SOLUTION OVERVIEW
Solution overview
Use reviewed policy and service knowledge for consultation, routing, and guided processing, moving from “can answer” to “can guide” while keeping people accountable.
AI-Enabled Guidance and Public-Service Solution
Use reviewed policy and service knowledge for consultation, routing, and guided processing—moving from answers to actionable guidance.
Delivery scope and practical service journeys
Help visitors understand which service is relevant, what to bring, and who can help next. Build maintainable knowledge from approved policies and service rules, provide guidance and policy matching, and hand questions requiring verification to staff. AI supports preparation and explanation; it does not replace formal processing or approval.
Delivery scope and responsibilities
Service stageResponsible usersSystems and equipmentOperational output
Knowledge preparation | service teams own the sources
Service teams organise policies, service guides, document lists, and common questions, with named reviewers and update owners.
Approved knowledge, service rules, and maintenance tools retain the relevant location, effective period, conditions, and source.
A source-traceable knowledge list identifies missing or conflicting information for review instead of filling a service-rule gap with a generated answer.
Service guidance | turn a question into documents and next steps
Visitors select a task and confirm relevant conditions; reception staff help interpret the request and provide missing information.
Online AI guidance, a digital assistant, or an on-site channel presents service explanations, document requirements, and processing locations for the confirmed conditions.
The visitor receives a task-specific document list and next steps, with human consultation for out-of-scope questions or uncertain answers.
Policy matching | explain why a policy may apply
Residents complete required authentication and information entry with staff help; service staff verify their circumstances and policy applicability.
A policy database, matching rules, and report output use authorised information relevant to the task, such as household, health, and employment circumstances.
The report lists relevant policies, conditions, documents, service points, and reasons for a match or non-match for visitors to understand and staff to review.
Video or counter assistance | move from guidance to action
Specialists take over complex questions, review documents, and explain the business-system steps; visitors follow the guidance for their application.
Video consultation, screen sharing, service guides, and counter service allow document review and demonstrations of available systems, such as Shanghai’s local service portal.
The visitor knows which authority processes the task, what documents to add, and how to continue; consultation advice is not presented as an approval decision.
Ongoing maintenance | make errors correctable
Knowledge owners handle policy changes, service staff review representative questions, and operators track unresolved requests and human handovers.
Knowledge versions, issue feedback, and answer sampling support a maintenance process for outdated content, staff confirmation, and republication.
Update ownership, review records, and resolutions are documented, and visitor-facing content corresponds to the reviewed version.
Policy matching: from household circumstances to a checkable service report
A resident asks which policies may be available; staff explain the purpose, information scope, and human-review stage.
With staff assistance, the resident completes identity authentication and enters the household, health, and employment information relevant to the matching task.
The system compares the information with the prepared policy conditions, distinguishing potentially relevant policies from conditions not met by the available information.
A report presents policy names, application conditions, required documents, service points, and reasons for a match or non-match.
Staff explain the report item by item and verify the resident’s circumstances, identifying supporting documents or further checks where required.
The resident and staff prepare the application documents and proceed to formal processing; the responsible authority applies its own eligibility and approval rules.
What the service producesThe output is understandable, checkable guidance that supports a real application—not an automatic award of benefits. Online guidance follows the same principle: for example, social-security card guidance first checks household-registration and insurance details before presenting documents and a service route.
How this works in real projects
PROJECT IN PRACTICE
Toyou AI smart policy medical examination system is newly launched in Tianlin Street, Xuhui District
A public report from 2026 describes Tianlin’s policy database and staff-assisted reports that help residents understand relevant public-support policies.
- After authentication, residents provide household, health, and employment information; the report shows policies, conditions, documents, and processing locations.
- The featured resident was guided towards hospitalisation mutual assistance for retirees, healthcare-cost relief, expenditure-related household assistance, and subdistrict holiday support.
- Staff explained the report and helped prepare application documents; the report could also explain why a policy matched or did not match.
The image shows the service centre, not a product interface. The named policies belong to this local case; other locations require approved local policies, and responsible authorities make formal determinations.
Read the full project →
PROJECT IN PRACTICE
Toyou’s new AI guidance + video assistant unlocking service experience is launched in Jing’an, Shanghai
The upgraded Fumin service mini-program connects service conditions and document guidance with video consultation, showing a route from initial explanation to staff assistance.
- In the social-security card example, the user confirmed household-registration and insurance information, received a document list, and followed the application guide.
- Video consultation allows visitors to show documents while staff share screens, demonstrate the local service portal, or display service-guide slides.
- The featured user received remote guidance on card activation and out-of-area medical registration while travelling, connecting consultation to practical action.
The image is a public project announcement, not a screenshot of chat, reporting, or the specialist console. The reported outcome is not a promise for other services, locations, or personal circumstances.
