Document Processing · IT Operations Automation
Pennsylvania wanted to reduce time spent on drafting, research, summarization, IT support, procurement, contract analysis, and other administrative work across state agencies while keeping human oversight, transparency, and labor protections in place for government use of generative AI.
Reported impact
The pilot showed strong employee acceptance and measurable productivity gains, with faster drafting, summarization, hiring-related work, procurement, contract review, policy research, and software-upgrade support, while reinforcing that AI should augment rather than replace government workers.
The Commonwealth of Pennsylvania, through the Office of Administration, ran a year-long generative AI pilot that gave 175 employees in 14 agencies access to ChatGPT Enterprise for real government workflows. Staff used the service for writing, research, document summarization, IT support, contract analysis, policy research, stakeholder communications, software-upgrade work, and training-content development. The implementation was governed by Executive Order 2023-19 and oversight bodies created for responsible AI use, while Carnegie Mellon University researchers collaborated on evaluation. Rather than embedding AI into a back-end transaction system, the pilot served as a secure end-user productivity layer: employees brought work materials and prompts into ChatGPT Enterprise, received drafts or summaries back, and then applied human review before use. Adoption was supported with prompt training, biweekly feedback sessions, interviews, and live demonstrations to measure impact and improve rollout.
Employees in 14 Pennsylvania agencies entered prompts and work materials such as emails, long documents, contracts, policy questions, and software-support tasks into ChatGPT Enterprise. The system returned drafts, summaries, research assistance, and troubleshooting help to those employees, who reviewed outputs before using them in government workflows. OA then collected biweekly feedback, interviews, and live demonstrations to assess impact and refine adoption.
"AI-assisted work helped employees improve hiring timelines, streamline procurement, more quickly analyze contracts, and research new policies faster."
"Employees emphasized that AI augments human expertise rather than replacing it, reinforcing the importance of human review and judgment."
"Nearly half of the participants, 48 percent, had never used ChatGPT before, yet over 85 percent reported a 'somewhat positive' or 'very positive' experience"
"Over the course of a year, 175 employees from 14 agencies used ChatGPT Enterprise to test AI’s impact."
"employees had a highly positive experience, reporting an average time savings of 95 minutes per day while using ChatGPT for writing, research, summarization, and IT support."
"the pilot program provided state employees with access to ChatGPT Enterprise to enhance efficiency, streamline workflows, and improve the delivery of services for Pennsylvanians."
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Capgemini Sample Jun 9, 2026 Federal Ministry of Digitalization and State Modernization (BMDS) BMDS needed to reduce long, complex planning and approval procedures for infrastructure projects that routinely exceeded legal timelines and burdened case workers with documentation-heavy manual workflows. more than six months beyond the legally defined limit 
OpenAI Sample May 29, 2026 Boston Children's Hospital Boston Children's needed to reduce the administrative burden of supply-chain, billing, and scheduling work while also helping clinicians reason through rare-disease cases where fragmented genetic data, incomplete… About 60,000 hours in time savings co-pilot geneticist · enterprise AI layer · secure internal ChatGPT environment 
Mistral Sample May 27, 2026 HTX HTX needed AI that understood Singapore's public-safety context, could be trained on sensitive data inside a secure environment, and would remain under government control rather than depending on generic off-the-shelf… 123B-parameter natively multimodal and multilingual foundation model AI Forge · AI Studio · NGINE 