{"stable_id":"bd4dc96eedda0f2d","slug":"pennsylvania-employee-genai-pilot","company":"Commonwealth of Pennsylvania","workflow_name":"Employee-centered ChatGPT Enterprise pilot with labor participation","function_code":"engineering","pattern_codes":["creator_to_judge"],"changed_assumption":"Hinge: state employees can use a governed generative-AI workspace for drafting, research, summarization, brainstorming, and IT support while employees retain judgment and organized labor participates in adoption design.","evidence_strength":"reported","publication_tier":"showcase","freshness":"current","reviewed_at":"2026-08-23","updated_at":"2026-08-22","source_quality_summary":"1 primary; publication outcomes are reported.","caveat_summary":"Self-reported by a voluntary, non-representative pilot cohort; not independently time-tracked. Survey sentiment is not a service-delivery outcome.","freshness_caveat":null,"workflow_summary":{"before":[{"actor":"State employee","action":"Researches, drafts, summarizes, or troubleshoots using conventional tools and manual synthesis.","actor_type":"control"},{"actor":"Reviewer","action":"Checks the employee's completed work product.","actor_type":"human"}],"hinge":"Hinge: state employees can use a governed generative-AI workspace for drafting, research, summarization.","after":[{"actor":"State employee","action":"Uses the approved enterprise assistant to brainstorm, draft, simplify, summarize, research, or support IT work.","actor_type":"system"},{"actor":"State employee","action":"Evaluates, corrects, and owns the AI-assisted work product before normal review.","actor_type":"system"}],"decision_mode":"shared","decision_marker":"Human moves from creator to judge"},"before":[{"order":1,"actor":"State employee","action":"Researches, drafts, summarizes, or troubleshoots using conventional tools and manual synthesis.","handoff_to":"Reviewer or service recipient","control":"Normal agency review and records policies."},{"order":2,"actor":"Reviewer","action":"Checks the employee's completed work product.","handoff_to":"Agency workflow","control":"Human approval."}],"after":[{"order":1,"actor":"State employee","action":"Uses the approved enterprise assistant to brainstorm, draft, simplify, summarize, research, or support IT work.","handoff_to":"State employee","control":"Approved environment, training, usage policy, and pilot monitoring."},{"order":2,"actor":"State employee","action":"Evaluates, corrects, and owns the AI-assisted work product before normal review.","handoff_to":"Reviewer or service recipient","control":"Human expertise and judgment remain required; labor-management group informs adoption."}],"decision_rights":"Employees decide whether and how to use an output and retain final responsibility; agency governance determines permitted tools and uses.","exception_path":"Employees reject inaccurate or unsuitable output and complete the task conventionally; sensitive or disallowed work remains outside the tool.","removed_work":["First-draft composition","Initial synthesis of long material","Routine reformulation and brainstorming"],"outcomes":[{"metric":"Self-reported time saved on days of use","baseline":"Equivalent tasks completed without ChatGPT Enterprise assistance.","result":"Participants reported an average 95 minutes saved per day of use.","period":"Year-long pilot concluded March 2025.","scale":"175 employees across 14 agencies; 136 provided direct feedback.","attribution_caveat":"Self-reported by a voluntary, non-representative pilot cohort; not independently time-tracked.","evidence_label":"reported"},{"metric":"Participant experience","baseline":"48% of participants had never used ChatGPT before the pilot.","result":"85% reported a positive experience.","period":"Year-long pilot concluded March 2025.","scale":"175 employees across 14 agencies.","attribution_caveat":"Survey sentiment is not a service-delivery outcome.","evidence_label":"reported"}],"executive_lesson":"A labor-aware pilot can quantify augmentation while making employee judgment, not replacement, the operating control.","anti_pattern":"Treating self-reported pilot savings as audited productivity or removing normal human review.","questions_for_leaders":["Which workflows are permitted or prohibited?","How will output accuracy and service outcomes be measured beyond self-report?","How will employees and labor representatives shape scale-up?"],"collections":["human-still-decides"],"bundle_version":"1.0.0","bundle_fingerprint":"sha256:c23c6cc2b88153d008ea8fda928f632ce0011fc2d4c5036672a16e5d895bab93","canonical_url":"https://brianletort.ai/transformations/pennsylvania-employee-genai-pilot"}