Kuitun City Brain
A social-security government-affairs assistant — let citizens talk to policy, keep sensitive data controlled.
Capability map
Evidence
Industry
Geography
Trust features
Let citizens talk to policy, keep sensitive data controlled — public serves all
Executive Summary
The Kuitun City Brain is a social-security government-affairs assistant for the city of Kuitun: citizens ask in one sentence about social security, medical insurance, retirement, work-injury, unemployment or maternity policy and get the process spelled out; service agents handle 12319 citizen requests efficiently. It precipitates policy items and years of cases from many departments into a governed government-affairs knowledge graph — answers are grounded in official documents, not made up by AI.
Evidence Snapshot
Real-World Signals
Organization
Kuitun City Brain
Pilot
Why It Matters
Public services have zero tolerance for error, yet citizens often don't know which department, what materials, or which process. The value of government AI is letting citizens talk to policy directly and letting agents do more with less legwork — and all of it rests on governance: public policy serves everyone, sensitive cases stay controlled.
Business Challenge
- Many city departments each own a slice; policy definitions are scattered
- Citizens calling the service hotline need cross-department lookups — slow and inconsistent
- Service guides and historical cases are scattered; no unified knowledge entry
- Public sector has zero tolerance for error, and citizen privacy is involved
Solution Architecture
Understand · Jinni
Citizens ask in natural language — 'how do I transfer my pension', 'which department handles a burst pipe' — and AI answers on target.
Remember · Knowledge Graph
Hundreds of policy items and years of citizen cases, structured into an item–materials–conditions–department knowledge graph; government-affairs memory in one place.
Govern · IMAI (RBAC)
Role-based access control (to-G, by sensitivity): public policy visible to all, sensitive cases only for authorized agents and leadership; answers grounded in official documents, auditable, zero-hallucination.
Act · IA
On the agent side, 12319 requests are handled efficiently — intake, routing and resolution closed loop, fully logged.
Trust & Governance
- Public policy for all; sensitive cases controlled (RBAC · to-G, by data sensitivity)
- Red line: no cross-citizen privacy — citizen A cannot see citizen B's complaint
- Answers grounded in official documents, not made up by AI; sources traceable
- Human oversight retained on sensitive decisions; fully auditable
Decision Provenance · Decision Trail
Every key node is auditable and traceable; critical decisions can optionally be recorded into a Trust Ledger.
Outcomes
Strategic Value
The Kuitun City Brain proves Trusted AI can enter public services — a zero-tolerance, high-privacy setting — where public policy serves all and sensitive cases stay controlled. It uses all four capabilities — Understand, Remember, Govern, Act — and its role-based access control (RBAC · to-G) is the government template of Entrust's Govern capability.
Next Phase
Expand policy and department coverage, deepen the 12319 intake loop, and gradually add incremental case streams and an internal staff knowledge base.