An AI-based cross-platform intelligent procurement sourcing and price comparison system

Abstract

This invention relates to the field of intelligent procurement technology, specifically an AI-based cross-platform intelligent procurement sourcing and price comparison system. It includes modules for cross-platform data acquisition, AI-powered intelligent sourcing, multi-dimensional price comparison analysis, intelligent approval, procurement execution, settlement and reconciliation, data review, and a central control module. Each module communicates bidirectionally with the central system via a distributed network. The central system includes a main control server, a distributed database, an API gateway, and a data encryption unit, ensuring unified data scheduling and security. The cross-platform data acquisition module standardizes data processing across multiple platforms; the AI-powered intelligent sourcing module accurately matches requirements; the multi-dimensional price comparison module comprehensively evaluates price, quality, and service; the intelligent approval module completes multi-level reviews according to rules; the procurement execution and settlement and reconciliation module achieves a closed-loop process; and the data review module provides decision support. This system boasts a high degree of intelligence, comprehensive process coverage, and can improve procurement efficiency and cost-effectiveness while ensuring compliance.

Description

Technical Field

[0001] This invention relates to the field of intelligent procurement technology, specifically to an AI-based cross-platform intelligent procurement sourcing and price comparison system. Background Technology

[0002] In the process of digital transformation of enterprise procurement, traditional procurement models face multi-dimensional bottlenecks, and existing procurement systems are unable to meet the needs of end-to-end intelligentization and cross-platform integration. On the one hand, the sourcing scope is limited to a single platform or fixed suppliers, lacking the ability to mine resources from multiple channels, and the cross-platform data formats are heterogeneous, requiring manual screening and integration, which is inefficient and prone to errors. On the other hand, price comparisons often focus on a single price dimension, ignoring key indicators such as quality and service, resulting in an imbalance in the cost-effectiveness of procurement, while lacking an AI-driven precise demand matching mechanism, leading to serious resource waste.

[0003] The approval process relies on manual transmission and judgment, lacking a standardized rule engine. This leads to delayed compliance verification, low approval efficiency, and a high risk of non-compliant procurement. In the procurement execution phase, order placement, logistics tracking, and goods receipt confirmation are disconnected, resulting in fragmented and difficult-to-trace data. Settlement and reconciliation require manual order aggregation and invoice verification, with time-consuming discrepancies handling and insufficient unified settlement capabilities. Furthermore, the fragmented data across the entire procurement process, lacking in-depth analysis and visualization, fails to provide data support for optimizing procurement strategies and hinders the achievement of cost reduction and efficiency improvement goals. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] To address the shortcomings of existing technologies, this invention provides an AI-based cross-platform intelligent procurement sourcing and price comparison system.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention provides the following technical solution: An AI-based cross-platform intelligent procurement sourcing and price comparison system of the present invention includes a cross-platform data acquisition module, an AI intelligent sourcing module, a multi-dimensional price comparison analysis module, an intelligent approval module, a procurement execution module, a settlement and reconciliation module, a data review module, and a system central control module. Each module communicates bidirectionally with the system central control module through a distributed network.

[0008] The cross-platform data acquisition module is used to acquire product data from multiple e-commerce platforms;

[0009] The AI ​​intelligent sourcing module matches the optimal commodity resources based on procurement needs;

[0010] The multi-dimensional price comparison analysis module is used for a comprehensive evaluation of product price, quality, and service;

[0011] The intelligent approval module completes the review of procurement applications based on a rules engine.

[0012] The procurement execution module is responsible for order placement and logistics tracking;

[0013] The settlement and reconciliation module enables monthly summaries and discrepancy processing;

[0014] The data review module provides in-depth analysis of procurement data.

[0015] Preferably, the central control module of the system includes a main control server, a distributed database, an API gateway, and a data encryption unit. The main control server adopts a multi-core processor to support parallel data processing; the distributed database is divided into a product database, an order database, a user database, and a historical purchase database, supporting massive data storage and fast retrieval; the API gateway uniformly manages the communication interfaces of each module, and the data encryption unit adopts the AES-256 encryption algorithm.

[0016] Each functional module establishes communication with the main control server through an API gateway. The main control server is responsible for scheduling data flow between modules and issuing instructions. The distributed database synchronizes the data generated by each module in real time, supports data backup and disaster recovery, and ensures data integrity.

