An AI-driven intelligent price inquiry decision-making method and system

Abstract

This invention provides an AI-driven intelligent price inquiry and decision-making method and system. The method includes: outputting the target product specifications and the target price inquiry platform to a mode decision module; the mode decision module generating a linkage control strategy to guide subsequent data collection and decision-making based on multiple pre-configured control modes and linkage rules; sending the linkage control strategy to a data aggregation module; the data aggregation module driving an automated execution component to acquire product data from each of the target price inquiry platforms according to the strategy; and synchronously inputting the acquired product data and the linkage control strategy into an intelligent decision engine, which determines the results based on the parameter weights in the linkage control strategy. According to this invention, by transforming vague user needs into specific scoring weights and execution priority parameters, subsequent data collection and decision-making are guided, making the results more closely match the user's true expectations.

Description

Technical Field

[0001] This invention relates to the field of computer data processing technology, and specifically to an AI-driven intelligent price inquiry and decision-making method and system. Background Technology

[0002] With the rapid development of e-commerce and the deepening of digital transformation in enterprise procurement, how to efficiently and accurately obtain the latest price and inventory information of components, spare parts and other commodities on numerous online trading platforms has become a core requirement in business processes such as distribution box pricing, supply chain management and cost control. Current solutions mainly consist of centralized data warehouse queries and multi-platform aggregated queries based on RPA (Robotic Process Automation).

[0003] The former uses data synchronization or ETL tools to periodically collect and store product information from multiple target platforms in a local or central database, but it cannot reflect the real-time price and dynamic inventory status of products. RPA, on the other hand, uses automated scripts to control the browser to simulate human operation, sequentially accessing the websites of each target platform, inputting search keywords, parsing the page structure and extracting product data, and finally summarizing and presenting the data locally. This solution solves the real-time problem to some extent.

[0004] However, its maintenance costs are high and its adaptability is poor. For example, once the webpage structure of each target platform is updated, all crawling scripts based on element positioning (such as XPath and CSS selectors) need to be manually re-debugged and deployed, resulting in heavy maintenance work. At the same time, its decision support is weak. Existing RPA solutions are essentially fixed process automation, only able to complete the basic functions of data collection and display, unable to understand complex user inquiry intentions, and unable to perform in-depth analysis and intelligent sorting of product information from different platforms. Therefore, an AI-driven intelligent inquiry decision-making method and system are invented to improve the above problems. Summary of the Invention

[0005] Therefore, the technical problem to be solved by the present invention is to overcome the above-mentioned defects, thereby providing an AI-driven intelligent inquiry decision-making method and system.

[0006] To address the aforementioned problems, this invention provides an AI-driven intelligent price inquiry decision-making method, comprising the following steps: S1: Receive and parse user inquiry requests through the request parsing module of the intelligent inquiry system, and extract the target product specifications and at least one target inquiry platform; S2: Output the target product specifications and the target inquiry platform to the mode decision module. The mode decision module generates a linkage control strategy to guide subsequent data collection and decision-making based on multiple pre-configured control modes and linkage rules. S3: The linkage control strategy is sent to the data aggregation module. The data aggregation module drives the automated execution component according to the strategy to obtain product data from each of the target inquiry platforms. The product data includes basic price, promotional information, logistics timeliness, activity time and product sales. S4: The acquired product data and the linkage control strategy are synchronously input into the intelligent decision engine. The intelligent decision engine calls the preset multi-dimensional comprehensive scoring model to calculate the product data according to the parameter weights in the linkage control strategy, and generates a comprehensive score and recommendation level for each product option. S5: Output the comprehensive score and recommendation level to the result output module, which then generates and outputs the optimized inquiry result list accordingly.

[0007] Preferably, in step S2, the mode decision module generates the linkage control strategy as follows: The mode decision module is pre-set with three basic control modes: time-priority mode, cost-priority mode, and comprehensive balance mode, and maintains a linkage rule library, which stores rules for combining or switching the basic control modes. The mode decision module receives user intent information from the request parsing module, selects at least one from the basic control modes based on the user intent information, processes the selected basic control modes according to the rules in the linkage rule base, and generates a linkage control strategy containing specific scoring weights and execution priority parameters. The linkage control strategy is simultaneously sent to the data aggregation module and the intelligent decision engine.

