【Technical White Paper】Chapter 4: Information Extraction Technology Industry Application Status and Cases (Part 1)

4. Status quo of industrial application

4.1 Industrial application of information extraction technology

Information extraction technology has been developed for many years, and related industries are becoming more and more mature. The following are several major industrial applications of information extraction:

• Business information extraction: through the development of a special information extraction system, analyze the big data of business information collected from various channels, extract information such as relevant company business information, public opinion status, risk monitoring, etc., and provide decision support information.
• Medical information extraction: Healthcare institutions and health insurance departments can use the information extraction system to obtain patients' symptoms, diagnosis, test results, and treatment in order to better provide medical services and insurance services.
• Government information extraction: Government departments use the information extraction system to obtain effective information from numerous government documents, making government services more accurate and efficient.

---

4.2 Industrial Application Cases of Information Extraction Technology

4.2.1 TextIn contract robot based on deep learning information extraction technology

Contract comparison is an essential link in the business process for enterprises to sign contracts and establish cooperation. Before the text is finalized, the contract often undergoes repeated revisions and version iterations, or there are differences in additions and deletions between the electronic version and the paper version. When signing a contract, it is necessary to ensure that the printed contract is consistent with the approved contract text, and the key information is complete, so as to avoid risks such as the use of templates, text modification, yin and yang contracts, and falsification of letters and certificates.

Hehe Information launched the TextIn contract robot . Based on STR identification and NLP algorithm, it has developed two core capabilities of contract key information identification and extraction and contract comparison. Help enterprises reduce contract risks and avoid major losses, while improving work efficiency, reducing duplication of labor, and reducing employment costs. It helps enterprises realize the full life cycle management of contracts, and has prominent applications in banking, insurance, securities, asset management, financial leasing, supply chain finance and other pan-financial businesses as well as corporate legal contract review scenarios.

In the process of contract review and management, different contracts often need to extract different key information. For example, purchase and sale contracts need to extract information such as Party A and Party B, contract amount, etc., lease contracts need to extract information such as the lessee, lessor, lease fee, and deposit. The contract needs to extract information such as the project name, project location, and start date, and the fund contract needs to extract information such as the fund client, manager, and fundraising period.

A unified and standardized extraction field cannot meet the extraction needs of all types of contracts. Therefore, TextIn Contract Robot has launched a new function of "extracting information by contract classification". Users can classify contract types in the "key information configuration", and customize and create the required extraction fields for each contract type.

After the contract is uploaded according to category, the contract robot can automatically extract the configured key information and seal information.

Core functions:

• Intelligent extraction of key information and seal information
  Supports identification of pictures (png, jpg, jpeg, tif, tiff), Word, PDF, Excel format contracts, intelligent extraction of key information and seal information. Supports custom setting of key information fields by contract classification. Seal information covers seal type and subject information. Seal types include: official seal, special seal for contracts, legal representative seal, special financial seal, special seal for invoices, and special business seal.
  

- Key information comparison

Support intelligent comparison of standard information extracted from the contract and self-configured information, quickly locate key information differences, and export Excel, Word, and PDF difference reports with one click.
Contracts often contain a large amount of table information, and table identification and comparison is also a difficult point in contract comparison.
Based on complex information extraction and text detection and recognition technologies, the contract robot can not only identify table information in documents of different formats with high precision, display the comparison differences of all fields, but also optimize the table comparison style, and compare content differences according to cells Alignment and list display are performed to facilitate clearer and more intuitive locking of contract differences.

Application Scenario

• Insurance contract management
  helps insurance companies compare the approved contract in the OA system with the printed contract uploaded in the risk system. If there is any discrepancy, an early warning is issued to avoid the risk of contract forgery; the contract submitted by the undertaker is compared with the contract template of the same type Compare the terms and remind the differences and missing situations. If the contract template is provided by the other party, extract key information and match it with the background rules.

• Bank credit approval
  helps banks to judge the existence of seals and handwritten signatures on leasing contracts, operating sites, rural property contract certificates and other materials, compare the contracts to be reviewed with the contract templates, and judge whether there are additions, deletions, or changes in clauses .

• Corporate Legal Review
  Helps corporate legal departments to control the iteration of each document version, which facilitates improving efficiency and reducing human errors in the process of repeatedly revising contracts between both parties.

• Procurement risk prevention
  Helps the procurement department of the enterprise to prevent the risk of "yin and yang contracts" and falsification of letter certificates in the early stage of contract negotiation, and identify and display the differences in paper contracts.

---

4.2.2 Tianyuan big data platform integrating knowledge graph based on deep learning information extraction

Financial risk control is facing the dual dilemma of information overload and information scarcity. On the one hand, due to the isolation of systems and data, isolated islands of information generally exist, making it difficult to obtain high-quality information or prone to information inconsistency, incompleteness, and one-sidedness, while the complexity of risk control requires a systematic approach to multi-dimensional information. collection and overall analysis; on the other hand, the information from various external channels and business lines is complex, it is difficult to judge the authenticity and validity of the data, and it is even more difficult to process a large amount of data on a large scale and mine the inner meaning of the data. connect.

Data in the financial field is a typical big data with "4V" characteristics (massive Volume, multi-structure and multi-dimensional Variety, huge Value, and timeliness Velocity).

The knowledge map can mine, analyze, construct, draw and display the interconnection between big data through visualization technology, which can well solve the difficulties of financial data application. Hehe Information has been deeply involved in the field of knowledge graphs for more than 10 years. Through complex information extraction technology based on deep learning, it has launched a variety of relationship graphs for the risk control needs of banks, securities, insurance, financial leasing, supply chain financial platforms and other institutions. These relationship graphs show the investment, employment, guarantee, litigation, upstream and downstream, suspected and other relationships between enterprises and natural persons in a multi-faceted manner, and risk control in pan-financial enterprises such as credit granting, investment, anti-money laundering, anti-fraud, due diligence, monitoring and early warning There are important applications in the scene.

• (1) Group relationship extraction
  Use the full amount of industrial and commercial equity data, combined with the associated customer identification method in the "Large Risk Exposure Management Measures" to model, extract equity investment relationships between enterprises, and analyze enterprise group factions.

• (2) Transaction relationship extraction
  Combined with the customer's internal transaction data, through information extraction and mining analysis of all enterprise transaction data information, and finally obtain the intra-industry enterprise transaction relationship.

• (3) Equity penetration extraction
  Based on big data calculation and information extraction, it accurately analyzes the equity structure behind the enterprise, and clearly displays the information of all partners of the enterprise disclosed to the natural person and legal person levels.

• (4) Litigation relationship extraction
  Utilizes the full amount of judicial litigation data, including case filing information, court announcements, court announcements, and judgment documents, and uses information extraction technology to mine corporate judicial litigation relationships

• (5) Event relationship extraction
  Using information extraction technology, according to the company's news trends and public opinion data, the event relationship between enterprises is calculated, and the negative index is high, and 14 types of public opinion relations in the past 6 months are output.
  

• (6) Extraction of upstream and downstream relations and competitive relations
  Use information extraction technology to display the relationship between the enterprise and upstream suppliers, downstream customers, and competitors through data such as supplier announcements, bidding, bidding, and bid winning information.

Relying on more than 100 billion pieces of real-time dynamic commercial big data from 230 million domestic enterprises and other organizations gathered by Qixinbao under its subsidiary Qixinbao, Hehe Information combines the business application solutions and technologies accumulated from the experience of big data platform construction projects of many leading enterprises Based on the self-developed knowledge graph technology that integrates information extraction based on deep learning, Tianyuan launched the Tianyuan big data application platform.

Data Quality Assurance Mechanism

Through a variety of data collection, cleaning and verification methods, relying on strict process control, the comprehensiveness and accuracy of data are improved.

Data Kanban

Dynamically present the layout and development of enterprises in the region, and provide financial enterprises with corporate marketing channels and marketing decision-making data support.

Business Home

Aggregated application portals, one-click search for enterprise and key personnel information.

core application

---

customer case

• A joint-stock bank
  A joint-stock bank joined hands with Hehe Information to build a big data risk portal, combining data such as credit and guarantees in the bank with Qixinbao big data under Hehe Information, realizing multi-dimensional search, intelligent risk association, and public opinion information analysis , map association analysis, information intelligence view, internal and external risk assessment and other applications.
  After the launch, nearly 10,000 employees in the bank have used the risk portal, including 4,500 account managers and 2,200 risk managers. A total of 55 branches and subsidiaries have used the risk portal to screen corporate risks. Employee feedback, the risk portal makes business personnel no longer need to query enterprise-related information through external third-party platforms, with higher efficiency and richer data dimensions, and through the integration of big data in the industry, the hidden guarantee circle chain and inter-enterprise transaction relations can be realized. Panoramic presentation across branches breaks down information islands.

  

• A leading agricultural commercial bank in a strong economic city
  A leading agricultural commercial bank in a strong economic city cooperates with Hehe Information to build a big data application portal to meet the unified data acquisition platform for business departments, operating departments, management departments, and risk control departments. , channels, and standards to empower the core strategy of the bank's digital transformation.
  The big data application portal is equipped with knowledge graph applications such as corporate portraits, corporate graphs, risk monitoring, and data visualization cockpit, as well as intelligent text recognition applications such as TextIn financial reporting robots, to meet the needs of multiple departments for risk control, anti-money laundering compliance, and marketing. For example, the in-depth perspective of the background of corporate customers, anti-money laundering compliance review, compliance supervision of branches and sub-branches by the Compliance and Legal Affairs Department, regional corporate customer marketing, pre-loan comprehensive review, post-loan comprehensive management, In business scenarios such as policy interpretation and corresponding product planning, identification of related parties and unified credit extension by the credit approval department, multi-dimensional credit verification, interpretation of regulatory laws and regulations, and insight into major risk issues, etc., through this big data application portal, multi-dimensional internal and external information can be obtained. data support.