[0017] Preferably, the cross-platform data acquisition module includes a platform adaptation interface unit, a data crawling unit, a data cleaning unit, and a data standardization unit. The platform adaptation interface unit has built-in dedicated interfaces for mainstream e-commerce platforms and supports the addition of new platform interfaces. The data crawling unit uses NLP-based web crawling technology to crawl product names, specifications, prices, supplier qualifications, sales volume, and review data in real time. The data cleaning unit uses AI algorithms to remove duplicate data and correct outliers. The data standardization unit converts heterogeneous data from different platforms into a unified format.

[0018] Preferably, the multi-dimensional price comparison analysis module includes a price analysis unit, a quality assessment unit, a service dimension unit, and a comprehensive scoring unit. The price analysis unit extracts real-time prices, historical price fluctuation data, and bulk purchase discount information for the product, and calculates the unit cost and total cost. The quality assessment unit combines product specifications, brand reputation, third-party testing and certification, and user review keyword extraction results to generate a quality score. The service dimension unit assesses the supplier's delivery timeliness, after-sales guarantee, return and exchange policy, and logistics coverage. The comprehensive scoring unit performs weighted calculations on price, quality, and service indicators based on preset weights and outputs a comprehensive score ranking.

[0019] Preferably, the intelligent approval module includes a rule engine unit, an approval process configuration unit, a compliance verification unit, and an approval result feedback unit. The rule engine unit has built-in standardized rules for purchase amount thresholds, supplier access standards, and budget compliance requirements, and supports enterprises to add custom approval rules. The approval process configuration unit can set multi-level approval nodes, including departmental approval, financial approval, and legal approval, and supports automatic matching of approval processes based on purchase amount and product type. The compliance verification unit compares the purchase application with the rule base to identify compliance risk points. The approval result feedback unit pushes the approval status to the applicant and approver in real time.

[0020] More preferably, the procurement execution module includes an order generation unit, a cross-platform order placement unit, a logistics tracking unit, and a goods receipt confirmation unit. The order generation unit automatically generates standardized purchase orders based on the approved price comparison results. The cross-platform order placement unit uses a platform adaptation interface unit to uniformly place orders across multiple platforms. The logistics tracking unit captures order logistics information in real time, including transportation status and estimated delivery time. The goods receipt confirmation unit allows users to upload receipt vouchers to confirm the goods receipt status.

[0021] The order generation unit establishes data interaction with the intelligent approval module and the distributed database. The cross-platform order placement unit reuses the communication link with the platform adaptation interface unit of the cross-platform data collection module. The logistics tracking unit obtains logistics data through the e-commerce platform interface. The receipt information is synchronized to the settlement and reconciliation module in real time.

[0022] Preferably, the settlement and reconciliation module includes an order amount summary unit, an invoice verification unit, a discrepancy processing unit, and a unified settlement unit. The order amount summary unit summarizes the payable amount for purchase orders on a periodic basis, including daily, weekly, and monthly. The invoice verification unit extracts invoice information using OCR recognition technology and automatically compares it with order data. The discrepancy processing unit identifies anomalies such as amount discrepancies and product discrepancies, generates a discrepancy list, and pushes it to the relevant responsible persons. The unified settlement unit supports integration with multiple payment methods.

[0023] Preferably, the data review module includes a data integration unit, an indicator analysis unit, a visualization unit, and a strategy optimization unit. The data integration unit extracts data from the entire procurement process, including sourcing, price comparison, orders, logistics, and settlement, forming a unified data mart. The indicator analysis unit calculates core indicators such as procurement cost reduction rate, sourcing efficiency, compliance rate, and supplier fulfillment rate, comparing year-on-year / month-on-month data. The visualization unit displays procurement category rankings, product rankings, and supplier rankings through charts, bar charts, line charts, or pie charts. The strategy optimization unit, based on machine learning algorithms, outputs suggestions for procurement budget allocation, supplier optimization, and procurement timing selection.

[0024] (III) Beneficial Effects

[0025] Compared with existing technologies, this invention provides an AI-based cross-platform intelligent procurement sourcing and price comparison system, which has the following beneficial effects:

[0026] This technical solution's cross-platform data acquisition module uses NLP web crawling technology to capture product information from mainstream e-commerce platforms in real time, and then uses AI algorithms for cleaning and standardization to ensure data accuracy and consistency. This module supports custom interface extensions, flexibly adapting to different e-commerce platforms and providing a comprehensive data foundation for subsequent analysis.

[0027] The AI-powered intelligent sourcing module matches optimal product resources based on procurement needs, utilizes deep learning models to predict supply and demand relationships, and optimizes procurement strategies. Compared to traditional manual screening methods, this module significantly improves sourcing efficiency and accuracy while reducing procurement costs.

