Recruitment service recommendation method, device and equipment based on multiple AI agents

Abstract

This application discloses a method, apparatus, and device for recommending recruitment services based on multiple AI agents. The method includes: acquiring first dynamic behavioral data of a target enterprise and its posted target positions; analyzing the first dynamic behavioral data using an AI diagnostic agent to determine if the target enterprise's recruitment business has entered a target node, where the target node includes at least one of a node with reduced exposure, a node with reduced screening efficiency, and a node with reduced communication response rate; matching at least one target recruitment value-added service corresponding to the target node from a pre-set service knowledge graph using an AI recommendation agent, and generating scenario-based service recommendation information based on the target enterprise's information and the target recruitment value-added service; recommending the service recommendation information to the target enterprise; and acquiring second dynamic behavioral data of the target enterprise and the target position after the target enterprise uses the target recruitment value-added service. Utilizing the embodiments of this application can improve the accuracy of recruitment service recommendations.

Description

Technical Field

[0001] This application relates to the field of Internet technology, as well as recruitment services and artificial intelligence technology, and in particular to a recruitment service recommendation method, apparatus, equipment, program product and storage medium based on multiple AI agents. Background Technology

[0002] As online recruitment platforms continue to deepen their service models, intelligently recommending recruitment value-added services to corporate users has become a core direction for business development. However, the recommendation methods widely used by current platforms often have the following limitations: First, they rely on preset fixed service packages or trigger mechanisms based on simple rules (such as time thresholds or single behavior counts). This approach is rigid and cannot accurately perceive the dynamic changes in a company's recruitment status and real-time pain points, leading to inaccurate recommendations and user dissatisfaction. Second, the lack of closed-loop tracking and quantitative evaluation of service usage effects makes it difficult to make the service value explicit, failing to provide a credible basis for companies' subsequent repurchase decisions and hindering the continuous optimization of recommendation strategies. Summary of the Invention

[0003] In view of this, embodiments of this application provide a recruitment service recommendation method, electronic device, computer-readable storage medium, and computer program product based on multiple AI agents, solving at least one technical problem.

[0004] This application provides a recruitment service recommendation method based on multiple AI agents, comprising: acquiring first dynamic behavioral data of a target enterprise and its published target positions; analyzing the first dynamic behavioral data using an AI diagnostic agent to determine that the target enterprise's recruitment business has entered a target node, the target node including at least one of a node with reduced exposure, a node with reduced screening efficiency, and a node with reduced communication response rate; matching at least one target recruitment value-added service corresponding to the target node from a pre-set service knowledge graph using an AI recommendation agent, and generating scenario-based service recommendation information based on the target enterprise's enterprise information and the target recruitment value-added service; recommending the service recommendation information to the target enterprise, and acquiring second dynamic behavioral data of the target enterprise and target positions after the target enterprise uses the target recruitment value-added service; comparing the first dynamic behavioral data and the second dynamic behavioral data using an AI effect tracking agent, and generating a service effect report based on the comparison results.

[0005] Optionally, according to the method of this application embodiment, the AI ​​diagnostic agent includes a recruitment process health time series model. The recruitment process health time series model can determine whether the recruitment business has entered a target node by analyzing the changing trend of first dynamic behavior data over time and comparing it with a preset threshold. The step of analyzing the first dynamic behavior data through the AI ​​diagnostic agent to determine whether the recruitment business of the target enterprise has entered a target node includes: when the recruitment process health time series model identifies that the first dynamic data meets the preset threshold, it determines that the recruitment business has entered a target node. The first dynamic behavior data includes at least one of the following: the growth rate of pageviews of the target position, the resume submission conversion rate, the set of keywords actively searched by the enterprise, and the communication response rate. Wherein, if the growth rate of pageviews is lower than the first threshold, it is determined that the recruitment business has entered a node of reduced exposure; if the resume submission conversion rate is lower than the second threshold, it is determined that the recruitment business has entered a node of reduced screening efficiency; if the communication response rate is lower than the third threshold, it is determined that the recruitment business has entered a node of reduced communication response rate.