[0008] Preferably, in step S4, before inputting the product data into the intelligent decision engine, the method further includes: The sales accumulation step is executed through a data enhancement module; Once the data aggregation module completes the collection and standardization of product data from all target inquiry platforms, the data enhancement module is triggered to start. The input of the data enhancement module is connected to the standardized data output of the data aggregation module to receive all standardized commodity data; The data enhancement module, based on the core specifications, brand and model information of the product, performs entity matching and summarization on product entries from different platforms, and identifies multiple product options pointing to the same physical product. For the same physical product from the same store on different platforms, the data enhancement module adds up the product's sales volume on different platforms to generate cross-platform cumulative sales data for the product. The output of the data enhancement module is connected to the extended data input of the intelligent decision engine, which is used to integrate the cross-platform cumulative sales data into the cumulative sales weight dimension and supplement the product data for multi-dimensional comprehensive scoring model to call and calculate.

[0009] Preferably, the intelligent decision engine uses a multi-dimensional comprehensive scoring model to perform the calculation as follows: In S4, the multi-dimensional comprehensive scoring model includes parallel price calculation unit, timeliness calculation unit, reputation calculation unit, promotion calculation unit, sales accumulation calculation unit, and time matching calculation unit; The time matching calculation unit receives promotional activity time information and logistics timeliness information from the data aggregation module. When it is identified that the promotional activity is in the preheating state, it estimates the arrival time according to the formula: arrival time = promotion start time + logistics estimated timeliness. The calculation result is compared with the user's demand time window, and a time matching score is output. The output scores of each calculation unit are weighted and summed according to the weight parameters in the linkage control strategy to obtain the comprehensive score.

[0010] Preferably, the result output module generates a list of inquiry results as follows: The result output module receives the comprehensive score, recommendation level, and time matching information generated by the time matching calculation unit for each product option from the intelligent decision engine. For product options associated with pre-sale promotions indicated by the time matching information, the result output module calls a comparison scheme generation unit to generate comparison data containing immediate execution schemes and delayed execution schemes, and integrates the comparison data into the display information of the product options to jointly form the inquiry result list.

[0011] Preferably, the data aggregation module drives the automated execution component to acquire product data as follows: The data aggregation module is connected to the automated execution component through a control command interface, and sends control commands containing the target platform access address and the simulated operation sequence to the automated execution component according to the linkage control strategy. The automated execution component has a built-in AI analysis unit, which is used to analyze the page structure after loading the target platform page, locate the search box, login form and product list elements, and simulate user operations to complete data collection. The data aggregation module also includes a data standardization unit, which is connected to the data output end of the automated execution component. The data standardization unit is used to convert the collected raw heterogeneous data into standardized data records with a unified field structure and send them to the intelligent decision engine.

[0012] Preferably, it further includes an anomaly detection step: The anomaly detection module performs the anomaly detection step, with its input connected to the output of the data aggregation module and the output of the intelligent decision engine, in order to obtain the product data and preliminary scoring results. The anomaly detection module is equipped with a causal knowledge base. When anomalies in price, inventory, or logistics data are detected, the causal knowledge base is called to perform root cause matching and inference, and an anomaly report and handling suggestions are generated. The output of the anomaly detection module is connected to the result output module, and is used to send the anomaly report and processing suggestions to the result output module, which then marks them as additional information in the corresponding item of the inquiry result list.

[0013] Preferably, the method further includes a cache management step: The cache management steps are executed through a cache management module, which is communicatively connected to the request parsing module, the intelligent decision engine, and the result output module. The cache management module associates and stores the key parameters of this inquiry task, the product data, and the final generated list of inquiry results in a distributed cache database, and sets a sliding expiration time for the cache item. When the request parsing module receives a new user inquiry request, the cache management module performs a similarity matching query in the distributed cache database. If a match is found and the cache has not expired, the cache result is directly returned through the result output module, and the sliding expiration time of the cache item is reset.

[0014] This invention also provides an AI-driven intelligent price inquiry system, characterized in that it includes the AI-driven intelligent price inquiry decision-making method described in any of the preceding claims, comprising: The request parsing module is used to receive and parse user inquiry requests, and extract the target product specifications and target inquiry platform; The mode decision module, whose input is connected to the output of the request parsing module, is used to generate a linkage control strategy based on multiple pre-configured control modes and linkage rules. The data aggregation module, whose control input is connected to the output of the mode decision module, is used to drive the automated execution component to collect commodity data based on the linkage control strategy. The intelligent decision engine has its data input end connected to the output end of the data aggregation module and its strategy input end connected to another output end of the pattern decision module, and is used to perform multi-dimensional comprehensive scoring of commodity data. The result output module, whose input is connected to the output of the intelligent decision engine, is used to generate and output an optimized list of inquiry results. The control mode management module has a configuration interface for external configuration of parameters for the time-priority mode, cost-priority mode, and comprehensive balance mode. Its output is connected to the rule update input of the mode decision module for injecting or updating rules into the linkage rule base of the mode decision module. An anomaly detection module, one input of which is connected to the data output of the data aggregation module, and the other input is connected to the intermediate result output of the intelligent decision engine, and its output is connected to the additional information input of the result output module; The data enhancement module has its trigger input connected to the task completion signal output of the data aggregation module, its data input connected to the standardized data output of the data aggregation module, and its data output connected to the extended data input of the intelligent decision engine. The cache management module has a first interface connected to the query interface of the request parsing module, a second interface connected to the result output interface of the intelligent decision engine, and a third interface connected to the cache read / write interface of the result output module.