[0028] The multi-dimensional price comparison analysis module combines price, quality, and service for comprehensive evaluation. It considers not only real-time price fluctuations but also incorporates quality indicators such as brand reputation, third-party certifications, and after-sales service guarantees. The comprehensive scoring unit uses preset weighted calculations to output supplier rankings, helping procurement personnel make quick decisions and improve procurement quality.

[0029] The intelligent approval module, based on a rules engine, automates approval process configuration, covering multiple stages including departmental approval, financial approval, and legal approval. The compliance verification unit identifies potential risks in real time, ensuring that every procurement request complies with internal policies and regulations, thus improving approval transparency and efficiency.

[0030] The procurement execution module enables end-to-end management of order generation, order placement, logistics tracking, and goods receipt confirmation. In particular, the cross-platform ordering unit unifies the processing of orders from multiple platforms, simplifying the operation process; the logistics tracking unit updates the transportation status in real time to ensure timely delivery of goods; and the goods receipt confirmation unit supports the uploading of vouchers to ensure the accuracy of receipt information.

[0031] The settlement and reconciliation module uses OCR technology to automatically extract invoice information and compare it with order data, quickly identifying and resolving discrepancies to avoid financial disputes. The unified settlement unit supports multiple payment methods, improving the efficiency of fund transfers.

[0032] The data review module helps companies quantify procurement performance and optimize procurement strategies through in-depth data mining and visualization of the entire procurement process. Machine learning algorithms recommend budget allocation, supplier selection, and procurement timing based on historical data, assisting companies in continuously improving their procurement management.

[0033] In summary, this system integrates intelligent management functions across the entire process from data collection to procurement execution and settlement reconciliation, significantly improving procurement efficiency, reducing operating costs, and enhancing supply chain flexibility. It is suitable for enterprises of all sizes and has broad market application prospects. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the system application process architecture of the present invention. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] Please see Figure 1 The present invention provides an AI-based cross-platform intelligent procurement sourcing and price comparison system, comprising a cross-platform data acquisition module, an AI intelligent sourcing module, a multi-dimensional price comparison analysis module, an intelligent approval module, a procurement execution module, a settlement and reconciliation module, a data review module, and a system central control module. Each module communicates bidirectionally with the system central control module through a distributed network.

[0037] The cross-platform data acquisition module is used to acquire product data from multiple e-commerce platforms;

[0038] The AI ​​intelligent sourcing module matches the optimal commodity resources based on procurement needs;

[0039] The multi-dimensional price comparison analysis module is used for a comprehensive evaluation of product price, quality, and service;

[0040] The intelligent approval module completes the review of procurement applications based on a rules engine.

[0041] The procurement execution module is responsible for order placement and logistics tracking;

[0042] The settlement and reconciliation module enables monthly summaries and discrepancy processing;

[0043] The data review module provides in-depth analysis of procurement data.

[0044] The working principle of this technical solution is as follows: the system's central control module serves as the core scheduling center; the cross-platform data acquisition module integrates multi-source commodity data; the AI ​​intelligent sourcing module accurately matches procurement needs; the multi-dimensional price comparison analysis module achieves comprehensive evaluation; the intelligent approval module ensures compliant procurement; the procurement execution module completes order and logistics management; the settlement and reconciliation module achieves financial closure; and the data review module provides decision support. All modules communicate bidirectionally through a distributed network, forming a fully intelligent management system encompassing demand input, sourcing and price comparison, approval execution, and settlement review, ensuring efficient, low-cost, and compliant procurement.

[0045] The system's central control module, acting as the system's brain, is responsible for issuing instructions to each module, scheduling data flow, ensuring data storage security and smooth communication, and forming the foundation of the entire closed-loop process.

[0046] The main control server uses a multi-core processor, such as the Intel Xeon 8375C, which supports multi-threaded parallel processing and can respond to requests from multiple modules simultaneously, ensuring data processing efficiency. The communication interfaces of each module are managed uniformly through an API gateway, and standardized interface protocols, such as RESTful, are used to avoid communication conflicts between modules.

[0047] The distributed database is partitioned by function for storage: the product database stores standardized product data collected across platforms, the order database records order information throughout its entire lifecycle, the user database stores data such as enterprise organizational structure and user permissions, and the historical procurement database archives past procurement records and price comparison results, supporting massive amounts of data, such as tens of millions of entries, for fast retrieval.

[0048] The data encryption unit uses the AES-256 encryption algorithm to perform double encryption on data transmission processes, such as between modules, between modules and e-commerce platforms, and on stored data. It also supports data backup and disaster recovery to ensure that data is not lost or leaked.