---

4.2.3 Supply Chain Finance Solutions

Supply chain finance refers to starting from the overall supply chain industry chain, using scientific and technological means to integrate information such as logistics, capital flow, and information flow, and building the integration of core enterprises that dominate the supply chain and upstream and downstream enterprises in the context of real transactions. The advanced financial supply system and risk assessment system provide systematic financial solutions to quickly respond to the comprehensive needs of enterprises in the industrial chain for settlement, financing, and financial management, reduce enterprise costs, and enhance the value of all parties in the industrial chain.

Based on the capabilities of intelligent text recognition technology, Qixinbao big data mining and knowledge graph, Hehe Information has helped many supply chain financial companies to realize the recognition and entry of certificates and documents, credit risk control of financing enterprises, and analysis of industrial chain supply chain. The business process is automated and intelligent, and effectively improves the credit risk control ability of upstream and downstream financing enterprises.

Intelligent identification and entry of certificates and bills

When upstream and downstream enterprises with financing needs independently submit financing applications through the supply chain financial platform, they need to submit enterprise certificates, acceptance bills and other materials to facilitate the platform's review.

Hehe Information provides 100+ types of card identification modules, covering common types of licenses in financial scenarios, such as: ID cards, bank cards, real estate certificates, business licenses, etc. Among them, ID card recognition supports automatic judgment of front and back, automatic judgment of original, color copy, and copy; bank card recognition supports debit cards and credit cards from different countries and industries in the world, VISA, JCB, UnionPay cards, Master cards, American Express Cards and other domestic and foreign bank cards can be accurately identified; business license identification supports old and new formats, and automatically distinguishes whether it is a copy or whether it is an electronic business license. It can also call Qixinbao big data through the API interface to automatically verify the authenticity of the business license , To prevent the risk of certificate fraud and invalidation.

In addition, Hehe Information also supports the intelligent classification of various types of bills such as electronic bank acceptance bills, commercial acceptance bills, and value-added tax invoices in different formats, and automatically enters them into the system after high-precision identification. The value-added tax invoices support automatic verification, which greatly improves the Improve business efficiency and release manpower.

Platform data combined with external data for flexible credit granting

The amount of financing application submitted by different enterprises is different, and the supply chain financial platform needs to evaluate the financing amount according to the credit of the enterprise. The evaluation of enterprise credit in supply chain finance relies on the transaction data between upstream and downstream financing enterprises and core enterprises on the one hand, thus solving the dilemma of small and medium-sized enterprises upstream and downstream with less credit information and relatively opaque information. These transaction data are provided by the supply chain The collection and standardization of financial platforms is the core credit data. On the other hand, it is also necessary to combine external big data to conduct due diligence on the industrial and commercial, judicial litigation, and business risks of financing companies.

Due diligence on enterprise risk panorama

Qixinbao, a subsidiary of Hehe Information, gathers more than 100 billion pieces of real-time dynamic commercial big data from 230 million domestic enterprises and other organizations, helping the supply chain financial platform to see the panoramic information of corporate customers, multi-dimensional risk scanning, covering corporate industry and commerce, and judicial litigation , business, taxation, qualifications, intellectual property rights, corporate relations, negative public opinion and other information.

The supply chain financial platform needs to pay special attention to the tax risks, judicial litigation, administrative penalties and other information of upstream and downstream small and medium-sized enterprises. will have a significant impact on business operations.

Qixinbao, a subsidiary of Hehe Information, provides multi-dimensional enterprise dynamic risk information inquiries such as administrative penalties, environmental protection penalties, serious violations and dishonesty, case filing information, court hearing announcements, persons subject to enforcement, and tax arrears information, and can conduct 7×24 multi-dimensional risk analysis Hourly real-time monitoring, four-level risk early warning mechanism, custom setting of risk level division, if a risk signal occurs, timely push early warning through SMS, email, WeChat, API interface, etc.

Contract key information extraction, intelligent decision engine

Supply chain finance involves the entry and review of multi-type contract information between multiple parties, such as purchase and sale contracts between core enterprises and upstream and downstream enterprises, pledge contracts between financing enterprises and supply chain financial platforms, core enterprises and supply chain finance Guarantee contracts between platforms. These contract information need to be checked for consistency with each other, and at the same time, they must be compared with transaction data such as order amounts and accounts receivable amounts.

Contract Key Information Extraction

Hehe Information launched the TextIn contract robot, which can use deep learning-based text detection and recognition and information extraction algorithms to intelligently extract key contract information. TextIn contract robot has built-in preset extraction of 42 standard key information fields, and supports custom configuration extraction fields.

The supply chain financial platform can efficiently extract the contract information required by the business based on the contract robot, such as the name of the buyer, the name of the seller, the validity period of the contract, and the contract amount in the purchase and sale contract. The supply chain financial platform can carry out intelligent decision-making and approval on the extracted contract information according to the audit items, such as "whether the buyer is a core enterprise, whether the seller is a financing enterprise, and whether the contract amount is greater than or equal to the amount of accounts receivable", so that the transaction background The veracity of the audit judgment.

In addition to reviewing the contract text information, the supply chain financial platform also reviews the contract seal. First of all, it is necessary to ensure that all the contract signing parties have stamped the seal, and the seal is clearly identifiable. In addition, it is necessary to check whether the name of the seal is consistent with the name of the contract signing party.

---

4.2.4 Full-process solutions for bank credit business

At present, the bank credit business presents the characteristics of diversified customer entities, group customers spanning regions, and business scope spanning industries. Insufficient identification of corporate information, uneven levels of customer credit status, increasingly complex corporate relationships, and time lag in post-loan risk management pose higher challenges to banks' big data application capabilities.

Due to the use of multiple independent systems for multiple business lines in the bank, there are isolated islands of data in the bank, and there is a lack of introduction of external data, so the data's ability to support the business is weak. Most of the data collection methods rely on manual sorting and entry, and there is a time lag in data update, and the efficiency is low and the labor cost is high.

Based on the capabilities of intelligent information extraction technology, Qixinbao big data mining and knowledge graph, Hehe Information Bank's full-process credit solution covers four major links: credit expansion, pre-loan access investigation, credit approval during loan, and post-loan early warning and monitoring , The whole process assists the bank credit business.

(1). Credit expansion

• 
  In the context of vigorously developing green finance, inclusive finance, science and technology innovation finance, and supply chain finance, regulators have requirements for assessment indicators for special credit such as green credit and small and micro enterprise credit . Based on the big data and industrial research of 230 million enterprises, Qixinbao, a subsidiary of Hehe Information, has launched a multi-type high-quality enterprise database through accurate information extraction technology to help banks accurately tap potential customers for special credit.
  Emerging Industry Database: Covering 14 categories of new agriculture, new energy, new generation information technology, energy conservation and environmental protection, and 1000+ sub-sectors of emerging industry companies and their geographical distribution.
  
  Green Industry Database: Covering 6 major categories of infrastructure green upgrades, green services, clean energy industries, energy conservation and environmental protection industries, ecological environment industries, and clean production industries, and 30+ sub-categories of green industry companies and their geographical distribution.
  
  Science and Technology Certification Enterprise Database: Covers the list of science and technology certification enterprises released by the Ministry of Industry and Information Technology, the Ministry of Science and Technology, and the National Development and Reform Commission, including high-tech enterprises, specialized and special innovations, technology giants, small and micro enterprises, technology-based small and medium-sized enterprises, Torch Program projects, and technological innovation demonstrations 9 categories of enterprise databases for enterprises, enterprise technology centers, and technology business incubators.

  Industrial chain: Provide 200+ industrial chain topics, covering more than 70,000 standard products, 8 million companies, and 100 million various industrial chains such as product-product upstream and downstream relationships, enterprise-product relationships, and enterprise-enterprise upstream and downstream relationships connection relation.

• Real-time capture of credit demand signals
  Favorable corporate events such as bidding, proposed projects, investment and financing, etc. may become high-quality credit demand signals. In traditional corporate marketing, the acquisition of information on these events relies on high-level interviews, on-site inspections, and inquiries from multiple parties, and the efficiency of clue acquisition is low.

Qixinbao "Potential Customer Subscription" under Hehe Information provides subscription services such as bidding, proposed projects, investment and financing information covered by key words such as industry and business, and supports subscription services based on region, registered capital, Qixin points, air space, etc. Screening is carried out on 10 dimensions including shell risk, default risk, subject enterprise risk, associated personnel risk, and associated enterprise risk. While obtaining credit leads, complete the preliminary screening of risk control for potential target companies.

Existing customers will continue to generate new credit demands. However, since account managers manage and maintain a large number of customers, it is difficult to pay attention to the potential demand changes of each customer, especially long-tail customers.
Qixinbao's "Century Extension" helps banks discover and follow up existing customers in a timely manner by pushing bank customers' registered capital changes, new branches, new financing information, real estate land acquisition and qualification certificates and other clue information new credit needs.

(2) Pre-loan access investigation

• KYC query
  KYC means knowing your customer, which is not only an anti-money laundering compliance requirement, but also the purpose of pre-loan due diligence. Qixinbao, a subsidiary of Hehe Information, launched the "KYC" function, providing a scene-based view of customer information, with key information at a glance, helping banks quickly understand the information of the main company, its affiliates, and affiliated companies, covering business, litigation, operation, and qualifications of the company , Multi-dimensional information such as personnel's foreign investment, employment, litigation, equity pledge, etc.

