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.
Collections: Human still decides
Executive brief
A labor-aware pilot can quantify augmentation while making employee judgment, not replacement, the operating control.
AI value · Self-reported time saved on days of use
Company-reportedParticipants reported an average 95 minutes saved per day of use.
Self-reported by a voluntary, non-representative pilot cohort; not independently time-tracked.
Before
Researches, drafts, summarizes, or troubleshoots using conventional tools and manual synthesis. → Checks the employee's completed work product.
After
Uses the approved enterprise assistant to brainstorm, draft, simplify, summarize, research, or support IT work. → Evaluates, corrects, and owns the AI-assisted work product before normal review.
Human boundary
Employees decide whether and how to use an output and retain final responsibility; agency governance determines permitted tools and uses.
Why it matters
Hinge: state employees can use a governed generative-AI workspace for drafting, research, summarization.
This case is company-reported. Use it for the operating-model shift; do not treat the numbers as independently measured.
Before
How the work ran before the change.
State employee
Researches, drafts, summarizes, or troubleshoots using conventional tools and manual synthesis.
ControlNormal agency review and records policies.
Reviewer
Checks the employee's completed work product.
ControlHuman approval.
What changed
Hinge: state employees can use a governed generative-AI workspace for drafting, research, summarization.
Decision rightHuman moves from creator to judge
After
How the same work runs now.
State employee
Uses the approved enterprise assistant to brainstorm, draft, simplify, summarize, research, or support IT work.
State employee
Evaluates, corrects, and owns the AI-assisted work product before normal review.
Exception path
Employees reject inaccurate or unsuitable output and complete the task conventionally; sensitive or disallowed work remains outside the tool.
Decision authority
Employees decide whether and how to use an output and retain final responsibility; agency governance determines permitted tools and uses.
| # | Actor | Action | Control |
|---|---|---|---|
| 01 | State employee | Researches, drafts, summarizes, or troubleshoots using conventional tools and manual synthesis. | Normal agency review and records policies. |
| 02 | Reviewer | Checks the employee's completed work product. | Human approval. |
| # | Actor | Action | Control |
|---|---|---|---|
| 01 | State employee | Uses the approved enterprise assistant to brainstorm, draft, simplify, summarize, research, or support IT work. | Approved environment, training, usage policy, and pilot monitoring. |
| 02 | State employee | Evaluates, corrects, and owns the AI-assisted work product before normal review. | Human expertise and judgment remain required; labor-management group informs adoption. |
Work that left the path
Human role before
Employees manually produced administrative work products end to end.
Human role after
Employees frame tasks, verify outputs, apply agency context, and remain accountable for final work.
AI role
Decision mode: assistive. Generates drafts, summaries, explanations, research starting points, and troubleshooting suggestions; it does not approve or issue government decisions.
Equivalent tasks completed without ChatGPT Enterprise assistance.Participants reported an average 95 minutes saved per day of use.
Year-long pilot concluded March 2025. · 175 employees across 14 agencies; 136 provided direct feedback.
Self-reported by a voluntary, non-representative pilot cohort; not independently time-tracked.
48% of participants had never used ChatGPT before the pilot.85% reported a positive experience.
Year-long pilot concluded March 2025. · 175 employees across 14 agencies.
Survey sentiment is not a service-delivery outcome.
Anti-pattern
Treating self-reported pilot savings as audited productivity or removing normal human review.
Questions
Portability conditions
Reputation risk
medium