[0015] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the AI-driven intelligent price inquiry decision method as described in any of the preceding claims.

[0016] The AI-driven intelligent price inquiry decision-making method and system provided by this invention have the following beneficial effects: 1. This invention analyzes user intent and generates linkage control strategies through a pattern decision module, which can transform vague user needs into specific scoring weights and execution priority parameters, thereby guiding subsequent data collection and decision-making, making the results more in line with the user's true expectations, and realizing the understanding and execution of the user's complex inquiry intent; 2. This invention also uses a pre-set multi-dimensional comprehensive scoring model combined with the weights of linkage strategies. The multi-dimensional comprehensive scoring model includes multiple calculation units such as price, timeliness, reputation, promotion, market popularity, and time matching. It can perform unified and interpretable comprehensive quantitative scoring and ranking of all candidate products, getting rid of simple result listing, helping users quickly locate the optimal option, and providing in-depth and quantifiable intelligent decision support. 3. The present invention also calculates the estimated time of receipt through a time matching calculation unit, and generates intuitive comparison data between immediate execution plan and delayed execution plan through a result output module, so that users can clearly assess the actual value and time cost of waiting for future offers, further improving comprehensive and forward-looking decision-making, and realizing intelligent processing and clear presentation of future promotional activities. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the system structure of the present invention. Detailed Implementation

[0018] like Figure 1 As shown, this invention provides an AI-driven intelligent price inquiry decision-making method, aiming to solve the shortcomings of existing technologies such as difficult RPA script maintenance, inability to understand complex user intentions, and weak decision support. It includes the following steps: S1: Receive and parse user inquiry requests through the request parsing module of the intelligent inquiry system, and extract the target product specifications and at least one target inquiry platform; S2: Output the target product specifications and the target inquiry platform to the mode decision module. The mode decision module generates a linkage control strategy to guide subsequent data collection and decision-making based on multiple pre-configured control modes and linkage rules. S3: The linkage control strategy is sent to the data aggregation module. The data aggregation module drives the automated execution component according to the strategy to obtain product data from each of the target inquiry platforms. The product data includes basic price, promotional information, logistics timeliness, activity time and product sales. S4: The acquired product data and the linkage control strategy are synchronously input into the intelligent decision engine. The intelligent decision engine calls the preset multi-dimensional comprehensive scoring model to calculate the product data according to the parameter weights in the linkage control strategy, and generates a comprehensive score and recommendation level for each product option. S5: Output the comprehensive score and recommendation level to the result output module, which then generates and outputs the optimized inquiry result list accordingly.

[0019] Specifically, user requests input through the front-end interface are typically in natural language or structured forms, such as web pages or API interfaces. The request parsing module first performs semantic analysis on the input, identifying and extracting key entities. For example, for a request to find a circuit breaker of a certain brand and model, comparing prices on platforms A and B, and preferably with delivery within three days, the module will extract: the target product specifications are {brand: [Brand Name], model: [Model Name]}, and the target inquiry platforms are [Platform A, Platform B]. In addition, the module parses the implicit user intent, such as the three-day delivery requirement indicating a timeliness requirement, and converts it into structured information to be passed to subsequent modules.

[0020] In some embodiments, in S2, the mode decision module generates a linkage control strategy as follows: There are three basic control modes, namely, the time - efficiency - priority mode, the cost - priority mode, and the comprehensive - balance mode, preset in the mode decision module, and a linkage rule library is maintained. Rules for combining or switching the basic control modes are stored in the linkage rule library. The mode decision module receives user intention information from the request parsing module, selects at least one from the basic control modes based on the user intention information, processes the selected basic control modes according to the rules in the linkage rule library, generates a linkage control strategy including specific scoring weights and execution priority parameters, and the linkage control strategy is simultaneously sent to the data aggregation module and the intelligent decision engine.