[0049] The cross-platform data acquisition module breaks down the data barriers of a single platform, enabling one-stop processing of product data collection, cleaning, and standardization from multiple e-commerce platforms, providing high-quality data support for subsequent sourcing and price comparison.

[0050] The platform's interface adaptation unit includes official API interfaces from mainstream platforms such as Taobao, JD.com, Pinduoduo, and Xianyu. It also provides custom interface development tools, allowing businesses to add new e-commerce platforms by configuring API parameters, resulting in strong adaptability. The interfaces employ dynamic adaptation technology, automatically adapting to platform interface version updates without requiring frequent modifications.

[0051] The data crawling unit is a web crawler based on NLP technology. It can parse unstructured data on the platform's product detail pages, such as product descriptions and user reviews, and accurately extract core fields such as product name, specifications, price, supplier qualifications, sales volume, and review keywords. It supports both scheduled crawling and triggered crawling modes. The scheduled crawling frequency can be set, such as once per hour, while triggered crawling starts immediately in response to procurement needs.

[0052] The data cleaning unit uses AI algorithms, such as K-means clustering and outlier detection, to remove duplicate data, correct price anomalies (e.g., erroneous data that is far below market price), and fill in missing parameters. The data standardization unit converts heterogeneous data from different platforms, such as "kilogram" and "kilogram weight," "specification" and "model," into a unified format, generating a standardized commodity data dictionary to ensure data consistency across modules.

[0053] The AI-powered intelligent sourcing module accurately analyzes procurement needs, matches optimal resources from massive amounts of standardized commodity data, and simultaneously filters high-quality suppliers, reducing interference from ineffective resources and improving sourcing efficiency.

[0054] The procurement requirement analysis unit uses NLP technology, such as the BERT model, to identify the natural language requirements input by the user, such as "procure 200 yellow V-shaped Grainger safety helmets". It extracts key information such as product name, brand, specifications, quantity, and quality requirements to generate structured requirement instructions, including keyword tags and parameter thresholds.

[0055] The intelligent matching unit uses machine learning algorithms, such as collaborative filtering and cosine similarity calculation, to semantically match structured demand instructions with standardized data in the product database and discover products that meet the specifications. It also supports fuzzy matching and can identify synonymous product names, such as "protective gloves" and "work gloves", to avoid missing any requirements.

[0056] The supplier screening unit accesses supplier information from the historical procurement database and product database, and assigns weighted scores based on dimensions such as qualifications (e.g., business license, testing and certification), historical cooperation records (e.g., fulfillment rate, after-sales feedback, user reviews (e.g., positive review rate, complaint rate), and fulfillment capabilities (e.g., delivery timeliness, inventory adequacy). Suppliers matching products are then ranked and sorted, such as A / B / C levels, with A-level suppliers being given priority.

[0057] The multi-dimensional price comparison analysis module breaks through the limitations of single-price comparison, comprehensively evaluates the cost-effectiveness of products from multiple dimensions such as price, quality, and service, and provides comprehensive data support for procurement decisions.

[0058] The price analysis unit extracts real-time prices, historical price fluctuation curves for the past 6 months, and bulk purchase discount rules, such as a 10% discount for purchasing 100 units, and calculates the unit cost and the total cost of the target purchase quantity. At the same time, it compares with historical purchase prices to generate the price advantage percentage, such as "4.29% lower than the lowest historical price".

[0059] The quality assessment unit integrates multi-dimensional data: comparing product specifications and parameters with industry standards, such as protection level, material compliance, and brand reputation scores; based on industry rankings and user perceptions, third-party testing and certification, such as SGS certification, and user review keyword extraction, such as the proportion of positive words like "wear-resistant" and "durable," to generate a quality score from 0 to 100.

[0060] The service dimension unit evaluates the supplier's service capabilities: delivery time, using average delivery time; after-sales guarantee, such as warranty period, repair policy, return and exchange policy, whether there is no-reason return and exchange, logistics coverage, whether it supports delivery to designated areas, and generates a service score according to weight.

[0061] The comprehensive rating unit is based on preset weights, with a default weighting of 40% for price, 35% for quality, and 25% for service. It supports customized adjustments by enterprises, calculates the weighted scores of the three components, outputs a comprehensive rating ranking, and generates a price comparison report, including a price comparison table and details of quality / service ratings.

[0062] The intelligent approval module is based on enterprise-customized rules to automate and standardize the approval of procurement applications, ensuring compliance of procurement activities while simplifying the approval process and improving approval efficiency.