Read the full project →How to verify the delivered workflow
CheckWhat to testExpected evidence
Answers have a service-rule basis
Test approved common and boundary questions against source policies, location, effective period, and document lists.
A knowledge-source list, question test set, service-team review results, and handover records for uncertain questions.
Reports explain conditions without replacing approval
Use service-team-approved examples to check matches, unmet conditions, documents, and service points, including human review and formal-processing notices.
A policy-condition comparison, sample reports, correction records for incorrect matches, and a statement of service boundaries.
Human handover leads to a usable next step
Test complex questions, missing documents, and out-of-scope requests; confirm usable video or counter routes and that receiving staff understand the issue requiring review.
Human-takeover routes, service departments and hours, end-to-end walkthrough records, and offline alternatives.
Policy changes reach the service channels
Update a test knowledge item through the agreed process and check review, version, visitor-facing content, and old-content handling, together with personal-data fields and permissions.
A knowledge-maintenance responsibility matrix, update and review records, version list, and permissions for necessary information.
Full construction brief
The retained construction material describes product options and reference projects. Historical figures and delivery periods apply to their original project scope, not a universal performance or delivery promise.
1. Human-AI collaboration
Counter captioning, digital assistants, and intelligent customer service support understanding and information organisation, with staff retaining formal interpretation and fallback responsibility.
2. Intelligent forms
OCR captures document data and checks forms against authorised service rules, while voice-assisted forms improve accessibility.
3. Policy knowledge services
AI Policy Matching identifies relevant elderly-care, healthcare, and social-assistance support.
4. Operations and workforce development
24/7 AI telephone service, personalised training, assessment, and hall patrol support staff capability and operations.
5. Application boundaries
Define approved knowledge, authorised data, verifiable outcomes, update ownership, and human-takeover conditions so generative capability is not confused with formal business rules.
Measured support
- DeepSeek and other large models as the core engine
- 40% business growth in the first half of 2025
- Telephone, training, and assessment products deployed in Shanghai and expanding to other cities
Sensitive data remains encrypted inside the government environment, with encrypted streams and project-defined masking.
Delivery priorities
Define the workflow, staff responsibilities and existing systems, then select the products for each service stage.
01
Service guidance and document preparation
Use reviewed eligibility rules, document lists, and service routes for questions and next-step guidance, handing uncertain requests to staff.
02
Digital assistants and on-site guidance
Deliver questions and guidance through digital assistants or robots alongside hall displays, considering noise, visitor needs, and staff assistance.
03
Policy matching and reports
With staff assistance and authorised information capture, match elderly-care, healthcare, and assistance policies and show conditions, documents, and service points for review.
04
From guidance to human assistance
Connect questions and document preparation to video or counter services, where staff verify circumstances, explain rules, and guide formal applications.
05
Knowledge maintenance and application scope
Define source review, updates, answer sampling, and human takeover. Policy matching and generated answers support service work; responsible authorities determine formal eligibility.
How the service moves from start to finish
The steps connect visitors, service teams, and configured systems.
- 01
Review the knowledge base
Service teams review policies, items, documents, and common questions, with defined sources, locations, and update owners.
- 02
Understand the visitor request
Confirm the conditions relevant to the request, explain and authorise necessary personal-data use, and allow staff-assisted input.
- 03
Provide guidance that can be checked
Present document lists, service points, and next steps. Policy-matching reports assist review, while staff confirm applicability.
- 04
Hand over and maintain the service
Transfer complex requests to video assistance or counter staff and use feedback to correct knowledge, update expired content, and improve service.
Plan around your setting and existing foundation
Suitable service needs
Halls and online channels with frequent policy questions, complex document preparation, and difficult service selection, particularly where human assistance is required.
Knowledge required before planning
Approved policies, service guides, document examples, and questions, with update dates, applicability, and responsible reviewers.
How to configure the channels
Combine online guidance, digital assistants, on-site robots, and policy-matching entry points as needed, connecting them to video and counter services.
How to verify the service
Test common questions, missing documents, and out-of-scope requests for answer support, human handover, updates, and staff ability to track and correct issues.
Policy matching and intelligent guidance support consultation and document preparation. Responsible authorities carry out formal processing, eligibility review, and approval. Site photos and project announcements are not software-interface screenshots.
Related products
Products are drawn from the original public website archive. The final combination depends on site conditions, existing systems, and project scope.
Project evidence
Archived images from publicly presented projects show how the solution fits real service environments.
Related project cases
Explore the project context, public delivery information, and archived site material.
PROJECT DISCUSSION
Discuss this solutionDefine the solution around your site and existing systems
Share the service context, existing foundation, and target outcome with our team.