[0021] Specifically, at least three basic control modes are preset in the mode decision module, and each mode is associated with a set of preset scoring weights and execution preferences. Among them, for the cost - priority mode, the weight configuration inclines towards the price dimension and the promotion intensity dimension, and when collecting data, it may crawl coupon information more deeply. For the time - efficiency - priority mode, its weight configuration inclines towards the logistics time - efficiency dimension, and when executing, it preferentially queries products marked with same - day delivery or next - day delivery. The weight of the comprehensive - balance mode is relatively balanced among various dimensions, and other dimensions such as reputation and sales volume are introduced as references. The mode decision module maintains a linkage rule library. When receiving user intention information, the mode decision does not mechanically select a single mode. For example, for the intention of the lowest cost but not exceeding five days at the latest, the mode decision starts a linkage control strategy mainly based on the cost - priority mode and supplemented by the time - efficiency - priority mode according to the rules. Each strategy contains a parameter set of specific instructions, such as: {primary mode: cost - priority, secondary mode: time - efficiency - priority, weight combination: {price: 0.4, time - efficiency: 0.3, promotion: 0.15,...}, execution priority: [self - operated store on Platform A, brand store on Platform B]}, and the corresponding strategy is simultaneously sent to the data aggregation module and the intelligent decision engine to coordinate their behaviors.

[0022] In some embodiments, in S3, the data aggregation module drives the automated execution component to obtain product data as follows: The data aggregation module is connected to the automated execution component through a control instruction interface, and sends a control instruction including the target platform access address and the simulated operation sequence to the automated execution component according to the linkage control strategy. The automated execution component is built - in with an AI analysis unit, which is used to analyze the page structure after loading the target platform page, locate elements such as the search box, login form, and product list, and simulate user operations to complete data collection. The data aggregation module also includes a data standardization unit, which is connected to the data output end of the automated execution component, and is used to convert the collected original heterogeneous data into standardized data records with a unified field structure and send them to the intelligent decision engine.

[0023] Specifically, the data aggregation module performs cross-platform data collection. It does not rely on fixed webpage element positioning scripts. Instead, it executes the process through command issuance, AI dynamic interaction, and data extraction and standardization. When a command is issued, the data aggregation module sends the target platform access address and search keywords generated based on product specifications to the automated execution component via the control command interface. During AI dynamic interaction, the automated execution component has a built-in AI analysis unit that controls the browser to open the target platform. After the page loads, it sends the complete page DOM (Document Object Model) information to a lightweight visual language model for analysis. The visual language model dynamically identifies elements on the page, such as search input boxes, search buttons, login boxes, and product list items, and generates command sequences that simulate human operations, such as clicking, inputting, and scrolling. This method can effectively adapt to partial page redesigns. Subsequently, the raw data collected during the data extraction and standardization process is heterogeneous, with different page structures on each platform. The data standardization unit within the data aggregation module converts the raw data into standardized data records with unified fields according to a predefined template. This results in the final output product data being a standardized list containing fields such as basic price, promotional information, logistics timeliness, activity time, and product sales volume.

[0024] In some implementations, before inputting the product data into the intelligent decision engine in step S4, the method further includes: executing the sales accumulation step through a data enhancement module; triggering the data enhancement module to start after the data aggregation module completes the collection and standardization of product data from all target inquiry platforms; the input end of the data enhancement module is connected to the standardized data output end of the data aggregation module to receive all standardized product data; the data enhancement module performs entity matching and summarization on product items from different platforms based on the core specification parameters, brand, and model information of the product, and identifies multiple product options pointing to the same physical product; for the same physical product from the same store on different platforms, the data enhancement module accumulates its product sales on different platforms to generate cross-platform cumulative sales data for the product; the output end of the data enhancement module is connected to the extended data input end of the intelligent decision engine to integrate the cross-platform cumulative sales data into the sales accumulation weight dimension and supplement it to the product data for multi-dimensional comprehensive scoring model to call and calculate.

[0025] Specifically, before inputting the product data into the intelligent decision engine, the method also includes a sales accumulation step: executed by a data enhancement module, whose input is a standardized product data list output by a data aggregation module; the data enhancement module executes a core entity matching algorithm to identify whether entries on different platforms point to the same physical product; it comprehensively compares the product's model, brand, and key specifications, and can also use the product's main image for auxiliary comparison; for the same identified product, for example, a certain MCCB-16A / 1P circuit breaker is sold in store A on platform A, store B on platform A, and store C on platform B, the data enhancement module accumulates their sales on all platforms to generate cross-platform cumulative sales, which is used as a market popularity indicator and attached to the standardized data record of the product, and then sent to the intelligent decision engine.