• Due diligence on related relationships
  In the pre-loan due diligence stage, banks need to identify and identify the company's equity structure, shareholder background, actual controllers, ultimate beneficiaries, related parties, directors, supervisors, and senior positions to check related risks.
  Based on big data analysis and mining and knowledge map capabilities, Hehe Information has launched knowledge maps such as equity penetration, controller relationship, beneficial owner identification, related party identification, personnel map, etc., and can combine with big data within the bank to unify business information across branches Analysis, and further improve the map of guarantee relationship, transaction relationship, supply chain upstream and downstream relationships, etc.

Based on the multi-dimensional relationship graph, the bank can penetrate the complex company equity structure layer by layer, accurately identify the actual controller, beneficial owner, shareholder, and related parties, identify the identity of the customer, and meet regulatory compliance requirements; deeply grasp the basic situation of the company, business operation The credit status, credit status, guarantee status, etc. provide a more comprehensive information dimension for credit approval.

Unified credit extension for group customers
Group companies need to carry out unified credit extension, and all companies in the group family tree share the credit line to prevent multiple credit extensions, excessive credit extension, and increased risk exposure.
Banks often have two types of risks in the practice of group customer identification: 1. Due to the difficulty of determining the group relationship, member companies are separated from the group's credit. 2. A large amount of equity transfer occurred in a member company, which led to a change in the group to which it belonged, and the bank lagged behind in obtaining the change information, resulting in the division of the company into the original group.
Hehe Information can build a group relationship graph based on equity investment, directorship, and bank internal customer data. When equity and other data change, the system automatically adjusts the group relationship, avoids the risk of time lag in information acquisition, and promotes bank group genealogy governance. Unified credit extension for group customers.

Pre-loan review report
After the account manager completes the pre-loan due diligence, he needs to summarize the information obtained from online and offline investigations into a report. The channel sources, credibility, and presentation styles of these information may be inconsistent, and they need to be sorted out uniformly. Account managers often spend a lot of time working on cases.

Qixinbao, a subsidiary of Hehe Information, can automatically generate a variety of due diligence reports, including: enterprise value-added credit report, enterprise basic credit report, KYC report, directors, supervisors and senior management investment and employment reports, enterprise equity structure report, equity penetration report, enterprise Business data reports, corporate chattel mortgage reports, etc., systematically sort out complex information, and comprehensively improve the efficiency of investigations.

(3) Loan credit approval

• Financial report identification
  Credit approval requires attention to corporate financial information, and financial reports are the primary source of financial information. Bank credit requires companies to provide financial reports for the last three years, and through analysis of financial reports to reveal the company's financial status, operating results and cash flow for a certain period of time. Based on self-developed intelligent text recognition, form recognition, and natural language processing technologies, Hehe Information launched the Textin financial report robot, which intelligently collects, verifies, and outputs financial report data in a structured manner. It only takes 2-3 seconds to identify a page of financial report information on average. The Textin financial report robot is easy to operate, has a high recognition accuracy rate, supports disconnected tables, and can accurately identify financial statements with few or no table lines. Through intelligent identification and matching of built-in financial standards, subject matching, and trial balance verification, financial data in different file formats and report formats can be efficiently output in a unified standard format, greatly shortening the financial report entry time and standardizing the data output format.

• Verification and exploration of enterprise relationship To approve
  credit in loans, it is necessary to review the customer information and investigation report submitted by the account manager, so as to avoid the operational risk and moral hazard caused by the lack of due diligence in the pre-loan investigation. Qixinbao, a subsidiary of Hehe Information, provides a corporate panorama relationship map - corporate chain map, a 360-degree panoramic view of corporate relationships, identifying and presenting shareholders, executives, foreign investment, judgment documents, court announcements, historical shareholders, suspected relationships, etc. Dimensions to facilitate the verification of the business relationship stated in the investigation report.

Many companies may have implicit associations. Implicit relationship is a kind of relationship between enterprises that does not reveal the relationship on the surface but actually implies an investment relationship or a control or influence relationship in business decision-making, capital allocation or production activities. Hidden related risks are often difficult to identify due to hidden and deliberately concealed relationships, and involve one-to-many and many-to-many investigations, and the investigation efficiency is very low.

Qixinbao, a subsidiary of Hehe Information, provides the "relationship search" function, which can add the names of companies or people in batches, and quickly obtain the relationship graph between target companies and natural persons based on employment, investment, suspected, upstream and downstream relationships, presenting a variety of connection paths , to deeply mine the implicit association relationship.

• Enterprise Credit Rating
  After comprehensive due diligence and credit review of the enterprise's basic information, credit information, operating information, financial information and other information, the bank can rate the customer and assess the credit line.
  Qixinbao, a subsidiary of Hehe Information, launched Qixinbao by cleaning and mining the big data of the whole enterprise, establishing an enterprise scoring model, and calculating Qixin based on data such as comprehensive development potential, intellectual property rights, enterprise scale, risk status, operating quality, and capital background. Score the comprehensive ranking in the industry, quantify the enterprise risk level, and give risk assessment suggestions to assist credit reviewers in the enterprise rating.

Hehe Information also supports the integration of Qixinbao's business, judicial, and business data with internal customer data (such as order data, flow data, financial data, etc.) The credit amount can be adjusted in real time according to the changes of the enterprise to prevent excessive credit or insufficient credit.

• Credit contract management
  In order to prevent credit risks arising from contracts, banks need to conduct prudent review and management of credit contracts to avoid confusion between old customers' current credit plans and previous plans, and lack of key information on credit applicants. risk. Hehe Information launched the Textin contract robot, based on STR recognition and NLP algorithm, which supports intelligent structured extraction of key contract information and comparison of key information, as well as full-text comparison of contracts, supporting multiple formats of pictures, PDF, and word.
  The key information extracted from the credit contract can be transmitted to the bank credit system through the API interface, automatically fill in the form, and accumulate data assets efficiently. It can also automatically call Qixinbao big data according to the extracted credit application enterprise name field, and prompt enterprise risks in real time.

(4) Post-loan early warning and monitoring

• Risk Public Opinion Monitoring
  Post-lending management is the last step in the credit risk control process. Practice has shown that customers often have signs before the deadline, and early detection of risk signals can make early analysis decisions and early decisive actions. In order to realize early warning of risks, banks need to closely monitor the trends and risks of borrowers. Not only can they provide early warnings before overdue occurrences, predict customer prospects, and pre-empt risks; they can also act quickly after overdue occurrences, Judging the repayment willingness and repayment ability according to the real-time dynamics of customers, formulating strategies and plans to collect and reduce the occurrence of non-performing loans.

  Qixinbao, a subsidiary of Hehe Information, supports 24/7 monitoring of risk information in more than 160 dimensions, including business changes, administrative penalties, judicial risks, violations of laws and regulations, abnormal operations, and negative public opinion, including public opinion information monitoring, Multi-faceted public opinion services including network-wide event analysis, keyword cloud analysis, sentiment analysis, etc. All risk types support custom risk classification, four-level early warning mechanism, and real-time early warning reminders. When a risk signal occurs, the early warning risk manager and account manager can be pushed through various methods such as SMS, email, WeChat, and API interface.

• Grasp the property clues to effectively clear
  the debt If the loan is overdue and enters the collection process, or it reaches the court's compulsory enforcement link, grasping the debtor's property clues will provide the bank with the initiative to recover the debt.
  Qixinbao, a subsidiary of Hehe Information, launched "property clues" to help banks grasp the flow of property clues of defaulting customers and assist in collection work. Property clues provide information on suspected inflows and outflows of eight types of assets: capital, equity, movable property, real estate, commercial income, intangible assets, litigation-related assets, and external claims. At the same time, it provides suggestions on the difficulty of property execution, assists in collection work, and reduces bank losses.

• Blacklist management and credit risk infection prevention and control
  Qixinbao, a subsidiary of Hehe Information, provides blacklist management functions. Banks can set non-performing loan customers as blacklist customers, unify data output caliber through blacklist labels, and create omni-channel blacklists Early warning mechanism. Different business lines can obtain unified customer information on the big data platform in the industry to avoid other business departments from establishing business relationships with blacklisted customers due to isolated information islands.

If a company defaults after the loan, the companies associated with it may also default. This is a typical causal credit risk contagion. Hehe Information empowers banks to shape risk networks based on holding relationships, group relationships, upstream and downstream relationships in the industrial chain, supply chain relationships, and cooperative relationships, intelligently analyze risk transmission paths and influence scores when post-loan risks occur , and timely identify risks There are borrowers in the bank that may be transmitted to , quantify and judge changes in corporate credit risk, and give early warnings to prevent large-scale, multi-subject defaults after credit defaults by large companies.

(5) Customer case

• Nanhai Rural Commercial Bank
  Nanhai Rural Commercial Bank cooperates with Hehe Information to build a Haiying big data application portal, unify the platforms, channels, and standards for data acquisition by business departments, operating departments, management departments, and risk control departments, and empower The core strategy of the digital transformation of the whole bank.

  The Haiying big data application portal is equipped with knowledge graph applications such as corporate portraits, corporate graphs, risk monitoring, and digital large screens, as well as Textin financial reporting robots. Nanhai Rural Commercial Bank can intelligently expand customers in batches through map search and potential customer subscription; through corporate portraits and labels, association relationship and beneficiary identification, risk and public opinion monitoring, etc. Instead of traditional business personnel manually integrating and processing the original data of financial reports, it improves the efficiency of financial report entry and analysis; the digital large screen provides big data analysis tools for relevant personnel to conduct overall planning and assist decision-making, and fully supports the development of various lines of business.