[0063] The rules engine unit has built-in standardized rules, such as requiring financial and legal approval for purchases of ≥50,000 yuan, and requiring suppliers to have Level 3 or above qualifications. Enterprises can add custom rules through a visual interface, such as adding "hazardous chemicals require approval from the safety department". The rules are synchronized to the compliance verification unit in real time.

[0064] The approval process configuration unit can set up multi-level approval nodes, including departmental approval, financial approval, legal approval, and final approval. It supports approval processes based on purchase amount (e.g., ≤10,000 yuan requires only departmental approval, 10,000-50,000 yuan requires financial approval) and product type (e.g., electronic components require technical department approval). The system automatically matches the corresponding approval process without manual allocation.

[0065] The compliance verification unit calls the rule base of the rule engine unit to compare the purchase application, including the price comparison report, supplier information, and budget information, and automatically identifies compliance risk points, such as supplier qualifications not meeting the standards or purchases exceeding the budget, and marks the risk level, such as general / serious, with an explanation.

[0066] The approval result feedback unit pushes the approval status in real time through system terminals, such as PCs / mobile devices, such as pending approval / in progress / approved / rejected, to the applicant and approver. The approver can view the price comparison report and compliance verification results, and provide online approval comments; the reason for rejecting the application must be stated so that the applicant can modify and resubmit it.

[0067] The procurement execution module transforms approved procurement requirements into actual orders, enabling unified order placement across platforms, real-time logistics tracking, and goods receipt confirmation, ensuring a smooth and controllable procurement execution process.

[0068] The order generation unit receives the approval instruction from the intelligent approval module, calls the price comparison results, supplier information, and procurement demand data in the distributed database, and automatically generates a standardized purchase order, which includes the order number, product details, purchase quantity, price, payment method, and delivery address. The order format meets the company's financial reimbursement requirements.

[0069] The cross-platform order placement unit reuses the platform adaptation interface unit of the cross-platform data acquisition module, eliminating the need for additional interface development. It converts standardized orders into a platform-compatible format through the corresponding platform interface, enabling unified order placement across multiple platforms. It also supports batch ordering, allowing multiple orders to be placed on multiple platforms simultaneously, saving operation time.

[0070] The logistics tracking unit captures order logistics information in real time through the e-commerce platform interface, such as courier company, waybill number, transportation status, current location, and estimated delivery time. It updates the information to the order database at a preset frequency, such as once every 2 hours. When the transportation status changes, such as "shipped", "in transit" or "signed for", it automatically pushes reminders to the user.

[0071] The goods receipt confirmation unit allows users to upload receipt vouchers, such as photos / electronic documents, via mobile devices to manually confirm whether the quantity and quality of goods meet the standards and complete the goods receipt status confirmation; the confirmation information is synchronized to the settlement and reconciliation module in real time as the basis for reconciliation.

[0072] The settlement and reconciliation module enables the summarization of purchase order amounts, invoice verification, discrepancy handling, and unified settlement, simplifying the financial reconciliation process, reducing human error, and ensuring the accuracy of financial data.

[0073] The order amount summary unit automatically retrieves received order data from the order database according to the enterprise's set period, such as daily / weekly / monthly, summarizes the amount payable, the amount paid, and the amount unpaid, and generates a periodic reconciliation overview, such as a monthly reconciliation report. It supports summarizing by supplier and product type.

[0074] The invoice verification unit uses OCR recognition technology, such as the Baidu OCR interface, to extract key information from paper / electronic invoices, such as invoice number, invoice amount, product details, and invoice date, and automatically compares it with the corresponding order data in the order database to ensure consistency of the information.

[0075] The discrepancy processing unit automatically identifies anomalies after comparison: amount discrepancies, such as invoice amount not matching order amount; product discrepancies, such as invoice product details not matching order; and unmatched discrepancies, such as having an invoice but no order or having an order but no invoice. It generates a discrepancy list, including the discrepancy type and the relevant order / invoice information, and pushes it to the finance personnel for processing. It supports discrepancy annotation and processing record retention for easy traceability.

[0076] The unified settlement unit supports integration with multiple payment methods, such as corporate bank transfers, Alipay Enterprise Edition, and WeChat Pay Merchant Account. Finance personnel can initiate payments in batches based on the reconciliation overview and discrepancy processing results; the payment status is synchronized to the order database in real time, marking the order as "paid".