[0026] In some implementations, the intelligent decision engine calls a multi-dimensional comprehensive scoring model for calculation as follows: In S4, the multi-dimensional comprehensive scoring model includes a parallel price calculation unit, a timeliness calculation unit, a reputation calculation unit, a promotion calculation unit, a sales accumulation calculation unit, and a time matching calculation unit; the time matching calculation unit receives promotional activity time information and logistics timeliness information from the data aggregation module, and when it identifies that the promotional activity is in a preheating state, it estimates the arrival time according to the formula, compares the calculation result with the user's demand time window, and outputs a time matching score; the output scores of each calculation unit are weighted and summed according to the weight parameters in the linkage control strategy to obtain the comprehensive score.

[0027] Specifically, the intelligent decision engine comprises various units forming a rule engine. Its price calculation unit calculates the final price of the product (original price - promotional discount) and performs a normalized score based on the lowest price among all candidate products. Its timeliness calculation unit scores based on the promised delivery time; the shorter the delivery time, the higher the score. Its reputation calculation unit calculates a reputation score based on store ratings and platform certification marks, such as brand flagship stores. Its promotion calculation unit evaluates the discount strength; the higher the discount rate and the more direct the discount, the higher the score. Its cumulative sales calculation unit utilizes cross-platform cumulative sales data provided by the data augmentation module; the higher the sales volume, the higher the market popularity score. The time matching calculation unit processes future effective discount conditions. When a product is identified as having a related pre-sale activity, the time matching calculation performs a key calculation: estimated delivery time = promotion start time + estimated logistics timeliness. This time is then compared with the user's demand time window. If the time window is not specified, the system default is used, and a time matching score is output. If the estimated delivery time is earlier than the demand deadline, the matching degree is high; if it is later but not significantly different, the matching degree is medium; if it is much later, the matching degree is low.

[0028] Specifically, each computing unit outputs a 0-1 dimension score. The intelligent decision engine performs a weighted sum of the scores of each dimension according to the specific weight parameters issued by the linkage control strategy. Finally, all product options are sorted according to the comprehensive score and a recommendation level is assigned, such as: >0.8 is strongly recommended, and 0.6-0.8 is recommended.

[0029] In some implementations, the result output module generates the price inquiry result list as follows: the result output module receives the comprehensive score, recommendation level, and time matching information generated by the time matching calculation unit for each product option from the intelligent decision engine; for product options whose time matching information indicates association with pre-sale promotion activities, the result output module calls a comparison scheme generation unit to generate comparison data containing immediate execution schemes and delayed execution schemes, and integrates the comparison data into the display information of the product options to jointly constitute the price inquiry result list.

[0030] Specifically, the results output module is responsible for generating reports visible to the end user. For products identified by the time matching calculation unit as being associated with pre-sale promotions, the results output module calls the internal comparison scheme generation unit to generate dual-scheme comparison data. The immediate execution scheme allows the user to ignore future discounts and purchase immediately under the current conditions, displaying the current real-time price and estimated delivery time. The delayed execution scheme allows the user to wait for future discounts to take effect before purchasing, clearly displaying the promotion name, start time, estimated price after discount, and estimated delivery time. At the same time, the system automatically calculates and prompts the benefits of waiting, such as waiting for 3 days, which is expected to save 15 yuan, a 20% saving. Through parallel comparison, users can intuitively weigh the time cost against the economic benefits, achieving in-depth decision support. Finally, all product options, along with their ratings, recommendation levels, and possible dual-scheme comparison data, are organized into a well-structured and information-rich inquiry results list, such as in JSON format or a visual webpage, and returned to the user.

[0031] In some implementations, an anomaly detection step is also included: the anomaly detection module executes the anomaly detection step, the input of the anomaly detection module is connected to the output of the data aggregation module and the output of the intelligent decision engine to obtain the product data and preliminary scoring results; the anomaly detection module is equipped with a causal knowledge base, and when anomalies in price, inventory or logistics data are detected, the causal knowledge base is called to perform root cause matching and inference, and an anomaly report and handling suggestions are generated; the output of the anomaly detection module is connected to the result output module, which sends the anomaly report and handling suggestions to the result output module, and the result output module marks them as additional information in the corresponding item of the inquiry result list.