• A joint-stock bank
  A joint-stock bank joined hands with Hehe Information to build a big data risk portal, combining data such as credit and guarantees in the bank with Qixinbao big data under Hehe Information, realizing multi-dimensional search, intelligent risk association, and public opinion information analysis , map association analysis, information intelligence view, internal and external risk assessment and other applications.

After the launch, nearly 10,000 employees in the bank have used the risk portal, including 4,500 account managers and 2,200 risk managers. A total of 55 branches and subsidiaries have used the risk portal to screen corporate risks. Employee feedback, the risk portal makes business personnel no longer need to query enterprise-related information through external third-party platforms, with higher efficiency and richer data dimensions, and through the integration of big data in the industry, the hidden guarantee circle chain and inter-enterprise transaction relations can be realized. Panoramic presentation across branches breaks down information islands.

---

4.2.5 Intelligentization of the five major business scenarios of securities

Within securities firms, middle-office departments such as compliance department and risk management department and front-end business departments such as self-operated business, asset management business, investment banking business, brokerage business, credit business, and research institutes have a large number of target company information inquiries, contract Big data needs such as regulation, risk control, due diligence, investment research, marketing, and public opinion monitoring.

Since each business department generally uses different systems , and there are various roles in each business, such as listed companies, financiers, investors, custodian banks, etc., a large amount of data is generated in isolated systems, which are difficult to share and cannot be aligned. .

At the same time, the internal information of securities companies often cannot meet all the data needs of the business , and it needs to be manually collected and summarized on various third-party websites. Information silos, repeated collection of information, and excessive reliance on labor and expert experience not only reduce business efficiency and increase labor costs, but also easily lead to pain points such as missing important risk information and the inability to reuse expert experience to grassroots risk control posts.

Hehe Information supports the combination of Qixinbao 's 230 million panorama and full-scale enterprise big data with the internal data of securities firms, based on data processing, machine learning, graph storage and calculation, knowledge engineering, information retrieval, information extraction, text detection and recognition And other cutting-edge technologies , integrate internal and external data resources into multi-dimensional knowledge graphs that can be used in depth and empower businesses, and build a unified risk control and compliance big data platform within securities companies.

Each business department of a securities company can efficiently obtain unified multi-dimensional data of the enterprise, deeply explore the relationship between enterprises, and effectively monitor risks and public opinion information, so as to realize the optimization of risk control management technology, improve the efficiency of enterprise analysis, refine the stratification of enterprise customer groups, and improve business quality of decision-making, etc.

(1) Credit risk management

● Business Unit

Risk Management Department, Credit Business Department, etc.

Unified query and display of internal and external data (including risk signals)
Brokers need to accurately identify, dynamically monitor, respond in a timely manner, and manage the entire process of credit risk in their operations. Detailed information and dynamic changes of customers, counterparties, etc.

Hehe Information can integrate the internal accounts, transactions and other data of securities firms with Qixinbao's full industrial, commercial, judicial and taxation databases, and unify the channels for business departments to query corporate data and the data output caliber. Business personnel can quickly locate target companies through enterprise name search, keyword fuzzy search, multi-condition combination screening, etc.

After searching for a company and entering the company details page, business personnel can inquire about the business, shareholders, directors, supervisors, branches, equity investment, financing and other dimensions of the company, as well as the internal information of the securities company. The internal and external information is displayed in one stop without switching System summary information. Risk signals and the identity attributes of customers in various businesses of securities companies are embedded in the company details page in the form of tags to help business personnel quickly grasp the various business relationships between target customers and the company. The details page collects risk signals from multiple sources such as business abnormalities, judicial litigation, administrative penalties, integrity and dishonesty, and bankruptcy liquidation in one stop.

Risk network shaping, in-depth capture of risk transmission

In addition to paying attention to the credit risk of the target company itself, the securities business department also needs to pay attention to the possibility of the credit risk of related companies being transmitted to the target company. Hehe Information empowers securities companies to shape risk networks based on holding relationships, group relationships, upstream and downstream relationships in the industry chain, supply chain relationships, and cooperative relationships. Through the risk network, it can deeply capture risk transmission and intelligently analyze risks when internal customers have credit risks. The conduction path and influence score are used to quantify and judge the changes in the credit risk of the target enterprise.

identified by the same customer

The identification of the same customer is one of the key points of credit risk management for securities companies, especially for financing businesses with relatively prominent risks. It is necessary to provide unified credit to group customers. All companies in the group family tree share the credit line to prevent repeated credit and excessive credit. Risk exposure widens.

Hehe Information provides a group relationship map, uses the full amount of industrial and commercial equity data, and combines the "Large Risk Exposure Management Measures" to carry out modeling, dig deep into the equity investment relationship between enterprises, analyze the factions of enterprise groups, and help securities companies accurately identify groups client.

Intelligent identification of financial reports to improve the efficiency of financial analysis

Financial analysis is an essential part of customer credit risk assessment. The financial report can truly and systematically reveal the financial status, operating results and cash flow of a company for a certain period of time, and then help securities companies analyze customers' profitability, solvency, investment income, development prospects, etc., and provide data reference for various business decisions.

Based on self-developed intelligent text recognition, table recognition, and natural language processing technologies, Hehe Information launched the TextIn financial report robot, which intelligently collects, verifies, and outputs financial report data in a structured manner. It only takes 2-3 seconds to identify a page of financial report information on average, and can Efficiently output financial data in different file formats and report formats in a unified standard format, greatly shorten the time for financial report entry, standardize the data output format, and automatically output capital structure, solvency, Analyze profitability, cash flow and other indicators to improve the efficiency of financial analysis.

Automatic generation of corporate credit reports
Qixinbao, a subsidiary of Hehe Information, supports automatic generation of corporate credit reports, including business, shareholders, foreign investment, judicial, risk, annual reports, intellectual property rights, operations and other data of the company, helping securities companies to systematically sort out complex information, Reduce desk work and comprehensively improve the efficiency of credit evaluation.

02 Compliance Supervision

● Business departments
Compliance and Legal Affairs Department, Risk Management Department, etc.

Identification of actual controllers and beneficial owners
According to the requirements of the "Anti-Money Laundering Law" and the Central Bank's "Notice on Strengthening Anti-Money Laundering Customer Identification (No. 235)", financial institutions need to effectively carry out customer identification for customers who have established business relationships , especially to penetrate and verify the equity structure of non-natural person customers, and to understand the information of the actual controller and beneficial owner of the customer.

For foreign-invested enterprises with complex businesses, especially the subsidiaries or joint ventures established by large multinational enterprise groups in China, the mutual shareholding relationship between their foreign shareholders and their related parties is often dense like a cobweb. Ownership structure is an extremely complex process.

Qixin Baoke, a subsidiary of Hehe Information, generates relationship graphs such as equity penetration, controller/suspected actual controller relationship, etc., and based on the Central Bank’s No. 235 document and anti-money laundering background, it automatically identifies and outputs the beneficial owner behind the company’s backstage, providing judgment The reason and path help securities companies to accurately identify, verify, and monitor the information of the actual controller and beneficial owner of the enterprise, and can be connected to the reporting system to meet the regulatory requirements for anti-money laundering compliance.

Abnormal transaction and abnormal account monitoring

The "Administrative Measures for the Reporting of Large-value Transactions and Suspicious Transactions by Financial Institutions" stipulates that financial institutions need to establish a large-value and suspicious transaction reporting system. , should be submitted to the China Anti-Money Laundering Monitoring and Analysis Center in a timely manner.
Based on the knowledge graph capability, Hehe Information can combine the internal transaction data of securities companies, through mining and analysis of enterprise accounts and transaction data information, and finally obtain the enterprise transaction relationship graph, which helps securities companies to achieve compliance such as multi-person verification, supervision account verification, account correlation analysis, etc. Verification, assisting compliance and risk control personnel to identify and determine abnormal transactions and abnormal accounts.

03Public opinion query and monitoring

● Business Department
Risk Management Department, Credit Business Department, Investment Banking Department, Securities Proprietary Department, Asset Management Department, Office, etc.

Public Opinion Risk Big Data Intelligent Collection
Brokers' multiple business departments need to continuously monitor corporate public opinion with different focuses: investment banking business needs to monitor whether sponsor companies have major violations, information disclosure fraud and other negative public opinion; self-operated and asset management businesses need to monitor Whether the issuer has shareholders, directors, supervisors, senior executives, etc., such as reduction of holdings, changes in senior executives; margin financing and securities lending business needs to focus on monitoring negative public opinion information that affects customers' repayment ability... Traditional public opinion risk information is collected
through multiple authoritative media, industry Keyword search on the website, manual extraction of information, plus reports written after on-site investigations, manual collection of online and offline information, and then credit risk judgments by experts. Hehe Information helps securities companies realize the automatic collection and aggregation of enterprise risk dynamics across the network, expands the source of online risk information, intelligently integrates and analyzes it with offline investigation reports, transforms expert experience into a risk control model, and automatically studies and judges credit risks, greatly reducing Information collection, judgment, and application costs.

One-stop query and monitoring of corporate public opinion

Hehe Information creates a one-stop public opinion query, monitoring and early warning platform for securities companies that better matches business scenarios. It can provide network-wide public opinion information summary and visual analysis covering 230 million companies, including more than 90 dimensions of network-wide public opinion monitoring and topics. Word cloud analysis, positive and negative sentiment analysis, source authority level analysis, and negative public opinion index analysis.

Brokerage business personnel input the name of the monitoring company or its affiliated companies, and they can directly query the public opinion of the relevant company. Each piece of public opinion information has positive and negative neutral sentiment labels, source authority labels, and monitoring dimension labels, and the negative degree of negative public opinion is indexed. , which is convenient for quick screening and review of public opinion. In addition, it also supports customizing the special retrieval of exchange penalty information, inputting the names of key personnel and stock names to query corporate public opinion and other functions.