[0006] Optionally, according to the method of this application embodiment, the step of matching at least one target recruitment value-added service corresponding to the target node from a preset service knowledge graph includes: querying the mapping relationship between services and target node types recorded in the service knowledge graph according to the type of the target node into which the recruitment business enters; retrieving at least one candidate recruitment value-added service associated with the target node type based on the mapping relationship; and selecting at least one target recruitment value-added service that matches the enterprise information from the at least one candidate recruitment value-added service in conjunction with the enterprise information of the target enterprise; wherein the enterprise information includes at least one of the enterprise's industry, enterprise size, and historical service purchase records.

[0007] Optionally, according to the method of this application embodiment, the step of generating scenario-based service recommendation information based on the enterprise information of the target enterprise and the target recruitment value-added services includes: determining the recruitment urgency of the target enterprise based on the first dynamic behavior data; performing combination logic matching between the matched at least one target recruitment value-added service and the recruitment urgency to generate a customized service package solution; generating explanatory text for the service package solution, the text including at least a description of the target node, the core pain points that the service package solution can solve, and the effect improvement indicators estimated based on the enterprise's historical data or industry benchmark data; and encapsulating the service package solution, its corresponding explanatory text, and selection entry point into the scenario-based service recommendation information.

[0008] Optionally, according to the method of this application embodiment, recommending the service recommendation information to the target enterprise includes: embedding the service recommendation information in a non-intrusive manner and presenting it in the recruitment business function interface currently being operated by the target enterprise, so as to guide the target enterprise to complete the service selection.

[0009] Optionally, according to the method of this application embodiment, the step of comparing the first dynamic behavior data and the second dynamic behavior data through the AI ​​effect tracking agent and generating a service effect report based on the comparison result includes: the AI ​​effect tracking agent extracting service-related target indicators from the first dynamic behavior data as baseline data, and extracting corresponding indicators within the same time window from the second dynamic behavior data as effect data; the AI ​​effect tracking agent performing causal inference analysis or difference significance test on the baseline data and the effect data to quantify the net effect of the use of the target recruitment value-added service on each target indicator; and the AI ​​effect tracking agent integrating the net effect, the magnitude of indicator change, and the achieved business results to generate a structured and visualized service effect report, wherein the business results include at least one of the following: the number of resumes submitted due to service use, the improved communication response rate, or the increased number of interview invitations.

[0010] This application provides a recruitment service recommendation device based on multiple AI agents, comprising: an acquisition module for acquiring first dynamic behavioral data of a target enterprise and its published target positions; an analysis module for analyzing the first dynamic behavioral data using an AI diagnostic agent to determine that the target enterprise's recruitment business has entered a target node, the target node including at least one of a node with reduced exposure, a node with reduced screening efficiency, and a node with reduced communication response rate; a matching module for matching at least one target recruitment value-added service corresponding to the target node from a pre-set service knowledge graph using an AI recommendation agent, and generating scenario-based service recommendation information based on the enterprise information of the target enterprise and the target recruitment value-added service; a recommendation module for recommending the service recommendation information to the target enterprise, and acquiring second dynamic behavioral data of the target enterprise and the target position after the target enterprise uses the target recruitment value-added service; and a comparison module for comparing the first dynamic behavioral data and the second dynamic behavioral data using an AI effect tracking agent, and generating a service effect report based on the comparison results.

[0011] This application provides an electronic device, which includes a processor and a memory storing computer program instructions; the electronic device executes the computer program instructions to implement the method described above.

[0012] This application provides a computer program product, which includes computer program instructions that, when executed, implement the method described above.

[0013] This application provides a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the method described above.

[0014] This application's embodiments utilize an AI diagnostic agent to continuously analyze dynamic behavioral data, enabling real-time identification of bottlenecks encountered in the recruitment processes of different enterprises, significantly improving the accuracy and timing of recommendations. Furthermore, by using an AI effect tracking agent to monitor product performance, it not only obtains timely service results but also optimizes recommendation strategies, thereby providing effective support for the intelligent and refined operation of enterprises. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings of the embodiments of this application will be briefly described below.

