Dynamic evaluation-based AI virtual counter marketing method and system

Abstract

The invention discloses an AI virtual counter marketing method and system based on dynamic evaluation, and the method comprises the steps: constructing a marketing function at a server, accessing an online business system through a standardized interface, and receiving operation parameters configured through a console; when the target user enters the ordering page, triggering and popping up a virtual counter interaction interface; collecting commodity data, calculating attraction scores of commodities, and forming a to-be-replaced commodity list based on the attraction scores; based on the attraction score and the real-time display state of the commodity, counter display effect evaluation is carried out to carry out commodity batch replacement judgment; and updating the commodities displayed in the counter according to the judgment result of the commodity batch replacement judgment and the to-be-replaced commodity list. An innovative attraction calculation algorithm and a display effect evaluation algorithm are introduced to eliminate low-efficiency commodities in real time, the display effect of the whole counter is quantitatively evaluated, and batch replacement is intelligently triggered.

Description

Technical Field

[0001] This invention relates to the field of internet marketing technology, specifically to an AI virtual counter marketing method and system based on dynamic evaluation. Background Technology

[0002] With the deepening development of internet services, benefit malls (such as memberships, points, and digital service subscriptions) have become important platforms for enterprises to enhance user stickiness and monetize value. In such scenarios, the core operational objective is to accurately recommend suitable benefit products to users, thereby increasing subscription rates and user satisfaction.

[0003] Currently, most common marketing recommendation solutions are based on list recommendations using collaborative filtering or content recommendation, or traditional customer service chatbots. For list recommendations, these technologies cannot respond in real-time to changes in market strategies and shifts in user interests; traditional customer service chatbots focus on answering questions and struggle to deeply integrate with product displays and real-time user behavior data, failing to achieve dynamic and accurate product recommendations and updates based on dialogue content. Existing solutions only focus on the performance of individual products (such as CTR), lacking a quantitative evaluation and optimization mechanism for the synergistic effect of the entire product portfolio; product display updates rely on fixed periods or manual operation, failing to automatically and intelligently optimize display content based on real-time user feedback data (clicks, purchases), leading to a decline in the effectiveness of counter displays over time. Summary of the Invention

[0004] The purpose of this invention is to provide an AI virtual counter marketing method and system based on dynamic evaluation, so as to solve the problems mentioned in the background art, such as the static nature of existing recommendation strategies and the lack of overall effect evaluation.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: According to one aspect of the present invention, an AI virtual counter marketing method based on dynamic evaluation is provided, the method comprising: building marketing functions on the server side and connecting to an online business system through a standardized interface, and receiving operating parameters configured through a console; When the target user enters the order page, a virtual counter interaction interface is triggered and pops up; Collect product data and calculate the attractiveness score of each product, and form a list of products to be replaced based on the attractiveness score; Based on the attractiveness score and the real-time display status of the products, the counter display effect is evaluated to determine the batch replacement of products. Based on the determination result of the batch replacement of goods and the list of goods to be replaced, the goods displayed in the counter are updated.

[0006] Based on the aforementioned scheme, the operating parameters include, but are not limited to, user configuration, number of products displayed on the counter, display period, attention threshold, single product replacement threshold, and number of products replaced in a single transaction.

[0007] Based on the aforementioned solution, the user configuration includes providing a user filtering interface that supports combined queries and filtering based on one or more of the following conditions: user basic attributes, ordered products, and most recent login time.

[0008] Based on the aforementioned scheme, the formula for calculating the attraction score is: ; α is the attraction score, w y C represents the weighting factor within the time period y. y B represents the number of product clicks within the time period y. y S represents the number of times a product is purchased within a time period y. y This represents the number of times a product is displayed within a time period y, where σ is the decay coefficient and t is the current time; where w y σ and σ are preset parameters.

[0009] Based on the aforementioned scheme, the attractiveness score of a product is compared with a preset single-item replacement threshold. If the attractiveness score is lower than the single-item replacement threshold, the product is marked as a product to be replaced, and the products are sorted according to the attractiveness score to form a list of products to be replaced.

[0010] Based on the aforementioned solution, the evaluation of the counter display effect includes calculating a counter display effect score: ; F represents the score for the display effect of the counter, and n represents the total number of products in the counter. Let i be the attractiveness score of product i at time t. It is a binary function representing whether product i is displayed at time t.