When the monitoring company has negative public opinion, real-time early warning, dynamic push, push to the business person in charge through various methods such as WeChat, email, SMS, etc., so as to quickly discover and timely respond to hot public opinion and emergencies of customers.

Brokerage reputation risk management

In addition to monitoring the public opinion of corporate customers, securities firms also need to pay attention to their own public opinion trends to prevent reputational risks. The brokerage office can add companies, branches, and subsidiaries to the list of monitored companies. The system monitors public opinion 24/7, statistically analyzes the emotional bias and quantity changes of public opinion, and sends early warnings of negative public opinion in a timely manner.

04Due Diligence & Investment Research

Business Department
Investment Banking Department, Research Institute and other
15+ enterprise relationship graph penetration verification

In the securities issuance and listing sponsorship business of securities companies, it is necessary to conduct due diligence on sponsor companies to control and reduce sponsorship risks; when securities research institutes issue investment research reports, they need to conduct in-depth investigation and analysis of target listed companies. These businesses require a penetrating understanding of multi-dimensional information such as genealogy, equity relationship, investment relationship, executives, shareholders, actual controllers, and beneficial owners of the target company. Due to the intricate relationship between the equity and control structures of many listed companies, penetration verification is difficult and time-consuming.

Based on big data analysis and mining, Hehe Information provides a variety of corporate relationship maps, including: shareholding penetration, accurate analysis of the shareholding structure behind the company, and a clear display of all partners disclosed to the natural person and legal person levels. The identification of related parties is based on the identification rules for related parties in the Shanghai Stock Exchange, Shenzhen Stock Exchange, and the Accounting Standards for Business Enterprises to calculate the relationship between the related parties of the enterprise. The enterprise chain map, based on the enterprise's industrial and commercial and judicial information, calculates the enterprise's shareholders, executives, foreign investment, judicial cases and other information to form a panoramic overview of the enterprise. The enterprise relationship diagram, combined with the personnel's investment relationship and the enterprise's investment relationship data, calculates the enterprises or natural persons related to the target enterprise.



In addition, there are 15+ types of corporate relationship graphs, such as controller relationship, beneficial owner relationship, group relationship, transaction relationship, guarantee relationship, suspected relationship, upstream and downstream relationship, competitor relationship, etc., to help due diligence investment research business to efficiently obtain a corporate panorama information, saving time for in-depth investigations.

Personal Relationship Map Analysis Key Personnel
In addition to building a corporate relationship map, Hehe Information can also help securities companies build a personal relationship map for key personnel such as actual controllers, beneficial owners, shareholders, and executives of the company. Personal relationship graphs can be used in various business scenarios such as due diligence, compliance risk control, and key person marketing.

Hehe Information supports the combination of key personnel's positions and historical positions, personal relationship graphs of legal persons, executives, and shareholders with internal information of securities companies (such as customer mobile phone numbers, addresses, customer numbers, etc.), and can be based on the relevance of internal and external information Analyze and display various reasoning relationships among people (colleagues, relatives and friends, etc.).

05Marketing customer acquisition

Business Department
Risk Management Department, Investment Banking Department, Brokerage Business Department, etc.

Customer portrait construction
Hehe Information relies on big data analysis and mining and knowledge map construction capabilities. It can combine internal and external data to construct existing customer portraits of brokerages. Through modeling mining, recommend potential customers with similar portraits, recommend matching customers and products, and can be used in brokerages. Scenarios such as customer marketing analysis, customer group discovery, customer screening, potential customer mining, etc.

Accurate customer expansion through multiple channels
The multi-business departments of securities companies need to expand corporate customers with different attributes and characteristics. For example, the FA business of investment banks mainly serves start-ups and growing unlisted companies; the bond business is more inclined to central enterprises and state-owned enterprises. Hehe Information provides a variety of customer development tools to help brokers obtain a list of potential customers in batches and accurately, and efficiently contract business development.

---

4.2.6 Intelligent Solutions for Energy Industry

With the orderly advancement of the "dual carbon" goal and the "14th Five-Year Plan", building a clean, low-carbon, safe and efficient energy system has become a key strategy for my country to ensure energy security and promote high-quality economic development. "Digitalization + renewable energy" has become the main theme of the development of the energy industry, which means that the integration of the energy revolution and the digital revolution is an inevitable trend.

Digital technology is profoundly changing the energy system. By optimizing the resource allocation capabilities, security capabilities, and intelligent interaction capabilities of energy production, transmission, trading, and consumption, energy companies can achieve digital and intelligent operation and management.

Recently, CGNPC, China Railway Group, Sinopec, State Grid and many other energy central enterprises have made it clear that digital transformation will be a key task during the "14th Five-Year Plan" period, and at the same time announced the road map, setting an example for other energy companies and reference.

The smart energy solution of Hehe Information, based on digital technology represented by artificial intelligence and big data, drives energy companies to restructure their business models and change through digital transformation in the fields of new energy industry layout, customer management, user service, and internal management. Management mode to realize industrial transformation and upgrading and value growth.

01Adjust the energy structure and deploy new energy

Under the double carbon target, photovoltaic, wind power, nuclear power and other new energy markets will undoubtedly usher in a broader space for development. For large-scale energy companies, it is necessary to do a good job in planning and planning in advance, actively adjust the energy structure, and lay out new energy development, new technology R&D and application, energy services and other fields.

Qixinbao, a subsidiary of Hehe Information, launched 240+ industrial chain topics, including: photovoltaic, wind power, biomass power generation, nuclear power, geothermal energy, shale gas, hydrogen energy and other new energy industry chains.
Energy enterprises can interactively view the upstream, midstream and downstream links of each industrial chain, corresponding enterprises, and geographical distribution, etc., and accurately grasp the overall layout of the industrial chain, industry-enterprise affiliation relationship, product-product upstream and downstream relationship, and enterprise-enterprise upstream and downstream relation. It can also combine enterprise ratings and enterprise labels to comprehensively screen out high-quality industrial chain enterprises, so as to expand upstream and downstream partners, empower suppliers to find sources, and develop cooperation projects.

Taking the photovoltaic industry chain as an example, the upstream of the industrial chain is minerals, silicon materials, silicon products, auxiliary materials, and photovoltaic equipment; the midstream is cells, cell auxiliary materials, photovoltaic modules, and module packaging auxiliary materials; the downstream is photovoltaic power stations, photovoltaic buildings , Photovoltaic power generation system integration, photovoltaic power generation system support equipment, photovoltaic product testing and certification. Qixinbao's photovoltaic industry chain topic includes a total of 49,997 companies, and the number of companies involved in the upstream, midstream and downstream links has increased significantly in turn, and the industrial chain has a pyramid-shaped structure.

From a national perspective, Jiangsu Province has 7,359 photovoltaic industry chain enterprises, accounting for 14.72% of the country, and has a clear advantage in the upstream and midstream links. In terms of downstream applications, Shandong Province takes the lead, with 3,999 downstream enterprises in the photovoltaic industry chain, especially in the fields of centralized photovoltaic power stations and solar power operation and maintenance.

Qixinbao’s industry chain topic can display the list of companies corresponding to each link of the industry chain in a panoramic manner, and supports sorting by industry chain relevance, Qixin score, and registered capital, and can also be screened based on enterprise qualifications to help energy companies quickly and efficiently lock in Potential high-quality partners in the target area, reducing the cost of selection time. Click on the company name to check the company details: industrial and commercial information, business information, intellectual property rights, risk information, etc., one-stop insight into the strength and risks of the company, and complete the preliminary access investigation while searching for sources and customers.

02Partner risk management

Large-scale energy companies are often comprehensive groups with a wide range of business coverage, which also leads to numerous systems within the energy group, even thousands. For example, it is reported that there are more than 1,000 systems in Sinopec, which are generally implemented by the technical department after the business department puts forward the requirements. However, after the system is completed, one after another is formed. The information cannot be shared and the cost remains high.

Diversified business types make energy companies have a large number of partners, but the lack of a unified data platform also makes it difficult for partners in various departments to communicate with each other. There are often conflicts, and it is difficult to achieve unified management and risk control.

Hehe Information helps energy companies realize risk management throughout the life cycle of partners based on the capabilities of partner qualification certificate OCR, panoramic enterprise information query, relationship investigation, and risk monitoring. Through the establishment of a unified big data query platform at the group level, all branches/subsidiaries and business departments can query and obtain multi-dimensional enterprise business information, and keep the dynamic update of information consistent with the full caliber, enabling full business scenarios.

Access investigation and qualification review
In the access stage, it can see through the panoramic enterprise information of partners, covering multiple dimensions such as industry and commerce, justice, taxation, intellectual property rights, assets, public opinion, etc., and efficiently review access qualifications, including whether the industry and commerce and qualification information meet the requirements. Shortlisting requirements, whether the qualification certificates are complete, whether there are potential major risks, etc.
Batch OCR intelligent text recognition can be performed on various qualification certificates provided by partners, information is automatically collected and entered into the system, and the enterprise information in Qixinbao is called according to the enterprise name, and the system intelligently compares the basic access requirements with the qualifications of partners Does it match. It can also automatically generate supplier/distributor access audit reports to improve the efficiency of due diligence.

Bidding compliance management
In the bidding stage, it is possible to conduct compliance checks on various internal and external relationships of bidding suppliers, such as: check the relationship between suppliers and existing suppliers to avoid dominance by one company; check suppliers and blacklisted companies to prevent blacklisted companies from using affiliated companies to bid; to check the relationship between suppliers and internal shareholders and executives of the company to avoid risks such as commercial bribery and profit transfer; to check the cross-shareholding among multiple bidding suppliers, and the interaction between directors, supervisors and senior management Positions, relationships with group companies, etc., to avoid risks such as bid rigging, bid collusion, and accompanying bids.