[0032] Specifically, the anomaly detection step is executed by the anomaly detection module. The anomaly detection module monitors the data flow, and when it finds that the price of a product deviates from the historical range or the market average price, the inventory is falsely high, or the logistics timeliness is abnormal, it will query its built-in causal knowledge base. The causal knowledge base stores domain rules, for example: the possible reasons for a new store with a price 30% or more below the historical average are sales-boosting marketing or data anomalies; it is recommended to carefully evaluate the store's reputation, and the generated anomaly report will be attached as a label to the results list.

[0033] In some embodiments, the method further includes a cache management step: the cache management step is executed by a cache management module, which is communicatively connected to the request parsing module, the intelligent decision engine, and the result output module; the cache management module associates and stores the key parameters of this inquiry task, the product data, and the final generated inquiry result list in a distributed cache database, and sets a sliding expiration time for the cache item; when the request parsing module receives a new user inquiry request, the cache management module performs a similarity matching query in the distributed cache database. If a match is found and the cache has not expired, the cache result is directly returned through the result output module, and the sliding expiration time of the cache item is reset.

[0034] Specifically, the cache management module stores each user request and the final list of optimized results in a distributed cache database, such as Redi, and sets a sliding expiration time, such as 30 minutes. When the request parsing module receives a new request, the cache management module first performs a similarity matching query. If a match is found and the cached result has not expired, the cached result is returned directly, which greatly improves the response speed of high-frequency repeated queries and resets the expiration time of the cached item.

[0035] This invention uses AI to dynamically analyze page structure and simulate interactions. Compared with traditional scripts based on fixed XPath / CSS selectors, this allows the system to automatically adapt when the page structure of the target platform undergoes regular updates, without the need for manual rewriting and deployment of crawling scripts, thus reducing the maintenance cost of automated scripts.

[0036] This invention analyzes user intent and generates linkage control strategies through a pattern decision module, which can transform vague user needs into specific scoring weights and execution priority parameters, thereby guiding subsequent data collection and decision-making, making the results more in line with the user's true expectations, and realizing the understanding and execution of the user's complex inquiry intent.

[0037] This invention utilizes a pre-defined multi-dimensional comprehensive scoring model combined with the weights of a linkage strategy. The multi-dimensional comprehensive scoring model includes multiple calculation units such as price, timeliness, reputation, promotion, market popularity, and time matching. It can perform unified and interpretable comprehensive quantitative scoring and ranking of all candidate products, moving beyond simple result listing and helping users quickly locate the optimal option, providing in-depth and quantifiable intelligent decision support.

[0038] This invention calculates the estimated arrival time through a time matching calculation unit and generates intuitive comparison data between immediate and delayed execution plans through a result output module. This allows users to clearly assess the actual value and time cost of waiting for future offers, further enhancing comprehensive and forward-looking decision-making and enabling intelligent processing and clear presentation of future promotional activities.

[0039] This invention uses a data augmentation module to perform cross-platform product entity matching and sales accumulation. The resulting cross-platform cumulative sales data more accurately reflects the market acceptance of products than sales on a single platform, providing a more reliable market popularity input dimension for the scoring model and accurately assessing the market popularity of products.

[0040] like Figure 2 As shown, this invention also provides an AI-driven intelligent price inquiry system, including the AI-driven intelligent price inquiry decision-making method described in the preceding claim. The AI-driven intelligent price inquiry system includes a request parsing module, a pattern decision-making module, a data aggregation module, an intelligent decision engine, a result output module, a control pattern management module, an anomaly detection module, a data enhancement module, and a cache management module. The modules interact with each other through interfaces to exchange data and instructions, collaboratively completing the entire process from request parsing to result output, including: The request parsing module is used to receive and parse user inquiry requests, and extract the target product specifications and target inquiry platform; The mode decision module, whose input is connected to the output of the request parsing module, is used to generate a linkage control strategy based on multiple pre-configured control modes and linkage rules. The data aggregation module, whose control input is connected to the output of the mode decision module, is used to drive the automated execution component to collect commodity data based on the linkage control strategy. The intelligent decision engine has its data input end connected to the output end of the data aggregation module and its strategy input end connected to another output end of the pattern decision module, and is used to perform multi-dimensional comprehensive scoring of commodity data. The result output module, whose input is connected to the output of the intelligent decision engine, is used to generate and output an optimized list of inquiry results. The control mode management module has a configuration interface for external configuration of parameters for the time-priority mode, cost-priority mode, and comprehensive balance mode. Its output is connected to the rule update input of the mode decision module for injecting or updating rules into the linkage rule base of the mode decision module. An anomaly detection module, one input of which is connected to the data output of the data aggregation module, and the other input is connected to the intermediate result output of the intelligent decision engine, and its output is connected to the additional information input of the result output module; The data enhancement module has its trigger input connected to the task completion signal output of the data aggregation module, its data input connected to the standardized data output of the data aggregation module, and its data output connected to the extended data input of the intelligent decision engine. The cache management module has a first interface connected to the query interface of the request parsing module, a second interface connected to the result output interface of the intelligent decision engine, and a third interface connected to the cache read / write interface of the result output module.