[0016] Figure 1 This is a schematic diagram of the system architecture of an embodiment of this application.

[0017] Figure 2 This is a flowchart of a recruitment service recommendation method based on multiple AI agents, according to an embodiment of this application.

[0018] Figure 3 This is a schematic diagram illustrating the processing of a recruitment service recommendation method based on multiple AI agents, as described in an embodiment of this application.

[0019] Figure 4 This is a schematic structural block diagram of a recruitment service recommendation device based on multiple AI agents, according to an embodiment of this application.

[0020] Figure 5 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0021] The principles and spirit of this application will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided to make the principles and spirit of this application clearer and more thorough. The exemplary embodiments provided herein are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this application.

[0022] Embodiments of this application relate to terminal devices and / or servers. Implementations of this application can be a system, terminal, device, method, computer-readable storage medium, or computer program product, and can be specifically implemented as entirely hardware, entirely software, or a combination of hardware and software. Figure 1 This diagram illustrates a system architecture according to an embodiment of the present application, including a terminal device 102 and a server 104. The terminal device 102 may include at least one of the following: a smartphone, tablet computer, laptop computer, desktop computer, smart TV, various wearable devices, augmented reality (AR) devices, virtual reality (VR) devices, etc. A client application, such as an app, mini-program, or browser-based client, can be installed on the terminal device 102. Users can input commands through the client, and the terminal device 102 can send request information containing the commands to the server 104. Upon receiving the request information, the server 104 performs corresponding processing and returns the processing result information to the terminal device 102. The server 104 may be a local server or a cloud server, and may be a single server or a server cluster, etc.

[0023] In this document, the terms "first," "second," "third," etc., are used only to distinguish one entity (or operation) from another in textual description, and do not require or imply any sequential order between these entities (or operations).

[0024] Figure 2 This paper illustrates a flowchart of a recruitment service recommendation method based on multiple AI agents according to an embodiment of this application. The method includes the following steps: S101: Obtain first dynamic behavioral data on the target company and the target job postings it has published; S102: Analyze the first dynamic behavior data through the AI ​​diagnostic agent to determine the target enterprise's recruitment business has entered the target node. The target node includes at least one of the following: exposure reduction node, screening efficiency reduction node, and communication response rate reduction node. S103: The AI ​​recommendation agent matches at least one target recruitment value-added service corresponding to the target node from the pre-set service knowledge graph, and generates scenario-based service recommendation information based on the target company's enterprise information and the target recruitment value-added service. S104: Recommend service information to the target company, and obtain the second dynamic behavioral data of the target company and the target position after the target company uses the target recruitment value-added service; S105: The AI ​​effect tracking agent compares the first dynamic behavior data and the second dynamic behavior data, and generates a service effect report based on the comparison results.

[0025] This application proposes a data processing and decision generation method driven collaboratively by multiple dedicated intelligent agents. Firstly, in this application embodiment, it is necessary to continuously collect data on the target company's operational behavior on the recruitment platform and data related to the recruitment effectiveness of the target position, forming first dynamic behavioral data. This dynamic behavioral data can be represented as a series of time-varying interactive events, such as the temporal changes in job pageviews, the keyword set of proactive search behavior, and the frequency and depth of communication conversations. Through real-time analysis of this sequence data by an AI diagnostic agent, abnormal nodes can be identified, such as stagnant exposure growth, decreased resume screening efficiency (reduced number of resumes remaining after screening), or a decrease in the number of responses received from communication with job seekers. At least one of these nodes constitutes a target node. The target company's recruitment business entering a target node also indicates that the recruitment business has encountered a bottleneck, for example, a decrease in the number of times the target position is exposed (encountering an exposure bottleneck), a decrease in resume screening efficiency (encountering a resume screening bottleneck), or a decrease in the response rate to communication with job seekers (encountering a communication bottleneck).