Risk monitoring and related risk transmission

In the contract performance stage of partners, carry out 7×24-hour risk monitoring on suppliers’ industrial and commercial changes, administrative penalties, judicial risks, business abnormalities, negative public opinion and other dimensional information, customize four-level risk early warning, real-time early warning, risk pre-warning, and prevent The supply chain is suddenly broken due to a major risk event of a partner, and the reputation of the company is damaged.

In the event of a risk signal, it will be pushed to the person in charge of the business through various methods such as WeChat, email, SMS, and internal letter. It can also intelligently judge the possible spread of risk events caused by partners based on the risk transmission network constructed by various corporate relationships. .

03Online business processing to improve user experience

In order to better serve energy end users, many energy companies have launched APP services. On the APP, users can independently complete operations such as opening services, paying fees, and inquiring, without having to go to offline business halls to complete operations, realizing online The one-certification process has greatly improved customer satisfaction and reduced the pressure on offline services.

Based on the self-developed intelligent text recognition technology, Hehe Information helps energy companies to integrate the recognition ability of bank cards, ID cards, business licenses, value-added tax invoices and other credential materials in the APP, and realizes the rapid and automatic extraction and entry of information. Users only need to upload material images in the APP, and the system can complete a series of automatic operations such as image optimization, structured extraction, information filling, authenticity verification, etc., eliminating the cumbersome and inefficient offline processing and manual data processing , It is also more in line with the trend of "contactless service" during the epidemic, and has realized the upgrade of digital infrastructure.

Typical Service Scenario

• Resident scene
  Residents can upload and identify ID cards and bank cards, face recognition and biometrics, register APP real-name users, quickly complete information authentication and card binding, and can perform online recharge, payment, query and derivative insurance, shopping and other services handle.

• Enterprise scenario Enterprises
  and institutions can easily open and use services such as enterprise payment and electronic invoices through uploading and identifying business licenses and bills, and can provide derivative services such as energy consumption analysis and finance.

• Scenarios for new energy customers
  PV customers can upload and identify their ID cards and bank cards, and obtain services such as site construction consultation, PV application and installation, and subsidy settlement provided by energy enterprise APPs. 04 Data assetization, more efficient internal management.

Energy companies will generate a large number of business documents during their operations, including: documents, invoices, certificates, contracts, etc. If these documents only exist in the form of paper or scanned pictures, it will not only be difficult to manage, but also difficult to retrieve and unable to In order to maximize the value of data, digitization is urgently needed.

For example, only in the business of supplier management, when a supplier registers, it is necessary to enter the information of the supplier's business license and qualification certificate into the system; when signing a contract with the supplier, it is necessary to review the procurement contract and electronically manage the printed contract; When the merchant clears accounts, it is necessary to enter the invoice information, and perform three-way verification and approval of the order, delivery note, and invoice information.

If the above-mentioned business process relies solely on manual information collection, entry, and verification, it will consume a lot of manpower, be inefficient and error-prone.
Based on self-developed complex scene text recognition, complex form recognition technology and advanced AI algorithm, Hehe Information provides energy companies with a variety of mature OCR products, covering general text recognition, general form recognition, 100+ kinds of document recognition, bill recognition , Contract identification and comparison, etc. It has high-precision recognition ability for printed and handwritten characters in pictures, face sheets, certificates, documents, forms and other scenes.

Support one-stop identification of value-added tax invoices, machine-printed invoices, quota invoices, bank receipts and bills, pound slips, POS receipts and other bills; identification of ID cards, bank cards, military certificates, driver's licenses, driving licenses, and business licenses and other certificates; identify complex forms such as multiple forms, no frame lines or tight form lines; for documents containing seals, it can also perform seal existence judgment and seal strengthening, flattening, extraction, elimination, etc., to help energy companies realize efficient Internal documents are digitized, business certificate review and information entry and filing are automated, which greatly improves business efficiency.

For non-standard and personalized documents and vouchers, Hehe Information's text recognition training platform can also provide the ability to independently create, train, and deploy OCR models, helping energy companies to conveniently and efficiently realize the structured extraction of document and voucher information.
In addition, the AI ​​management platform of Hehe Information integrates AI capability access, API management, business distribution, monitoring statistics and other capabilities, which can help energy companies to manage OCR products in a unified way, and meet the needs of branches/subsidiaries to call different OCR capabilities Demand, improve the stability of OCR operation, create a safe AI service ecosystem, and automatically bill based on the amount of calls, which is convenient for enterprise cost management.

04 Customer case

• A large energy group
  A large energy group started from the city gas business. After continuous business expansion and strategic upgrading, it has formed a complete natural gas industry chain of mining, storage, transportation, and distribution, as well as quality products covering tourism, culture, health, and real estate. life product chain.
  The group's internal documents and certificates are numerous and complicated. With the help of the Hehe information bill robot, the automatic classification and identification of the group's bills has been realized, and the inefficient and cumbersome traditional manual review and entry has been changed in the areas of cost control management, tax analysis, purchase and sales item management, and financial reimbursement. , Significantly reduce costs and increase efficiency, enabling digital transformation of enterprises.
  In order to strengthen the supply chain management capability, the group has also joined hands with Hehe Information to create a digital and integrated supply chain risk control platform, which can provide panoramic perspective and real-time monitoring of key information such as supplier credit status, personnel changes, and risk public opinion within the supply chain. Dimensional review of the behavior of target companies and affiliated companies, realizing intelligent and efficient corporate due diligence.

• A management investment group

  The main business of a management-oriented investment group covers commodity trading, medicine and health, real estate investment and financial services. In contract signing, merger and acquisition transactions and other businesses, it is necessary to control and manage corporate credit risks.

In order to solve the pain point of incomplete and untimely acquisition of corporate information, the group cooperated with Hehe Information to create an enterprise risk information management system used within the group, which realized corporate credit investigation, related relationship investigation, risk dynamic monitoring, member companies Management and other core functions reduce the management costs of corporate credit risk, archives risk and operational risk. Through artificial intelligence and big data, "wrap up more than 300,000 contracts" to enhance the value of legal affairs, so that "information will not be distorted, leaders will not fail to monitor, and enterprises will not lose control."

---

4.2.7 Bank Account Service Solutions for Small and Micro Enterprises

The central bank issued the "Guiding Opinions of the People's Bank of China on Optimizing Bank Account Services and Risk Prevention and Control for Small and Micro Enterprises" (hereinafter referred to as the "Guiding Opinions").

The "Guiding Opinions" pointed out that it is necessary to adopt differentiated due diligence methods for small and micro enterprises, promote the easy account opening of small and micro enterprises, and at the same time strengthen the risk prevention and control capabilities of bank accounts to solve the problem of difficulty in opening accounts for small and micro enterprises.

It is necessary to simplify the difficulty of account opening for small and micro enterprises, reduce the requirements for supporting certification materials, and strengthen risk prevention and control, and identify accounts involved in gambling. This undoubtedly puts forward higher requirements for banks' ability to implement big data and new technology applications.

Based on more than 100 billion pieces of real-time dynamic commercial big data and big data analysis, mining and visualization capabilities of more than 100 billion pieces of real-time dynamic business big data and big data analysis and visualization capabilities gathered by Qixinbao, a subsidiary of Qixinbao, it can help banks quickly obtain Core information such as corporate business, taxation, judicial litigation, etc., prompts major risks, avoids requiring too many proof materials from companies, and at the same time, through business licenses, legal person ID cards and other documents and bank flow OCR identification, speeding up the account opening process and improving services At the same time of quality and efficiency, it efficiently implements due diligence and manages the bank accounts of small and micro enterprises throughout the life cycle.

1. Simple account opening, efficient due diligence and quick investigation of suspicious features

Take a differentiated approach to customer due diligence. Banking financial institutions (hereinafter referred to as banks) shall, on the basis of "knowing your customers", follow the "risk-based" principle to determine the specific methods of due diligence on customers of small and micro enterprises (including individual industrial and commercial households, the same below), and shall not " "One size fits all" requires customers to provide supporting certification materials, and shall not make unreasonable or beyond necessary identity verification requirements to customers.

Promote the simple account opening service for small and micro enterprises. Banks shall review the account opening certification documents of small and micro enterprises in accordance with regulations, and shall open an account if the purpose of opening the account is reasonable and there is no obvious reason to suspect that the account is engaged in illegal and criminal activities. According to the needs of small and micro enterprises, simple account opening services can be provided, the requirements for auxiliary certification materials can be simplified, and background data verification can be strengthened. Account function settings should match the degree of customer identity verification and account risk level. In the future, account functions can be upgraded according to customer due diligence. Banks are encouraged to refer to the "Guidelines for the Simple Account Opening Service of Bank Accounts for Small and Micro Enterprises" before the end of 2021, and complete the comprehensive implementation of the simple account opening service on the basis of the previous pilot.
--"instructions"

This simple account opening service refers to a bank that simplifies the requirements for supporting supporting materials after reviewing the account opening certification documents of small and micro enterprises in accordance with the "Administrative Measures for RMB Bank Settlement Accounts" and other regulations. The account opening function matches the customer's identity verification degree and account risk level. The basic deposit account meets the needs of customers for opening accounts.

Simple account opening does not mean "no requirements" for small and micro enterprises to open accounts. Banks still have to follow the "risk-based" principle, fulfill anti-money laundering, anti-terrorist financing, and anti-tax evasion obligations, implement the account real-name system, and effectively identify, evaluate, monitor, and control accounts business risk. If obvious suspicious features are found during the account opening process, the simple account opening service is not applicable.