[0011] Based on the aforementioned scheme, the counter display effect score is compared with a preset attention threshold. If the counter display effect score is less than the preset attention threshold, a batch replacement of products is triggered.

[0012] Based on the aforementioned solution, updating the products displayed in the counter also includes periodically performing a forced refresh of the counter products according to a preset display cycle.

[0013] Based on the aforementioned solution, the virtual counter interaction interface includes a virtual teller and a dynamic product display counter; the virtual teller engages in multi-round dialogue with the user through natural language processing technology, identifies the user's intent, and recommends products.

[0014] According to another aspect of the present invention, an AI virtual counter marketing system based on dynamic evaluation is provided. The system includes: a configuration management module, a virtual counter interaction module, an attractiveness calculation module, a display effect evaluation module, and a product update module. The configuration management module is used by operators to configure operating parameters; The virtual counter interaction module is used to dynamically display products, engage in multi-round dialogue with users, identify user intent, and recommend products. The attraction calculation module is used to calculate the attraction score of a product based on product data and generate a list of products to be replaced. The display effect evaluation module is used to calculate the counter display effect score and determine the batch replacement of goods; The product update module performs product replacement and update operations based on the determination result of the batch product replacement judgment and the list of products to be replaced, or according to a preset display cycle.

[0015] As can be seen from the above technical solution, the present invention has at least the following advantages and positive effects compared with the prior art: by accurately positioning the triggering time of the virtual counter at the ordering or renewal page where the user's intention is strongest, intervention is carried out at the best decision point, and combined with multi-round interaction to deeply explore user needs, the accuracy and success rate of marketing conversion are significantly improved; an innovative attraction calculation algorithm (α) and display effect evaluation algorithm (F) are introduced to eliminate inefficient products in real time, quantitatively evaluate the display effect of the entire counter and intelligently trigger batch replacement, ensuring that the counter always maintains the best attraction; the combination of virtual tellers and dynamic counters provides a humanized and interactive shopping experience, which significantly enhances the user's service experience and satisfaction.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention. It is obvious that the drawings described below are merely some embodiments of the invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings: Figure 1 This is a schematic diagram of an AI virtual counter marketing method based on dynamic evaluation according to the present invention; Figure 2 This is a schematic diagram of the batch replacement process for the products of this invention. Detailed Implementation

[0018] To more clearly illustrate the purpose, technical solutions, and advantages of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein. On the contrary, these embodiments are provided so that the present invention will be more comprehensive and complete, and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0019] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a full understanding of embodiments of the invention. However, those skilled in the art will recognize that the technical solutions of the invention can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of the invention.

[0020] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0021] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily need to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0022] The present invention will now be described in detail with reference to specific embodiments: Example 1

[0023] like Figure 1 As shown, this embodiment provides an AI virtual counter marketing method based on dynamic evaluation. Taking the "Benefit Mall" application scenario as an example, the specific steps of the method are as follows: S1: Build marketing functions on the server side and connect to the online business system through standardized interfaces to receive the running parameters configured through the console.

[0024] In this embodiment, the marketing function specifically refers to a virtual counter marketing tool, which is an intelligent marketing software system and service platform designed specifically for the rights and benefits mall and similar business scenarios. It seamlessly integrates with existing business systems through predefined standardized interfaces and protocols, forming a unified management and execution marketing capability platform. The core of the tool includes an integrated management console and algorithm engine. Operators can intuitively configure user targeting, display rules, and algorithm parameters through the console interface, thereby customizing the interaction and recommendation strategies of the virtual counter. This tool, as the carrier entity of this method, encapsulates the data processing, algorithm calculation, decision-making, and front-end display functions into an automated process, providing the necessary technical foundation and operating environment for implementing this method. This embodiment builds the virtual counter marketing tool on the server side and integrates with the rights and benefits mall through predefined standardized interfaces.

[0025] Furthermore, operators configure parameters through a dedicated management console interface, setting user configuration, basic configuration, and advanced configuration to customize the behavior strategy of the virtual counter. The basic configuration consists of core operating parameters, which are the core engine parameters controlling the behavior of the virtual counter and algorithmic decisions. The advanced configuration is used to fine-tune the interactive experience and additional functions for front-end users.