[0026] Once the AI ​​diagnostic agent identifies a specific business node in the recruitment process, the process enters the service matching phase. The system pre-configures a structured service knowledge graph, which is essentially a networked data model representing the relationships between recruitment services, their functional attributes, and the business problems they can solve. The AI ​​recommendation agent receives the diagnostic results as input and performs graph traversal and pattern matching within the knowledge graph to map one or more target recruitment value-added services that functionally correspond directly to the business node. Furthermore, the agent optimizes its strategy by incorporating the target company's own corporate information data, ultimately generating a contextualized recommendation message that includes a specific service combination, expected technical effects, and reasons for suitability.

[0027] After the recommendation information is pushed to the target enterprise's user interface, the system continues to monitor subsequent behavior, collecting second dynamic behavior data generated after the enterprise uses the recommended service. The task of the AI ​​effect tracking agent is to perform before-and-after comparative analysis, using the first dynamic behavior data as a baseline and comparing it with the second dynamic behavior data in multiple dimensions and conducting causal inference analysis. The output is a structured service effect report. This report quantifies the changes in key indicators before and after service intervention from a technical perspective, such as the increase in exposure and the improved matching degree of resumes after screening.

[0028] The method provided in this application upgrades loose, reactive rule-based judgments into a pipelined processing flow consisting of multiple specialized AI agents connected via a standard interface. The AI ​​diagnostic agent transforms raw behavioral sequences into an understanding of business states; the AI ​​recommendation agent enables knowledge retrieval and personalized adaptation of solutions from problem diagnosis; and the AI ​​effect tracking agent establishes a quantifiable and verifiable feedback loop. This organic combination of technical steps ultimately achieves an automated and scalable recommendation mechanism at the system level that can dynamically perceive business states, accurately match solutions, and empirically evaluate intervention effects.

[0029] Optionally, in some embodiments of this application, the AI ​​diagnostic agent includes a recruitment process health time-series model. This model can determine whether the recruitment process has entered a target node by analyzing the changing trend of first dynamic behavior data over time and comparing it with a preset threshold. The step of analyzing the first dynamic behavior data using the AI ​​diagnostic agent to determine whether the target company's recruitment process has entered a target node includes: When the recruitment process health time series model identifies that the first dynamic data meets the preset threshold, it determines that the recruitment business has entered the target node. The first dynamic behavioral data includes at least one of the following: the target job's pageview growth rate, resume submission conversion rate, enterprise's active search keyword set, and communication response rate. Specifically, if the pageview growth rate is lower than the first threshold, the recruitment business is determined to have entered a node of reduced exposure; if the resume submission conversion rate is lower than the second threshold, the recruitment business is determined to have entered a node of reduced screening efficiency; and if the communication response rate is lower than the third threshold, the recruitment business is determined to have entered a node of reduced communication response rate.

[0030] In this application embodiment, the recruitment process health time-series model is a time-series data analysis algorithm or machine learning model designed for the characteristics of the recruitment business process. This model does not examine single-point data in isolation, but continuously receives and processes the first dynamic behavioral data stream generated by user operations, analyzing the evolution patterns and trend characteristics of key indicators over time.

[0031] Specifically, the time-series model incorporates a series of preset thresholds, representing metrics for different stages of the recruitment process. These thresholds can be technically calibrated based on industry benchmarks, historical data, or specific business objectives. During operation, the model continuously compares and matches real-time input behavioral data sequences (e.g., the sliding window calculation of the growth rate of target job views, real-time statistics of resume submission conversion rates, and dynamic changes in communication response rates) with these preset technical thresholds.

[0032] When the model identifies one or more behavioral data indicator sequences that consistently or significantly deviate from the preset healthy range (e.g., page view growth rate is below the first threshold for several consecutive days), it identifies this phenomenon as a specific bottleneck in the recruitment business, indicating that the target company's recruitment business has entered a target node. Specifically, this can be determined by comparing the page view growth rate with the first threshold to identify the recruitment business entering a node of reduced exposure; by comparing the resume submission conversion rate with the second threshold to identify the recruitment business entering a node of reduced screening efficiency; and by comparing the communication response rate with the third threshold to identify the recruitment business entering a node of reduced communication response rate.