[0041] Each functional module in this invention has a clearly defined responsibility and communicates through interfaces to achieve collaborative and efficient processing of data and decision flows. The control mode management module allows administrators or users to easily configure, adjust, or combine different basic control modes and their linkage rules according to business scenarios, enabling the system to quickly adapt to diverse inquiry strategy requirements. The anomaly detection module can monitor data anomalies in real time and provide possible root cause inferences and processing suggestions using a built-in causal knowledge base, feeding back to users as additional information to improve the credibility of the results and make the system reliable. The cache management module caches the results of high-frequency or repeated queries and sets a sliding expiration time, which can reduce repeated calls to underlying data aggregation and intelligent computing resources while ensuring a certain level of data timeliness, thereby speeding up the response speed to user requests.

[0042] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the AI-driven intelligent price inquiry decision method as described in any of the preceding claims.

[0043] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention. The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the protection scope of the present invention.

Claims

1. An AI-driven intelligent price inquiry decision-making method, characterized in that, Includes the following steps: S1: Receive and parse user inquiry requests through the request parsing module of the intelligent inquiry system, and extract the target product specifications and at least one target inquiry platform; S2: Output the target product specifications and the target inquiry platform to the mode decision module. The mode decision module generates a linkage control strategy to guide subsequent data collection and decision-making based on multiple pre-configured control modes and linkage rules. S3: The linkage control strategy is sent to the data aggregation module. The data aggregation module drives the automated execution component according to the strategy to obtain product data from each of the target inquiry platforms. The product data includes basic price, promotional information, logistics timeliness, activity time and product sales. S4: The acquired product data and the linkage control strategy are synchronously input into the intelligent decision engine. The intelligent decision engine calls the preset multi-dimensional comprehensive scoring model to calculate the product data according to the parameter weights in the linkage control strategy, and generates a comprehensive score and recommendation level for each product option. S5: Output the comprehensive score and recommendation level to the result output module, which then generates and outputs the optimized inquiry result list accordingly.

2. The AI-driven intelligent price inquiry decision-making method according to claim 1, characterized in that: In S2, the mode decision module generates the following linkage control strategy: The mode decision module is pre-set with three basic control modes: time-priority mode, cost-priority mode, and comprehensive balance mode, and maintains a linkage rule library, which stores rules for combining or switching the basic control modes. The mode decision module receives user intent information from the request parsing module, selects at least one from the basic control modes based on the user intent information, processes the selected basic control modes according to the rules in the linkage rule base, and generates a linkage control strategy containing specific scoring weights and execution priority parameters. The linkage control strategy is simultaneously sent to the data aggregation module and the intelligent decision engine.

3. The AI-driven intelligent price inquiry decision-making method according to claim 2, characterized in that: Before inputting the product data into the intelligent decision engine, step S4 further includes: The sales accumulation step is executed through a data enhancement module; Once the data aggregation module completes the collection and standardization of product data from all target inquiry platforms, the data enhancement module is triggered to start. The input of the data enhancement module is connected to the standardized data output of the data aggregation module to receive all standardized commodity data; The data enhancement module, based on the core specifications, brand and model information of the product, performs entity matching and summarization on product entries from different platforms, and identifies multiple product options pointing to the same physical product. For the same physical product from the same store on different platforms, the data enhancement module adds up the product's sales volume on different platforms to generate cross-platform cumulative sales data for the product. The output of the data enhancement module is connected to the extended data input of the intelligent decision engine, which is used to integrate the cross-platform cumulative sales data into the cumulative sales weight dimension and supplement the product data for multi-dimensional comprehensive scoring model to call and calculate.

4. The AI-driven intelligent price inquiry decision-making method according to claim 3, characterized in that: The intelligent decision engine uses a multi-dimensional comprehensive scoring model to calculate: In S4, the multi-dimensional comprehensive scoring model includes parallel price calculation unit, timeliness calculation unit, reputation calculation unit, promotion calculation unit, sales accumulation calculation unit, and time matching calculation unit; The time matching calculation unit receives promotional activity time information and logistics timeliness information from the data aggregation module. When it is identified that the promotional activity is in the preheating state, it estimates the arrival time according to the formula: arrival time = promotion start time + logistics estimated timeliness. The calculation result is compared with the user's demand time window, and a time matching score is output. The output scores of each calculation unit are weighted and summed according to the weight parameters in the linkage control strategy to obtain the comprehensive score.