This requires banks to conduct fast and efficient due diligence on small and micro enterprises during the account opening process. On the basis of basic materials, they can accurately verify customer identities and classify account risk levels based on query, matching, and verification of background data.

Qixinbao, a subsidiary of Hehe Information, gathers more than 100 billion pieces of real-time dynamic commercial big data from 230 million domestic enterprises and other organizations, and can provide more than 1,000 data feature labels including industry and commerce, equity, judicial involvement, dishonesty, public opinion, assets, etc. . In the account opening scenario, the teller can quickly query the multi-dimensional details of small and micro enterprises by entering the company name, including basic industrial and commercial information, business information, judicial information, tax information, etc., and can also check the relationship between the company and internal and external blacklisted companies. The enterprise relationship map is convenient for cross-checking information with existing materials and quickly troubleshooting suspicious features.

In addition, based on the KYC overview, you can quickly grasp the risk situation of the main company, affiliated personnel, affiliated companies, and suspected affiliated companies, and combine the Qixin score and the shell index to quantify and judge account risks, and complete corporate due diligence efficiently.

2. Account opening certification documents OCR documents are electronically streamlined and paper materials are reduced

Use technological means to improve the service level of corporate bank accounts. Encourage banks to open electronic channels for small and micro enterprises to open accounts and make reservations. Support small and micro enterprises to submit account opening certification documents online, minimize paper materials, and reduce the number of forms and signatures. Actively promote the application of electronic business licenses and electronic signatures in bank account opening and other links. On the premise of effectively identifying the customer's identity, it is encouraged to support the online handling of the change and cancellation of the bank account of small and micro enterprises.
--"instructions"

Whether it is a counter or an online account opening scenario, the company needs to submit basic materials such as business license, legal person ID card, official seal, financial seal, and account opening form. Based on advanced intelligent text recognition technology and AI algorithm, Hehe supports complex scene text recognition, complex form recognition, seal existence judgment and recognition, seal image optimization, and high-precision structured recognition of documents, forms, business licenses and other enterprises Documents, automatically enter document information in different file formats into the system, automatically fill in forms, realize the automation of account opening business processes, digitize documents throughout the account opening life cycle, and greatly improve the efficiency of account opening.

In addition, in order to prevent fraudulent corporate certificates and submitted certificates that are not the latest and valid certificates, etc., Hehe Information can also call its Qixinbao big data through the API interface, and realize "certificate OCR identification-automatic form filling-and Qixinbao" in the banking business system. Xinbao Big Data Consistency Verification—Mismatched Data Automatic Error Correction Prompt" is a one-stop enterprise account opening information entry and verification review, without switching or jumping to different system platforms, which greatly improves the efficiency of enterprise account opening certification materials review, strict Prevent certificate fraud and invalidation risks.

3. Master the multi-dimensional information account classification and hierarchical management of small and micro enterprises

Establish an account classification and grading management system. Banks should establish a classification and grading management system for small and micro enterprise bank accounts by the end of 2021, provide account functions that match the degree of customer identity verification and account risk level according to industry characteristics, enterprise scale, and operating conditions, and prudently agree with customers that non-counter Face-to-face business, and reasonably set the non-counter channel fund transfer limit, transaction number, verification method, etc., which can be dynamically adjusted according to the normal and reasonable needs of customers or temporary needs, account risk conditions, etc.
--"instructions"

To implement classified and hierarchical management of small and micro enterprise accounts, the key point is to understand the industry, enterprise scale and operating conditions of small and micro enterprises. Due to the serious information asymmetry of small and micro enterprises, it is often difficult for banks to clearly grasp the real situation of small and micro enterprises and make risk assessments.

Qixinbao, a subsidiary of Hehe Information, not only provides enterprise information inquiry services such as enterprise scale and business information, but also helps banks deeply understand the industry and characteristics of small and micro enterprises through rich enterprise labels, industry labels, and industry chain topics. It is an emerging industry and a scientific and technological innovation enterprise that is quite different from traditional enterprises.

insert image description here

In addition, based on the cleaning and mining of commercial big data, Qixinbao establishes an enterprise scoring model, and calculates the Qixin score based on data such as growth, intellectual property, enterprise scale, risk status, operating quality, and capital background. Ranking, quantifying enterprise risk levels, and giving risk assessment suggestions, providing credible data support for account classification and grading.

4. Abnormal transaction monitoring enterprise map association investigation

Identify and control fraudulent and gambling accounts. Banks should continue to improve the risk monitoring model based on the characteristics of fraud-related and gambling-related accounts, and include accounts with abnormal opening and transaction conditions into the scope of key monitoring. Suspicious accounts involving fraud and gambling that are discovered through monitoring and cannot be ruled out after verification, shall take appropriate control measures in a timely manner in accordance with laws and regulations, depending on the situation, and transfer them to the local public security organs.

Continue to carry out the verification of accounts involved in the case. Banks are supported in carrying out reverse investigations of the accounts involved in the case transferred by the public security organs, related investigations into other bank accounts opened by the enterprises involved in the case and their relevant personnel, and taking appropriate control measures for suspicious accounts. Encourage the establishment of a regular analysis system for accounts involved in cases, find loopholes in risk prevention and control and improve the risk prevention and control system.
--"instructions"

After the account is opened, the bank should continue to monitor the account transactions, discover, investigate, and report abnormal transactions and accounts in a timely manner. Hehe Information's bank statement identification and analysis products can identify bank statements based on high-precision OCR. Through the analysis of transaction counterparties, related parties, suspected related parties, and transaction content, it can intelligently judge the rationality of transactions and identify suspicious transactions in multiple dimensions. It can also mine and analyze corporate accounts and transaction data information based on the knowledge graph capability, and finally obtain a graph of corporate transaction relationships, and conduct more effective monitoring and analysis on the source, amount, frequency, flow direction, and nature of corporate transactions to assist banks in compliance Risk control personnel identify and determine accounts involved in fraud and gambling.

When carrying out investigations of accounts involved in the case, the company involved in the case and its personnel can be screened through various types of corporate relationship graphs and personnel graphs provided by Hehe Information, such as equity relationships, suspected controller relationships, beneficial owner relationships, group relationships, and corporate chain graphs. Other bank accounts opened by the relevant person.

Since the opening of the account involved in the case is often concealed, the actual associated person or affiliated enterprise may not constitute an associated relationship on the surface. Hehe Information also provides a graph of suspected relationships, presenting companies that are suspected to have relationships through features such as the same judgment documents, patents, phone numbers, email addresses, and domain names, and helping bank compliance personnel dig deep and trace possible relationships that may exist behind them.

5. Full life cycle management of small and micro enterprise accounts

Pre-verification / in-process monitoring / post-event investigation

Strengthen account lifecycle management. Banks shall establish and improve a full life cycle management mechanism for small and micro enterprise bank accounts before, during, and after. Use valid data to cross-verify the customer’s identity in advance, and the customer promises to use the account legally and compliantly; strengthen the identification and control of fraud-related and gambling-related transactions during the event; and strengthen the investigation and clean-up of existing accounts and the accountability of fraud-related and gambling-related accounts after the event.
--"instructions"

Optimizing small and micro enterprise account services is not only simple account opening, but also on the basis of simple account opening, the whole life cycle management of small and micro enterprise accounts, through information verification, risk identification and monitoring, investigation and cleaning in the whole process before and after the event, While simplifying the account opening process, strengthen account control and implement account compliance requirements such as anti-money laundering.

Qixinbao big data service under Hehe Information provides massive enterprise big data support and convenient management tools in the full life cycle management of small and micro enterprise accounts.

In advance, by using more than 100 billion pieces of real-time dynamic commercial big data and more than 1,000 data feature labels collected by Qixinbao from 230 million domestic enterprises and other organizations, you can in-depth query the multi-dimensional information of enterprises, combine account opening materials with structured identification, cross Verify customer identity and conduct efficient due diligence.

During the event, based on Hehe information bank flow identification and analysis of product and transaction relationship graphs, abnormal transactions are closely monitored, transactions and accounts involved in fraud and gambling are identified, and various relationship graphs can be used to check the related persons and affiliated enterprises of the accounts involved in the case. other bank accounts.

Afterwards, relying on Qixinbao account annual inspection, regular medical examination, risk dynamics and other functions, regularly check whether the industrial and commercial operation status of the enterprise to which the stock account belongs, whether the business period has expired, whether it has been included in the list of seriously illegal and untrustworthy enterprises, whether it has been listed Enter into the list of abnormal business operations, etc., and clean up suspicious accounts in a timely manner.

---

4.2.8 Credit Granting Solutions for Commercial Bank Group Customers

The unified credit management of group customers is the focus and difficulty in the current credit management work of commercial banks. Hehe Information once assisted a bank in maintaining group customer data. After sampling 10 group samples, it was found that nearly 1,000 group member companies were not designated by the bank account manager as a group, and another nearly 100 companies were mistakenly classified into other groups. group.

A group is a group of corporate law groups with a certain scale and organic connections (that is, a family tree) formed through capital investment, management control, or family associations, etc. Group.

Since high-quality group customers are the targets of all banks competing for cooperation, the competition is fierce, and the negotiating position of customers is constantly improving. It often happens that internal related parties do not transfer assets or profits according to the principle of fair price, which brings huge losses to the bank.

As a guiding document for group customer credit management, the "Commercial Bank Group Client Credit Business Risk Management Guidelines" proposes credit objects, credit business scope, credit principles, credit business risk management, information management and risk early warning. However, in banking practice, due to complex customer group relationships, frequent changes in equity relationships, frequent related transactions, hidden related guarantees, etc., it is extremely difficult for banks to implement unified credit management for group customers, and many banks have received approval from the China Banking Regulatory Commission for this ticket.