[0033] This application's embodiments transform raw, massive, and seemingly disordered user interaction logs into bottleneck signals with clear business semantics and technical definitions through a time-series analysis model. It solidifies the experiential knowledge of recruitment operations (which data patterns correspond to which problems) into automatically executable computational logic, upgrading the diagnostic process from relying on subjective human experience or simple static rules to highly real-time automated technical analysis based on dynamic data sequences and quantified thresholds, thus improving analysis efficiency.

[0034] In some embodiments of this application, optionally, at least one target recruitment value-added service corresponding to the target node is matched from a pre-set service knowledge graph, including: Based on the type of target node into which the recruitment business enters, query the mapping relationship between the services and target node types recorded in the service knowledge graph; Based on the mapping relationship, at least one candidate recruitment value-added service associated with the target node type is retrieved; Based on the target company's corporate information, at least one target recruitment value-added service that matches the corporate information is selected from at least one candidate recruitment value-added service; the corporate information includes at least one of the following: the industry to which the company belongs, the company size, and historical service purchase records.

[0035] In this embodiment, based on the type of the target node currently entered by the recruitment business, such as the exposure reduction node, the system accesses a pre-built service knowledge graph. This graph is a knowledge base that stores and represents data in a graph structure. Its nodes represent various recruitment value-added services (such as job refresh cards and resume precision exposure packages), and the edges represent the functional relationships between services and the types of business problems they can solve, as well as between services themselves. The entire matching process first uses the type of the target node diagnosed as the query key to search for and locate service nodes in the graph that have a direct mapping relationship with that type. This mapping relationship is predefined and fixed in the graph structure. For example, the exposure reduction node type in the graph will be explicitly linked to a set of services that have the function of increasing exposure. This completes the initial technical mapping from problem to solution, and the retrieved results are called candidate recruitment value-added services.

[0036] The next step in the processing flow is the personalized screening step. The system will concurrently access or call upon stored target company information. This information includes structured classification tags (such as industry and size) and time-series behavioral records (such as historical service purchase records). Here, the recommendation agent executes a multi-dimensional matching algorithm, comparing and weighting the applicability characteristics of candidate services (e.g., some services are more suitable for mid-sized companies in the internet industry, or have a functional complementarity with services already purchased by the user) with the specific information of the current company. For example, if a company has historically purchased a communication membership package, and the AI-powered intelligent screening package in the current candidate services can functionally complement it and improve overall process efficiency, then that service may receive a higher matching score. Finally, based on a set of screening rules or ranking algorithms, the system will output one or more target recruitment value-added services from the candidate set that best fit the current company's context.

[0037] This application's embodiments introduce a structured knowledge representation and retrieval tool—the service knowledge graph—and combine it with real-time filtering of multi-dimensional static enterprise information, making service recommendations no longer simple lists or fixed bundles. Furthermore, this application's embodiments establish a reasoning-based recommendation mechanism based on logical association and context awareness, significantly improving the alignment between recommendation results and users' actual needs.

[0038] In some embodiments of this application, optionally, scenario-based service recommendation information is generated based on the target company's enterprise information and target recruitment value-added services, including: Based on the first dynamic behavioral data, determine the urgency of recruitment for the target company; The at least one target recruitment value-added service obtained through matching is combined with the recruitment urgency using a combination logic to generate a customized service package solution. Generate explanatory text for the service package solution. The text should include at least a description of the target node, the core pain points that the service package solution can solve, and the performance improvement indicators estimated based on the company's historical data or industry benchmark data. The service package options, their corresponding explanatory text, and selection entry points are packaged into scenario-based service recommendation information.

[0039] In this application embodiment, the recruitment urgency is a level or numerical label representing time sensitivity, which is output after analyzing dynamic behavioral data (such as the passage of time after the job posting, the gap between the progress of the application volume and the expected goal, and the surge in the frequency of HR's proactive search behavior).