[0077] The data review module integrates data from the entire procurement process, conducts in-depth analysis and visualization, and outputs optimization suggestions to help companies optimize procurement strategies, reduce costs, and improve supply chain efficiency.

[0078] The data integration unit extracts full-process data from the distributed database, such as sourcing records, price comparison results, order data, logistics information, settlement data, and approval records. It then cleans and integrates the data according to time, product, supplier, and platform dimensions to form a unified data mart and eliminate data fragmentation.

[0079] The indicator analysis unit calculates core indicators based on the data mart: procurement cost reduction rate, compared with historical average cost, sourcing efficiency, such as sourcing time / traditional sourcing time, compliance rate, such as the number of compliant purchase orders / total number of orders, and supplier fulfillment rate, such as the number of on-time and high-quality delivery orders / total number of supplier orders. At the same time, it compares year-on-year / month-on-month data to identify data change trends.

[0080] The visualization unit uses charts such as bar charts, line charts, pie charts, and radar charts to intuitively display procurement data: procurement category ranking, based on the proportion of procurement amount for each category, product ranking, based on the top 10 single items in terms of procurement volume / amount, supplier ranking, based on the top 10 in terms of fulfillment rate / cost-effectiveness, price fluctuation trends, etc. It also supports filtering and viewing by time range, such as daily / weekly / monthly / quarterly.

[0081] The strategy optimization unit uses machine learning algorithms, such as the random forest model, to analyze historical procurement data and changes in core indicators, and outputs targeted optimization suggestions: procurement budget allocation, tilting towards high-cost-performance categories; supplier optimization, eliminating suppliers with low fulfillment rates, deepening cooperation with high-quality suppliers; and procurement timing selection, recommending low-price procurement periods based on price fluctuation curves. These suggestions can be directly imported into the procurement plan.

[0082] This technical solution has four major advantages in terms of feasibility: mature architecture, universal components, easy operation, and controllable performance.

[0083] The core technologies and components are all mature industrial-grade solutions: AI algorithms, such as NLP, machine learning, OCR recognition, distributed databases, and API gateways, are all commercially available technologies; there are publicly available solutions for e-commerce platform interfaces and payment interfaces, with a wide range of procurement channels and no technical barriers.

[0084] Simple module assembly and deployment: Each module communicates through a standardized API interface, supporting cloud server deployment, such as public cloud / private cloud deployment, allowing enterprises to flexibly choose the deployment method according to their scale; the system supports modular deployment, allowing the sourcing and price comparison modules to be deployed first, and then gradually expanding to approval, settlement, and review modules, reducing the implementation threshold.

[0085] The operation process is tailored to actual needs: the entire process from requirement input to data review is aligned with the actual procurement scenarios of enterprises. The user interface supports visual operation, such as rule configuration, weight adjustment, and chart viewing. Operators can become proficient in using it after simple training. It supports multi-device adaptation for PC / mobile terminals to meet the needs of office and mobile operation.

[0086] In summary, this technical solution has no obstacles in terms of technical architecture, component selection, implementation difficulty, and performance. It can be directly put into industrial application, is suitable for the procurement scenarios of large, medium and small enterprises, and has significant practical value and promotion prospects.

[0087] Detailed Workflow

[0088] System Deployment and Initialization Configuration

[0089] Module Deployment: Deploy each functional module on the cloud server, configure the main control server, distributed database, and API gateway of the system's central control module, and enable AES-256 data encryption; complete the communication connection between each module and the API gateway, and test the smoothness of data transmission.

[0090] Platform and rule configuration: Through the platform adaptation interface unit of the cross-platform data collection module, connect to target e-commerce platforms such as Taobao and JD.com, and configure custom platform addition channels; enter enterprise procurement rules, amount thresholds, supplier standards, approval nodes, and set initial weights for multi-dimensional price comparison in the intelligent approval module.

[0091] Data initialization: Import the company's historical procurement data, supplier information, and user permission data into the distributed database, complete the standardized configuration of the data dictionary, and unify product categories, specifications, and units.

[0092] Procurement requirement input and AI sourcing

[0093] Request entry: Users can enter their procurement requirements through PC / mobile system terminals, such as "purchase 500 pairs of 3M wear-resistant and non-slip work gloves", and can add requirements such as specifications, budget, and delivery time, and submit the request.

[0094] Demand Analysis and Data Acquisition: The system's central control module instructs the AI ​​intelligent sourcing module's procurement demand analysis unit to generate structured demand instructions through NLP technology; it then triggers the cross-platform data acquisition module to capture matching product data through various e-commerce platform interfaces, which is then cleaned, standardized, and stored in the product database.