[0026] For user configuration, which precisely defines the target users of the virtual counter, basic attributes such as username, user ID, registration time, most recent login time, user activity level, historical order volume, and number of items added to cart can be configured for filtering; as well as advanced criteria such as "ordered items include: [drop-down selection of items]", "unordered items include: [drop-down selection of items]", "login within the last [ ] days", "registration time before / after [date]", and "user who recently ordered [product name]". "ordered items include: [drop-down selection of items]" targets users who have already purchased a product for cross-selling; "unordered items include: [drop-down selection of items]" targets users who have not purchased a product for promotion; "login within the last [ ] days" targets recently active users; "registration time before / after [date]" sets different strategies for new and old users; and "user who recently ordered [product name]" targets users who have just completed an order for additional sales. A user filtering interface is provided, supporting combined queries and filtering based on one or more of the following criteria: basic user attributes, ordered items, and most recent login time. Different counter content can be displayed to user groups with different characteristics to achieve precise marketing.

[0027] For the basic configuration, you can configure the virtual teller image, the number of items displayed on the counter, the display period, the attention threshold (F threshold), the single-item replacement threshold (α threshold), and the number of items replaced per time. The virtual teller image is an AI digital human figure, and its clarity can be adjusted according to server load. The number of items displayed on the counter sets the number of items displayed at one time. The display period sets a fixed time interval (e.g., 10 minutes, 30 minutes, 1 hour) to periodically refresh the counter's inventory. The attention threshold (F threshold) sets a critical value for the counter's display effect score F; when the algorithm-calculated F value falls below this threshold, batch item replacement is triggered. The single-item replacement threshold (α threshold) sets a critical value for the item's attractiveness score α; when an item's α value falls below this threshold, the item is marked as "pending replacement." The number of items replaced per time determines the number of items replaced each time when batch item replacement is triggered.

[0028] For advanced configurations, features such as voice conversion and mode switching can be configured. Voice conversion controls whether the virtual teller's text responses are converted into voice broadcasts to the user. Mode switching allows switching the console's color theme, adjusting the font size and contrast of the front-end page, and configuring a list of supported languages.

[0029] S2: When the target user enters the order page, a virtual counter interaction interface is triggered and pops up.

[0030] In this embodiment, when a key business event of a user entering the order page is detected, an intelligent "virtual counter interaction interface" is automatically triggered and pops up. This interface combines "humanized dialogue" (virtual teller) with "visualized product display" (dynamic product display counter) to proactively provide users with precise marketing services at the best time and in the best way, thereby greatly improving the user experience and the marketing efficiency of merchants.

[0031] Order pages, including initial order pages and renewal pages, are key conversion nodes with specific business attributes and user intent. When a user reaches this page, it indicates they have completed the preliminary decision-making process of selecting and comparing products, and have a clear intention to purchase or renew, placing them in the final stage before a transaction is completed. At this point, the user's attention is highly focused, and they are more receptive to relevant recommendations. Order pages are designed to allow users to confirm and complete their order or renewal, typically including elements such as an order summary, payment method selection, and a final confirmation button. Initial order pages are confirmation pages for users purchasing a benefit product (such as a membership card or service package) for the first time. Renewal pages guide users to renew their existing benefit services when they are about to expire or have already expired. Triggering a virtual counter on this page for precise and efficient secondary marketing or cross-selling can significantly improve marketing conversion and success rates.

[0032] Furthermore, the virtual counter interface is a composite and intelligent interactive interface, including virtual teller interaction and a dynamic product display counter. The virtual teller interaction is an AI digital human interaction interface with multi-turn dialogue, intent recognition, and semantic understanding capabilities. Interaction content can include uncovering user needs, such as proactively greeting ("Hello, I see you are ordering service X. Would you like to learn about the recently popular service Y?") or answering user questions; understanding the user's potential needs based on the dialogue content to provide accurate and personalized product recommendations; and answering user questions about platform rules, product details, etc. Through human-like interaction, the virtual teller interaction builds trust, deeply explores user needs, and guides them to focus on recommended products.

