Anonymized Urban Planning Agency
Urban planning research
Datawhale
Government/Public Service
Explicit FDE engagement
Overview case | Use as a research lead
Evidence level A
In validation
A
Evidence level
Datawhale FDE Case 100 | No. 4: From manual analysis to intelligent urban-planning coordination
Evidence level measures whether a source can be located and reviewed; it does not mean vendor-reported claims were independently audited.
Overview case | Use as a research lead
5 / 12
Business context / 2
Transformation workflow / 2
Technical workflow2 / 2
Human roles & governance1 / 2
Measured outcomes / 2
Source traceability2 / 2
Remaining gaps: Business context, Transformation workflow, Human roles & governance, Measured outcomes
Business problem
Planning data, policies, reporting are fragmented and manual aggregation and research cycles are long.
Solution
Collect multi-source data and documents and generate them using AI-assisted retrieval, synthesizing, researching and reporting.
Technical architecture & production workflow
Step 01
SOP / History Case / Expert Interview
→
Step 02
Document Parsing, Cutting and Metadata Tags
→
Step 03
Embedding / semantic index
→
Step 04
Retrieving similar cases and rules based on current questions
→
Step 05
LLM is evidence-based advice and attachment
→
Step 06
• Rewrite the knowledge base of the new experience
Key technology & infrastructure components
Knowledge base/RAGEmbedding/vector searchPermissions and metadataLLMExpert review
Human roles & accountability
Planning specialists are responsible for judgement, policy interpretation and formal conclusions.
FDE delivery actions
- "When will the staff come to see the teacher?"
- Interviews with experts to convert tacit judgment into searchable cases and rules
- Design knowledge particles, labels, versions and privileges instead of simply uploading documents
- Recall rate/Application of answers using real questions
- Keep feeding back feedback, wrong answers and new cases.
Reusable delivery patterns
- HF, high-cost, verifiable narrow-flow selection
- Steps that can be validated with a certainty tool to prevent model self-assessment
- Retain manual responsibility for high-risk actions
- For each manual amendment to follow-up searchable context/rules
Business outcomes & delivery results
Reduction of desk analysis cycle; specific indicators not made public.
Evidence boundaries & verification notes
- Public information confirming business processes; technical components are abstract architecture based on public description
- The source can be directly located in the case; the value remains disclosed by the source and does not represent an independent audit.