Baidu AI Cloud Financial Services China Translated from Chinese

Agricultural Bank of China

Document Processing

Agricultural Bank of China needed to automate high-volume document and identity extraction across rural onboarding, credit and operations workflows where manual entry was slow, error-prone and expensive.

Batch statement entry improved 10x.
OCR accuracy reached above 98%.
The OCR intelligent-service cloud supports 13 major business scenarios.
The platform serves more than 400 million individual customers and 4 million corporate customers.

Reported impact

Business outcome

ABC improved recognition quality, cut document-entry effort and embedded OCR into a broad set of production banking scenarios, making smart-document handling part of its larger intelligent-bank architecture.

  • Cost Reduction
  • Customer Experience
  • Productivity Improvement

Solution

Agricultural Bank of China extended its Financial Brain program with an OCR intelligent-service cloud that standardizes document extraction across multiple business lines. The deployment uses a distributed, loosely coupled architecture under the bank's Financial Brain so OCR becomes a reusable capability rather than a point solution. Within that framework, Baidu OCR services cover general OCR, identity cards, bank cards, tables, business licenses and custom templates, enabling the bank to process both structured forms and mixed operational documents. The platform sits behind several production scenarios: rural users can register and bind cards faster through ID and face-recognition capture, credit officers can extract enterprise financial statements from large batches of forms, and operations teams can handle account opening, bill extraction and centralized processing with fewer manual steps. By folding OCR into Financial Brain's perception layer and unified access controls, the bank turned document understanding into a core service for smart banking rather than a one-off workflow tool.

Data flow

Identity documents, cards, tables and bank forms are routed into the OCR intelligent-service cloud, where recognition services extract fields and return them to onboarding, credit, mobile-banking and operations workflows for downstream processing.

Solution architecture

6 components · 1 layer
  1. Compute
    • General OCR Extracts text from common banking documents as part of the bank's OCR service cloud.
    • ID Card OCR Captures and parses identity information for onboarding and verification workflows.
    • Bank Card OCR Reads bank-card information for account and payment workflows.
    • Table OCR Extracts structured data from enterprise financial statements and other tables.
    • Business License OCR Processes business-license information for enterprise workflows.
    • Custom Template OCR Handles bank-specific forms and document layouts not covered by standard templates.

Architecture clues

  • OCR is treated as part of Financial Brain's perception engine
  • The OCR service cloud uses a loosely coupled distributed architecture
  • The same OCR stack serves rural onboarding, credit and centralized operations

Evidence from the source

In smart credit, hundreds of tables can be entered in seconds through OCR table recognition, greatly improving efficiency (original: 通过OCR表格识别和输入,几百张表格实现秒级录入,大幅提升信贷经理的工作效率).
The OCR intelligent-service cloud became a key second-phase project in Agricultural Bank of China's Financial Brain (original: 农行金融大脑二期重点项目OCR智能服务云正式投产).
The OCR service cloud includes six capabilities and reaches recognition accuracy above 98% (original: OCR智能服务云,包括六大能力……OCR智能服务云的识别准确率达到98%以上).
The system already supports 13 major scenarios and serves more than 4亿 personal customers and 400万 corporate customers (original: 已为农行的……13大业务场景中提供多样化的智能应用……为农行超过4亿个人客户、400万对公客户提供AI智能服务).
Transparency Evidence & methodology 100/100 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 value100/100
  • 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 impact25/25
    2 metrics describe measurable impact.
  • Deployment scale10/10
    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

    Agricultural Bank of China customer story

    First-party
    https://cloud.baidu.com/customer/case/nongyeyinhang.html
    Source type
    Vendor-published customer story
    Published
    May 14, 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 14, 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.

Related

Similar use cases

See all like this