Baidu AI Cloud Public Sector China Translated from Chinese

Guangdong Provincial Government Services and Data Administration (广东省政务服务和数据管理局)

Enterprise Knowledge Search

Guangdong needed to replace slow, manual, and fragmented investment-attraction processes with a digital service that could match enterprises to resources and explain frequently changing policies more accurately and at lower operating cost.

Current consultation accuracy exceeds 90%.
The Heyuan pilot was trained and launched within 2 months.

Reported impact

Business outcome

The province gained a scalable digital investment-attraction assistant that speeds consultation, improves policy interpretation, lowers service cost, and gives Guangdong a reusable public-sector large-model pattern for future digital-government services.

  • Cost Reduction
  • Customer Experience
  • Productivity Improvement

Solution

Guangdong's government services and data administration department worked with Baidu AI Cloud to build the 'Investment Guangdong' AI assistant as a province-level investment-attraction service. The solution refactored Baidu's UNIT dialogue platform with Wenxin large-model capabilities, then integrated large volumes of structured resource data covering parks, factories, office buildings, projects, land parcels, and events alongside unstructured policy documents. Qianfan-based summarization and generation improved the system's understanding of colloquial business questions, while a shared knowledge layer let the assistant perform precise matching against招商 resources and natural-language Q&A over policy text. In the Heyuan pilot, the team fine-tuned the model on local full-volume resource and policy data, enabling a single assistant to answer investor questions, summarize policy content, and return targeted resource recommendations without forcing citizens or enterprises to search across many disconnected systems.

Data flow

Structured data about parks, factories, buildings, land parcels, projects, and events is combined with policy documents inside the shared knowledge base. Investor questions enter through the dialogue platform, Qianfan interprets colloquial intent and policy text, and the assistant returns either matched investment resources or generated policy explanations for Guangdong and pilot city scenarios.

Solution architecture

4 components · 3 layers
  1. Compute
    • Qianfan Large Model Platform Provides the summarization and generation capabilities that improve policy understanding and investor question answering.
  2. Storage
    • Resource and Policy Knowledge Base Stores structured investment resources and unstructured policy content so the assistant can return grounded matches and answers.
  3. Serving
    • UNIT Intelligent Dialogue Platform Serves as the conversational foundation that was rebuilt and upgraded for the province-level investment assistant.
    • Investment Guangdong AI Assistant Delivers investor-facing consultation, resource matching, and policy interpretation services.

Architecture clues

  • The assistant was built as a province-level service and then fine-tuned on full-volume local data for a pilot city.
  • The implementation combines structured investment-resource matching with unstructured policy-document Q&A and summarization.
  • The system is intended to support citizens and enterprises asking policy and resource questions in natural language.

Evidence from the source

It can answer precise questions over structured investment resources and perform natural-language Q&A and generation over unstructured policy documents (original: 具备招商资源结构化数据的精准匹配问答;具备招商政策文本文档非结构化数据的自然语言问答和生成).
The Heyuan pilot completed model training and launch within two months and currently achieves more than 90% inquiry accuracy (original: 2个月内完成了模型训练和上线…现阶段问询准确率达90%以上).
The assistant rebuilt Baidu UNIT with Wenxin large-model capabilities and integrated massive resource and policy data (original: 依托百度文心大模型,对百度UNIT智能对话平台进行重构升级,集成广东省海量资源与政策数据).
The department and Baidu built the country's first province-level investment-attraction AI assistant (original: 广东省政数局联合百度智能云打造全国首个省级招商引资AI助手).
Transparency Evidence & methodology 85/100 reference value · 89% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

01

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value85/100
  • Problem clarity20/20
    Measures whether the customer's operational problem is specific enough to compare.
  • Implementation depth25/25
    Rewards named components, data flow, integrations, and architecture detail.
  • Outcome specificity20/20
    Measures whether the reported benefit is concrete rather than a generic claim.
  • Quantified impact15/25
    1 metric describe measurable impact.
  • Deployment scale5/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence89%
  • Source authority20/25
    The main evidence is an official vendor-published customer story.
  • Claim traceability21/25
    10 of 12 key fields link to a named source.
  • Direct evidence20/20
    4 direct source quotes retained.
  • Date provenance10/10
    The publication date comes from the source page or publisher.
  • Source transparency8/10
    1 source reference document publisher, URL, authority, retrieval, and supported fields.
  • Implementation specificity10/10
    Rewards source-backed components, data flow, quantified impact, and a specific outcome.
02

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

  1. Baidu AI Cloud

    Guangdong Provincial Government Services and Data Administration (广东省政务服务和数据管理局) customer story

    First-party
    https://cloud.baidu.com/customer/case/zhengshu.html
    Source type
    Vendor-published customer story
    Published
    May 15, 2025
    Supports
    • Customer
    • Business problem
    • Solution
    • Products
    • Solution components
    • Data flow
    • Integrations
    • technical area
    • use case category
    • Business outcome
    • outcome category
    • Metrics
    • Architecture
    • evidence quotes
    • Publication date

Date evidence

  • Published on source page May 15, 2025 · Baidu AI Cloud · First-party
    Open date source
Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.