[0033] Specifically, the virtual teller engages in multi-round interactions, using Natural Language Processing (NLP) technology to conduct multiple dialogues with users. During these dialogues, the virtual teller proactively uncovers user needs, identifies user intent, and recommends products, ultimately guiding the user towards precise recommendations and transactions for beneficial items. First, intent recognition is performed. The initial questions input by the user are quickly categorized to determine whether the current dialogue falls within the core business scenario of "product recommendation" or other scenarios (such as inquiries or complaints). The system receives text or voice input from the user (converted to text via ASR) and uses NLP technology to extract core vocabulary from the questions. For example, if a user asks, "Are there any good movie deals lately?", the system extracts keywords such as "movie" and "deal." These extracted keywords are then matched against a pre-configured "intent library," typically pre-configured by operations personnel and containing numerous business-related scenarios. The system determines whether the current dialogue intent belongs to a recommended or non-recommended scenario. Next, semantic understanding is performed. If the current dialogue intent is a non-recommended scenario, the system invokes an existing knowledge base (FAQ) or business process engine to directly respond to the user's specific question or guide the user through the process. If it is a recommended scenario, a multi-round question-and-answer interaction is initiated.

[0034] Specifically, the system initiates multiple rounds of question-and-answer interaction, extracting more specific "entity" information from the user's initial questions. Since the user's initial needs are often vague, follow-up questions are asked to confirm and supplement key information to build a complete user profile. The entity information confirmed in each round of dialogue is passed to the backend recommendation algorithm as real-time input parameters. Based on the increasingly clear user profile, the recommendation algorithm dynamically filters and sorts the most matching product list from the product library. This dynamically generated recommendation list refreshes the "dynamic product display counter" in real time. Users can add recommended products to their cart or place bulk orders directly without leaving the current dialogue interface.

[0035] The dynamic product display counter is a visual and interactive product recommendation area that can display multiple (e.g., 8-15) recommended products side-by-side in a grid, list, or 3D shelf format. Each product module includes core information such as image, name, key benefits, and price. The displayed products are not fixed but dynamically updated in real-time or near real-time based on algorithm calculations (attractiveness score α, counter display effect F) to ensure that only the most attractive products are displayed at any given time. Users can directly click on the counter interface to view details or add items to their cart / order in bulk with a single click.

[0036] S3: Collect product data and calculate the attractiveness score of the products, and form a list of products to be replaced based on the attractiveness score.

[0037] In this embodiment, core user behavior data for each product is collected in real time or at regular intervals within a specific time period, including the number of times the product is displayed, the number of times it is clicked, and the number of times it is purchased. The number of times the product is displayed indicates the number of times the product is shown to the user in the display case, the number of times the user clicks to view the product details, and the number of times the user ultimately purchases the product. Data collection is achieved through event tracking technology. When the user performs a corresponding action, the front end (virtual display case interface) sends a log record to the server. The back end service aggregates these logs to form the historical and real-time data streams for each product.

[0038] Furthermore, a product attractiveness score is calculated based on the collected product data to objectively measure a product's overall recent appeal to users. The calculation formula is as follows: ; α is the attraction score; w y The weighting factor for time period y is determined by operations personnel based on parameters pre-configured in the marketing calendar, and is applied during major promotional holidays (such as Singles' Day). y Set to 1.6, w for a regular promotional day y (Set to 1.2) All product clicks and purchases will be given higher weight, so that the best-selling products during the holiday can remain on the shelves for a period of time after the holiday, extending their popularity; Based on conversion rate, C represents a comprehensive measure of positive user behavior (clicks and purchases). y B represents the number of product clicks within the time period y; y S represents the number of times a product was purchased within a time period y; y This indicates the number of times the product is displayed within the time period y; The time decay factor amplifies the influence of recent data and weakens the influence of long-term data. t is the current time, and σ is the decay coefficient, which is configured by the operations team. The larger the value, the faster the forgetting speed. The value of σ ranges from 0.1 to 0.5, and it is configured to use a larger decay coefficient value for high-frequency users and a smaller decay coefficient value for low-frequency users.

[0039] Furthermore, the system determines whether to replace a product based on its attractiveness score. It compares the attractiveness score α of all products with the single-product replacement threshold set by the operators in the console, and sorts all products by attractiveness score α from low to high. Products with low attractiveness (α below the threshold) are automatically marked as "to be replaced" and a list of products to be replaced is generated. Furthermore, when a batch of products needs to be replaced in the future, the N products with the lowest attractiveness score α are replaced first.

