT, with 136 providing direct feedback on their experience. Participants represented a diverse cross-section of the workforce, varying in age, job role, seniority, union status, and prior exposure to AI tools—48% of whom had never used ChatGPT before.

Despite a wide range of use patterns, 85% of participants reported a positive experience. Many saw meaningful improvements in both efficiency and quality of their work, with some estimating time savings of up to 95 minutes per day. Importantly, ChatGPT was not seen as a replacement for human expertise, but as a valuable tool to enhance existing capabilities. Success with the tool depended more on users' ability to navigate adoption barriers—such as accuracy concerns, habit formation, and privacy questions—than on any specific job profile.

Top 5 Pilot Findings:

1. The majority of users described their experience with ChatGPT as “very positive.”

2. On average, participants estimated saving 95 minutes of work per day.

3. ChatGPT amplified—but did not replace—the expertise of public sector employees.

4. No single “superuser” profile emerged; adoption success varied across roles and individuals.

5. ChatGPT proved especially valuable in addressing complex, policy-driven, and communication-heavy challenges unique to government work.

Read on to learn more about the background, use cases and experiences and the challenges and barries to this pilot.