[0095] Intelligent matching and supplier screening: The intelligent matching unit matches products that meet the requirements from the product database, and the supplier screening unit sorts suppliers by level and generates a candidate product list, which includes product information and supplier level.

[0096] Multi-dimensional price comparison and procurement application

[0097] Multi-dimensional analysis: The multi-dimensional price comparison analysis module extracts candidate product data from the price, quality, and service units respectively, calculates price advantage, quality score, and service score, and generates a ranking of the comprehensive score units according to their weights, thus generating a price comparison report.

[0098] Purchase request submission: Users view the price comparison report, select target products and suppliers, and the system automatically generates a purchase request form, linking the price comparison report, supplier information, and budget. After user confirmation, the request is submitted for approval.

[0099] Intelligent approval and order placement

[0100] Compliance verification and process matching: The compliance verification unit of the intelligent approval module compares the procurement application with the rule base to identify risk points; the approval process configuration unit automatically matches the corresponding approval node and pushes the application form to the approver.

[0101] Multi-level approval: Approver can view the application form and price comparison report through the system terminal and approve it online, such as approve / reject; if rejected, the reason must be stated, and the applicant must modify and resubmit; after all nodes have passed the approval, the approval result is fed back to the procurement execution module.

[0102] Cross-platform ordering: The order generation unit of the procurement execution module automatically generates standardized orders, and the cross-platform ordering unit places orders through the platform adaptation interface. The order information is stored in the order database.

[0103] Logistics tracking and receipt confirmation

[0104] Logistics monitoring: The logistics tracking unit captures order logistics information in real time and synchronizes it to the order database. Status changes are automatically pushed to users as notifications.

[0105] Receipt Confirmation: After receiving the goods, users can upload receipts via mobile devices to confirm that the quantity and quality of the goods meet the standards and complete the receipt confirmation; the confirmation information is synchronized to the settlement and reconciliation module.

[0106] Settlement reconciliation and unified payment

[0107] Amount Summary and Invoice Verification: The order amount summary unit of the settlement and reconciliation module summarizes the amount payable for received orders on a monthly basis; the invoice verification unit recognizes invoice information through OCR and automatically compares it with order data.

[0108] Discrepancy handling: The discrepancy handling unit generates a discrepancy list and pushes it to finance personnel for processing; finance personnel mark the processing results and update the reconciliation data.

[0109] Unified payment: Based on the reconciliation results, finance personnel initiate payments in batches through a unified settlement unit. The payment status is synchronized to the order database, and the order is marked as "paid".

[0110] Data review and strategy optimization

[0111] Data integration and indicator analysis: The data integration unit of the data review module extracts data from the entire process to form a data mart; the indicator analysis unit calculates core indicators and compares year-on-year / month-on-month data.

[0112] Visualization and Strategy Output: The visualization unit generates procurement data charts for enterprise management to view; the strategy optimization unit outputs budget allocation, supplier optimization, and procurement timing suggestions, which enterprises can then incorporate into their next procurement plan.

[0113] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their likenesses.

Claims

1. An AI-based cross-platform intelligent procurement sourcing and price comparison system, characterized in that, It includes a cross-platform data acquisition module, an AI intelligent sourcing module, a multi-dimensional price comparison analysis module, an intelligent approval module, a procurement execution module, a settlement and reconciliation module, a data review module, and a system central control module. Each module communicates bidirectionally with the system central control module through a distributed network. The cross-platform data acquisition module is used to acquire product data from multiple e-commerce platforms; The AI ​​intelligent sourcing module matches the optimal commodity resources based on procurement needs; The multi-dimensional price comparison analysis module is used for a comprehensive evaluation of product price, quality, and service; The intelligent approval module completes the review of procurement applications based on a rules engine. The procurement execution module is responsible for order placement and logistics tracking; The settlement and reconciliation module enables monthly summaries and discrepancy processing; The data review module provides in-depth analysis of procurement data.

2. The AI-based cross-platform intelligent procurement sourcing and price comparison system according to claim 1, characterized in that, The system's central control module includes a main control server, a distributed database, an API gateway, and a data encryption unit. The main control server uses a multi-core processor and supports parallel data processing. The distributed database is divided into a product database, an order database, a user database, and a historical purchase database, supporting massive data storage and fast retrieval. The API gateway manages the communication interfaces of each module in a unified manner, and the data encryption unit adopts the AES-256 encryption algorithm. Each functional module establishes communication with the main control server through an API gateway. The main control server is responsible for scheduling data flow between modules and issuing instructions. The distributed database synchronizes the data generated by each module in real time, supports data backup and disaster recovery, and ensures data integrity.