[0040] S4: Based on the attractiveness score and the real-time display status of the products, evaluate the counter display effect to determine the batch replacement of products.

[0041] Counter display effectiveness evaluation refers to assessing the overall performance of the current product mix at the counter; the real-time display status of the products. , indicating whether product i is displayed in the counter at a specific time t, is represented by a binary function (0 or 1). Based on the attractiveness score calculated in step S3 and the real-time display status of the product, the counter display effect score F is calculated using the following formula: ; in, This means that at any time t, all currently displayed ( The attractiveness score of the product. The sum of these values ​​represents the instantaneous total attractiveness of all items on the counter at that moment. The logarithmic scores of the instantaneous attractiveness calculated at each discrete time point over a period of time (from t0 to t1) are summed together to assess the sustained performance over the entire period.

[0042] Furthermore, the calculated counter display effect score F is compared with the attention threshold set in the control panel beforehand; if F is greater than or equal to the threshold, it means that the overall performance of the current counter's product mix is ​​good and no batch update is required; if F is less than the threshold, batch replacement of products is triggered.

[0043] S5: Update the products displayed in the counter based on the determination result of the batch replacement of the products and the list of products to be replaced.

[0044] like Figure 2As shown, the triggering conditions for batch product replacement include: the counter display effect score F is less than a threshold, and a forced refresh is performed according to a pre-configured display cycle (e.g., every 10 minutes). When batch product replacement is triggered, all products in the current counter are sorted from lowest to highest based on their attractiveness score, generating a list of products to be replaced. Based on the operationally configured number of products to be replaced per batch (N), the N products with the lowest attractiveness scores in this list are selected for batch replacement. New products are selected from the activity product library, prioritizing new products in the same category as the replaced products. Within this category, they are sorted by the highest pageviews or other operational metrics, and the top N new products are selected. To avoid interfering with the user's current browsing and operation, and to prevent full-page refreshes, front-end partial refresh technologies (such as Ajax or WebSocket) can be used to update the page. For entirely new products, since there is no historical data, their initial attractiveness score is 0 or very low. They need to be manually added to the counter by operations personnel or replaced through a fixed display cycle refresh mechanism to gain display opportunities and accumulate data.

[0045] Furthermore, a new round of monitoring and decision-making cycle begins, continuously collecting product data and cyclically performing product attractiveness calculations, display effect evaluations, and product updates to achieve continuous adaptive optimization of product display strategies.

[0046] It should be noted that the method described in this invention is not only applicable to rights and benefits malls, but also to any online business platform that needs to recommend and market products, such as digital content subscription platforms, online membership service systems, and software-as-a-service (SaaS) product renewal scenarios. Example 2

[0047] This embodiment exemplifies an AI virtual counter marketing system based on dynamic evaluation, including a configuration management module, a virtual counter interaction module, an attractiveness calculation module, a display effect evaluation module, and a product update module.

[0048] The configuration management module is used by operations personnel to configure operating parameters. Operating parameters include, but are not limited to, user configuration, number of products displayed on the counter, display period, attention threshold, single product replacement threshold, and number of products replaced in a single transaction. User configuration includes providing a user filtering interface, which supports combined queries and filtering based on one or more of the following conditions: user basic attributes, ordered products, and most recent login time.

[0049] The virtual counter interaction module is used to dynamically display products to target users, engage in multi-round dialogue with users, identify user intent, and recommend products.

[0050] The attractiveness calculation module is used to calculate the attractiveness score of products based on product data and generate a list of products to be replaced; the formula for calculating the attractiveness score is: ; α is the attraction score, w y C represents the weighting factor within the time period y. y B represents the number of product clicks within the time period y. y S represents the number of times a product is purchased within a time period y. y This represents the number of times a product is displayed within a time period y, where σ is the decay coefficient and t is the current time; where w y σ and σ are preset parameters. The attractiveness score of a product is compared with a preset single-item replacement threshold. If the attractiveness score is lower than the single-item replacement threshold, the product is marked as a product to be replaced, and the products are sorted according to their attractiveness scores to form a list of products to be replaced.