The risk of group customers is systematic, concealed and destructive. However, at the same time, the credit granting to group customers is often inefficient due to the complex and heavy inter-agency collaboration workflow within the bank. How to balance the efficiency of credit granting to group customers and the Dual goals of risk concentration control?

01Construct a group family tree and update it dynamically with big data

Group relationship map
The first step in the unified credit granting of group customers is to grasp the overall picture of the group and identify which enterprises are group customers.
The scope of group customers is defined in the "Commercial Bank Group Customer Credit Business Risk Management Guidelines". However, just like the case mentioned at the beginning of this article, in banking practice, it is difficult for banks to identify group customers, and the proportion of "fish that slipped through the net" is very high. This is mainly because of two reasons:

• 1. Due to the complex group relationship of the enterprise, there are information asymmetry and technical difficulties in the identification, the account manager mistakenly judged the group relationship of the enterprise during the due diligence process, and regarded it as a single customer credit.
• 2. Since the equity of an enterprise may change at any time, from an industrial and commercial perspective, an enterprise can leave or join a group at any time. The bank lagged behind in capturing information on equity changes, did not update group customer information in a timely manner, and mistakenly divided the enterprise into the original group based on the original information.

The criterion for judging group customers is not whether the company is named after the "group", but the principle of "emphasizing substance over form" should be used to define group customers by association. The organizational form of a standard group company often has a parent company, the parent company sets up several subsidiaries, and forms a group enterprise through the shareholding relationship of the parent company and the subsidiary company.

In addition to this standard group customer, the bank should also dig out other companies with related or even implicit related relationships and determine them as group customers, such as: companies controlled by actual controllers through direct or indirect shareholdings, actual shareholders Shareholders entrust natural persons or corporate legal persons to hold shares and control enterprises on behalf of shareholders through entrustment agreements.

Hehe Information can build a group relationship map based on equity investment, directorship, and bank internal customer data (such as kinship), helping banks grasp the group's panoramic relationship and relationship path, and relying on big data to technically solve the problem of difficult group customer identification.

The group relationship map is updated in real time based on dynamic big data. For example, when the company's equity changes, the system automatically recalculates the equity path, adjusts the group relationship, avoids the risk of information acquisition time lag, promotes the bank group family tree governance, and solves the "difficulty in identifying group customers." "The problem.

02 Group customer financial report entry analysis

Credit granting to group clients needs to analyze the consolidated financial status of the group and the respective financial status of the member companies. It is necessary to consider the group’s main operating income, net profit, cash flow status and group net assets, debt ratio and debt structure, and to determine the group’s comprehensive For the credit line, it is also necessary to comprehensively consider the independent repayment ability of a single customer in the group, the risk of loan collusion, the risk of related guarantee capabilities, and the contagiousness of risks, and prudently determine the allocated credit line of each credit entity, taking into account the overall risk of the group and the risk of a single account control.

For this reason, the bank needs to enter the three consolidated financial statements of the group and the financial statements of the member companies in the last three years. The workload of financial statement entry is heavy, and a lot of manpower is required for tedious financial statement data entry. The efficiency is low, which is not conducive to the development, maintenance and Credit cooperation relationship with key high-quality customers.

Hehe Information launched the TextIn financial report robot, which can input a page of financial report in an average of 2-3 seconds. It collects and recognizes the data of the three financial tables in a structured manner, and has a high recognition accuracy. The report type can also be quickly and accurately identified and processed.

The TextIn financial report robot is easy to operate. By matching built-in financial rules, subject matching, and trial balance verification, it can efficiently output financial data in different file formats and report formats in a unified standard format, greatly shortening the financial report entry time and standardizing data. The output format improves the efficiency of financial analysis and credit approval.

03 Risk investigation of "financial report whitewashing"

In banking practice, it is not uncommon for group customers to "whitewash" financial reports, such as: transferring interests to a single entity, and using this entity to declare credit lines; concealing related transactions to inflate income and profits; hiding financial information of some companies within the group. These "financial report whitewashing" behaviors make the issued financial reports unable to truly reflect the operating conditions of the group's customers, leading to deviations in the bank's judgment on the group's financial situation and risks, which can easily lead to excessive credit extension.

1. Check the "consolidated scope"
consolidated financial statements, which refers to the financial statements that reflect the overall financial status, operating results and cash flow of the enterprise group formed by the parent company and all its subsidiaries.
Some companies within the group do not consolidate financial statements. If the financial reports of the member companies are simply summed up, due to the existence of double-measured assets and a large number of related transactions, it is easy to generate revenue and inflated profits. In some groups, in the consolidated financial statements, Failure to include all member companies, especially companies that appear to have no equity relationship but are controlled by the same actual controller, are easily excluded from the scope of consolidated financial statements.

Based on the group relationship graph, the bank can review the scope of consolidation of the group's consolidated financial statements, and check whether there are member companies that should be included in the consolidated statements but have not been included.

2. Investigate "connected transactions"
companies obtain glamorous performance in financial reports through self-buying and selling related transactions, which is especially common among listed companies. Group customers may regard the transfer of products between affiliated companies as external sales, falsely increase revenue, or use affiliated companies outside the scope of consolidation to increase product prices through transfer pricing, falsely increase profits, or transfer profits Beautify the financial statements of core enterprises, hoping to obtain high credit through core enterprises.
Qixinbao, a subsidiary of Hehe Information, can calculate the related party relationship of the enterprise according to the identification rules of the Shanghai Stock Exchange, the Shenzhen Stock Exchange, and the accounting standards for enterprises, and assist banks in identifying and verifying the proportion of related parties in the company's related transactions and transactions.

At the same time, with the help of Hehe Information's OCR capability, it can quickly identify documents and documents such as written agreements, transportation invoices, and warehouse receipts involved in related transactions, and focus on checking whether the transaction price is fair and whether the related transaction is true, so as to prevent enterprises from passing related transactions. Inflated sales and profits.

04Identify related mutual guarantees and guarantee circle chains

In order to meet the guarantee requirements for bank credit, group customers often establish mutual guarantee relationships among enterprises within the group, and obtain bank loans through related guarantees. However, due to the large number of related guarantees within the group, the internal enterprises form a chain of guarantee circles, and the systemic risks of the group enterprises cannot be effectively spread outward. Once a problem occurs in one enterprise, it may trigger a domino effect, and the risk will spread within the group through the guarantee chain. Circulation, transmission, and amplification lead to a crisis for the entire group.

Hehe Information can build a guarantee relationship map by mining enterprise guarantee information (including PBOC credit information), combined with bank internal customer guarantee information, and internal and external data, mining guarantee relationships layer by layer, identifying guarantee chains, guarantee circles, mutual guarantees, Special guarantee relationships such as self-insurance assist banks to comprehensively review and evaluate the risk mitigation capabilities of various credit guarantees in the unified credit granting of group customers, and avoid actual guarantee shortages and guarantees that are mere formalities caused by related mutual guarantees and guarantee circle chains .

05 Monitor the flow of funds and strictly prevent embezzlement of loans

In order to improve the efficiency of capital use, many groups will adopt the management mode of unified allocation of funds by the headquarters in finance. The credit funds of member companies enter the group's overall fund pool through various channels, and then the group as a whole arranges the use of funds. This model greatly increases the risk of loan misappropriation. There are also groups that exaggerate their actual credit needs, obtain low-cost funds that far exceed their own needs, and use them for inter-enterprise lending to obtain interest rate differentials. These behaviors have seriously increased the difficulty of credit risk management for banks. '

Therefore, banks need to closely monitor the flow of credit funds after they are issued. Hehe Information has launched a bank statement identification and analysis product, which identifies bank statements based on high-precision OCR, automatically verifies the authenticity and integrity of the transaction, and intelligently judges that the transaction is reasonable through the analysis of transaction counterparties, related parties, suspected related parties, and transaction content. Suspicious transactions are identified in multiple dimensions to avoid risks such as fund backflow, fund collection, illegal inflow of funds, illegal fund-raising, and money laundering.

In addition, Hehe Information can also help banks map their internal transaction data knowledge, build a transaction relationship map, and help banks more effectively monitor and analyze the source, amount, frequency, flow, and nature of customer transactions.

If it is found that credit funds are circulating within the group, or the funds are not used for production and operation, the actual use of funds is inconsistent with the loan contract, or even flow to key regulated industries such as real estate, the bank may stop lending or even reduce the credit line.

06Internal and external, cross-branch data connection

Large or very large groups across regions are cross-branch customers for banks. Group member companies have established business contacts with local bank branches, which led to each branch keeping some member company data, such as: the use of credit lines in the past, performance records, comprehensive returns, guarantee records, etc., all of which are Important reference data for unified credit granting to group customers.

If the cross-branch customer information and business information cannot be summarized and analyzed in a unified way to form a unified group relationship, guarantee relationship, transaction relationship, etc. of the whole bank, the unified credit management of group customers will lack the necessary data support.

Through privatization deployment, Hehe Information can help banks realize cross-branch data connection and integrated utilization on the one hand, output the business data of the whole bank in a panoramic and unified form, and systematically sort out data assets in the bank. The organic integration and cross-validation of data and internal data, the introduction of external databases such as industrial and commercial information, judicial litigation, enforcement of dishonesty, business information, tax information, intellectual property rights, negative information, etc., enrich the bank's data asset lake and big data application tools. Due diligence, credit approval, risk monitoring and other risk control links provide more powerful data support.

In addition to privatization deployment, Qixinbao big data products under Hehe Information also support various deployment methods such as visualization plug-ins, SaaS, and API interfaces.