5. The AI-driven intelligent price inquiry decision-making method according to claim 4, characterized in that: The result output module generates a list of inquiry results as follows: The result output module receives the comprehensive score, recommendation level, and time matching information generated by the time matching calculation unit for each product option from the intelligent decision engine. For product options associated with pre-sale promotions indicated by the time matching information, the result output module calls a comparison scheme generation unit to generate comparison data containing immediate execution schemes and delayed execution schemes, and integrates the comparison data into the display information of the product options to jointly form the inquiry result list.

6. The AI-driven intelligent price inquiry decision-making method according to claim 1, characterized in that: The data aggregation module drives the automated execution component to acquire product data as follows: The data aggregation module is connected to the automated execution component through a control command interface, and sends control commands containing the target platform access address and the simulated operation sequence to the automated execution component according to the linkage control strategy. The automated execution component has a built-in AI analysis unit, which is used to analyze the page structure after loading the target platform page, locate the search box, login form and product list elements, and simulate user operations to complete data collection. The data aggregation module also includes a data standardization unit, which is connected to the data output end of the automated execution component. The data standardization unit is used to convert the collected raw heterogeneous data into standardized data records with a unified field structure and send them to the intelligent decision engine.

7. The AI-driven intelligent price inquiry decision-making method according to claim 1, characterized in that: It also includes anomaly detection steps: The anomaly detection module performs the anomaly detection step, with its input connected to the output of the data aggregation module and the output of the intelligent decision engine, in order to obtain the product data and preliminary scoring results. The anomaly detection module is equipped with a causal knowledge base. When anomalies in price, inventory, or logistics data are detected, the causal knowledge base is called to perform root cause matching and inference, and an anomaly report and handling suggestions are generated. The output of the anomaly detection module is connected to the result output module, and is used to send the anomaly report and processing suggestions to the result output module, which then marks them as additional information in the corresponding item of the inquiry result list.

8. The AI-driven intelligent price inquiry decision-making method according to claim 1, characterized in that: The method also includes a cache management step: The cache management steps are executed through a cache management module, which is communicatively connected to the request parsing module, the intelligent decision engine, and the result output module. The cache management module associates and stores the key parameters of this inquiry task, the product data, and the final generated list of inquiry results in a distributed cache database, and sets a sliding expiration time for the cache item. When the request parsing module receives a new user inquiry request, the cache management module performs a similarity matching query in the distributed cache database. If a match is found and the cache has not expired, the cache result is directly returned through the result output module, and the sliding expiration time of the cache item is reset.

9. An AI-driven intelligent price inquiry system, characterized in that, The AI-driven intelligent price inquiry decision method, including any one of claims 1-7, comprises: The request parsing module is used to receive and parse user inquiry requests, and extract the target product specifications and target inquiry platform; The mode decision module, whose input is connected to the output of the request parsing module, is used to generate a linkage control strategy based on multiple pre-configured control modes and linkage rules. The data aggregation module, whose control input is connected to the output of the mode decision module, is used to drive the automated execution component to collect commodity data based on the linkage control strategy. The intelligent decision engine has its data input end connected to the output end of the data aggregation module and its strategy input end connected to another output end of the pattern decision module, and is used to perform multi-dimensional comprehensive scoring of commodity data. The result output module, whose input is connected to the output of the intelligent decision engine, is used to generate and output an optimized list of inquiry results. The control mode management module has a configuration interface for external configuration of parameters for the time-priority mode, cost-priority mode, and comprehensive balance mode. Its output is connected to the rule update input of the mode decision module for injecting or updating rules into the linkage rule base of the mode decision module. An anomaly detection module, one input of which is connected to the data output of the data aggregation module, and the other input is connected to the intermediate result output of the intelligent decision engine, and its output is connected to the additional information input of the result output module; The data enhancement module has its trigger input connected to the task completion signal output of the data aggregation module, its data input connected to the standardized data output of the data aggregation module, and its data output connected to the extended data input of the intelligent decision engine. The cache management module has a first interface connected to the query interface of the request parsing module, a second interface connected to the result output interface of the intelligent decision engine, and a third interface connected to the cache read / write interface of the result output module.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the AI-driven intelligent price inquiry decision method as described in any one of claims 1 to 8.

Images