[0051] The display effect evaluation module is used to calculate the counter display effect score and determine the batch replacement of goods; the counter display effect evaluation includes calculating the counter display effect score: ; F represents the score for the display effect of the counter, and n represents the total number of products in the counter. Let i be the attractiveness score of product i at time t. It is a binary function representing whether product i is displayed at time t; the counter display effect score is compared with the preset attention threshold. If the counter display effect score is less than the preset attention threshold, the product is replaced in batches.

[0052] The product update module performs product replacement and update operations based on the judgment results of the batch product replacement and the list of products to be replaced, or according to the preset display cycle.

[0053] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the claims. It should be understood that the invention is not limited to the precise structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. An AI-powered virtual counter marketing method based on dynamic evaluation, characterized in that, The method includes: Build marketing functions on the server side and connect to the online business system through standardized interfaces to receive the running parameters configured through the console; When the target user enters the order page, a virtual counter interaction interface is triggered and pops up; Collect product data and calculate the attractiveness score of each product, and form a list of products to be replaced based on the attractiveness score; Based on the attractiveness score and the real-time display status of the products, the counter display effect is evaluated to determine the batch replacement of products. Based on the determination result of the batch replacement of goods and the list of goods to be replaced, the goods displayed in the counter are updated.

2. The AI ​​virtual counter marketing method based on dynamic evaluation according to claim 1, characterized in that, The operating parameters include, but are not limited to, user configuration, number of products displayed on the counter, display period, attention threshold, single product replacement threshold, and number of products replaced in a single transaction.

3. The AI ​​virtual counter marketing method based on dynamic evaluation according to claim 2, characterized in that, The user configuration includes providing a user filtering interface that supports combined queries and filtering based on one or more of the following conditions: user basic attributes, ordered products, and most recent login time.

4. The AI ​​virtual counter marketing method based on dynamic evaluation according to claim 1, characterized in that, The formula for calculating the attraction score is: ； α is the attraction score, w y C represents the weighting factor within the time period y. y B represents the number of product clicks within the time period y. y S represents the number of times a product is purchased within a time period y. y This represents the number of times a product is displayed within a time period y, where σ is the decay coefficient and t is the current time; where w y σ and σ are preset parameters.

5. The AI ​​virtual counter marketing method based on dynamic evaluation according to claim 1, characterized in that, The attractiveness score of a product is compared with a preset single-item replacement threshold. If the attractiveness score is lower than the single-item replacement threshold, the product is marked as a product to be replaced, and the products are sorted according to the attractiveness score to form a list of products to be replaced.

6. The AI ​​virtual counter marketing method based on dynamic evaluation according to claim 1, characterized in that, The evaluation of counter display effectiveness includes calculating a counter display effectiveness score: ； F represents the score for the display effect of the counter, and n represents the total number of products in the counter. Let i be the attractiveness score of product i at time t. It is a binary function representing whether product i is displayed at time t.

7. The AI ​​virtual counter marketing method based on dynamic evaluation according to claim 6, characterized in that, The counter display effect score is compared with a preset attention threshold. If the counter display effect score is less than the preset attention threshold, a batch replacement of products is triggered.

8. The AI ​​virtual counter marketing method based on dynamic evaluation according to claim 1, characterized in that, The updating of the products displayed in the counter also includes periodically refreshing the products in the counter according to a preset display cycle.

9. The AI ​​virtual counter marketing method based on dynamic evaluation according to claim 1, characterized in that, The virtual counter interface includes a virtual teller and a dynamic product display counter; the virtual teller uses natural language processing technology to engage in multi-round dialogue with the user, identify the user's intent, and recommend products.

10. An AI virtual counter marketing system based on dynamic evaluation, characterized in that, It includes a configuration management module, a virtual counter interaction module, an attractiveness calculation module, a display effect evaluation module, and a product update module; The configuration management module is used by operators to configure operating parameters; The virtual counter interaction module is used to dynamically display products, engage in multi-round dialogue with users, identify user intent, and recommend products. The attraction calculation module is used to calculate the attraction score of a product based on product data and generate a list of products to be replaced. The display effect evaluation module is used to calculate the counter display effect score and determine the batch replacement of goods; The product update module performs product replacement and update operations based on the determination result of the batch product replacement judgment and the list of products to be replaced, or according to a preset display cycle.

Images