3. The AI-based cross-platform intelligent procurement sourcing and price comparison system according to claim 1, characterized in that, The cross-platform data acquisition module includes a platform adaptation interface unit, a data crawling unit, a data cleaning unit, and a data standardization unit. The platform adaptation interface unit has built-in dedicated interfaces for mainstream e-commerce platforms and supports the addition of new platform interfaces. The data crawling unit uses NLP-based web crawling technology to crawl product names, specifications, prices, supplier qualifications, sales volume, and review data in real time. The data cleaning unit uses AI algorithms to remove duplicate data and correct outliers. The data standardization unit converts heterogeneous data from different platforms into a unified format.

4. The AI-based cross-platform intelligent procurement sourcing and price comparison system according to claim 1, characterized in that, The multi-dimensional price comparison analysis module includes a price analysis unit, a quality assessment unit, a service dimension unit, and a comprehensive scoring unit. The price analysis unit extracts real-time prices, historical price fluctuation data, and bulk purchase discount information to calculate unit costs and total costs. The quality assessment unit combines product specifications, brand reputation, third-party testing and certification, and user review keyword extraction results to generate a quality score. The service dimension unit evaluates the supplier's delivery timeliness, after-sales guarantee, return and exchange policy, and logistics coverage; the comprehensive scoring unit calculates the price, quality, and service indicators based on preset weights and outputs a comprehensive score ranking.

5. The AI-based cross-platform intelligent procurement sourcing and price comparison system according to claim 1, characterized in that, The intelligent approval module includes a rule engine unit, an approval process configuration unit, a compliance verification unit, and an approval result feedback unit. The rule engine unit has built-in standardized rules for procurement amount thresholds, supplier access standards, and budget compliance requirements, and supports enterprises to add custom approval rules. The approval process configuration unit can set multi-level approval nodes, including departmental approval, financial approval, and legal approval, and supports automatic matching of approval processes based on procurement amount and product type. The compliance verification unit compares the procurement application with the rule base to identify compliance risk points. The approval result feedback unit pushes the approval status to the applicant and approver in real time.

6. The AI-based cross-platform intelligent procurement sourcing and price comparison system according to claim 1, characterized in that, The procurement execution module includes an order generation unit, a cross-platform order placement unit, a logistics tracking unit, and a goods receipt confirmation unit. The order generation unit automatically generates standardized procurement orders based on approved price comparison results. The cross-platform order placement unit uses a platform adaptation interface unit for unified order placement across multiple platforms. The logistics tracking unit captures order logistics information in real time, including transportation status and estimated delivery time. The goods receipt confirmation unit allows users to upload receipt vouchers to confirm the goods receipt status. The order generation unit establishes data interaction with the intelligent approval module and the distributed database. The cross-platform order placement unit reuses the communication link with the platform adaptation interface unit of the cross-platform data collection module. The logistics tracking unit obtains logistics data through the e-commerce platform interface. The receipt information is synchronized to the settlement and reconciliation module in real time.

7. The AI-based cross-platform intelligent procurement sourcing and price comparison system according to claim 1, characterized in that, The settlement and reconciliation module includes an order amount summary unit, an invoice verification unit, a discrepancy processing unit, and a unified settlement unit. The order amount summary unit summarizes the payable amount for purchase orders on a periodic basis, including daily, weekly, and monthly. The invoice verification unit extracts invoice information using OCR recognition technology and automatically compares it with order data. The discrepancy processing unit identifies anomalies such as amount discrepancies and product discrepancies, generates a discrepancy list, and pushes it to the relevant responsible persons. The unified settlement unit supports integration with multiple payment methods.

8. The AI-based cross-platform intelligent procurement sourcing and price comparison system according to claim 1, characterized in that, The data review module includes a data integration unit, an indicator analysis unit, a visualization unit, and a strategy optimization unit. The data integration unit extracts data from the entire procurement process, including sourcing, price comparison, orders, logistics, and settlement, forming a unified data mart. The indicator analysis unit calculates core indicators such as procurement cost reduction rate, sourcing efficiency, compliance rate, and supplier fulfillment rate, comparing year-on-year and month-on-month data. The visualization unit displays procurement category rankings, product rankings, and supplier rankings through charts, bar charts, line charts, or pie charts. The strategy optimization unit is based on machine learning algorithms and outputs suggestions on procurement budget allocation, supplier optimization, and procurement timing selection.

Images