[0040] After confirming the recruitment urgency, the system enters the combinatorial logic matching phase. This phase does not simply list services, but rather relies on a pre-defined combinatorial logic rule base. This rule base defines the correspondence between different levels of recruitment urgency and different service combination strategies. For example, for a high urgency tag, the rule may trigger the automatic bundling of multiple target services (such as job refresh cards and AI intelligent screening packages) into a scheme named "Rapid Recruitment Package".

[0041] Next, the system receives the following inputs: the target node (e.g., exposure bottleneck), the service package solution, and relevant data (e.g., historical performance data of the enterprise or industry benchmark data). Then, following a preset copy template and natural language generation rules, it automatically fills in and organizes the language, outputting a structured explanatory text. Technically, this text must include several key information elements: an objective description of the identified business bottleneck, i.e., the current stage of the recruitment process; the capabilities of the recommended service package for the target node type; and a quantified expected performance improvement indicator based on data (e.g., an estimated 80% increase in exposure).

[0042] Finally, the customized service package, automatically generated explanatory text, and an interactive selection entry point (such as a button) are all encapsulated together. This encapsulation refers to organizing all information elements into a complete, self-contained service recommendation information data package according to the data structure agreed upon by the system's front-end and back-end (such as a specific JSON object or protocol buffer message). The structure of this data package is designed to ensure precise delivery and rendering at the specific location on the user's current interface.

[0043] The embodiments of this application greatly improve the accuracy, personalization, and operability of recommendation information by using urgency assessment and combinatorial logic matching.

[0044] Optionally, in some embodiments of this application, service recommendation information is recommended to the target enterprise, including: Service recommendation information is embedded and presented in a non-intrusive manner into the target company's current recruitment business function interface to guide the target company in selecting services.

[0045] In this embodiment, when the system generates a scenario-based service recommendation information data packet, it does not push it through a strong interference method such as popping up an independent window, jumping to a new page, or covering the entire screen. Instead, the front-end interface integration module accurately locates and embeds the data based on the context of the recruitment business function interface that the user is currently active in.

[0046] The process first requires identifying the specific functional interface the user is currently interacting with, such as a job management panel, resume screening list, or communication history page. The system then selects a pre-defined, suitable "embedding point" based on the relevance of the recommended information content to the current interface's business logic. For example, for a identified point of decreased exposure, the recommended information is dynamically inserted into the same panel area displaying the job's view count trend, presented as a data card, sidebar tip, or annotation next to a data chart. This embedding operation is typically achieved through dynamic component mounting technology in the front-end framework. The embedded information elements maintain visual consistency with the native interface style and allow users to conveniently view, interact with, or even directly select the job without leaving their current main task.

[0047] This application's embodiments, through interface integration, transform isolated marketing behaviors that might otherwise disrupt task flow and fragment the experience into scenario-driven auxiliary decision support deeply coupled with business processes. This not only significantly reduces the user's workload and operational resistance caused by interface switching and task interruptions, but more importantly, by closely aligning the timing, location, and content of recommendations with the user's current specific work goals and the data they see, it greatly enhances the immediacy, relevance, and understandability of the recommendations.

[0048] Optionally, in some embodiments of this application, an AI effect tracking agent compares the first dynamic behavior data and the second dynamic behavior data, and generates a service effect report based on the comparison result, including: The AI ​​effect tracking agent extracts the target indicators related to the service from the first dynamic behavior data as baseline data, and extracts the corresponding indicators within the same time window from the second dynamic behavior data as effect data. The AI ​​effect tracking agent performs causal inference analysis or difference significance test on baseline data and effect data to quantify the net effect of the use of target recruitment value-added services on each target indicator. The AI-powered performance tracking agent integrates the net effect, the magnitude of changes in metrics, and the achieved business results to generate a structured, visualized service performance report. The business results include at least one of the following: an increase in the number of resumes submitted due to service usage, an improvement in communication response rate, or an increase in the number of interview invitations.

[0049] The AI ​​effect tracking agent in this embodiment can identify and extract several target indicators directly related to the expected effect of the service from the historically stored first dynamic behavior data, based on the functional characteristics of the purchased target recruitment value-added service. For example, if the service is a job refresh card, the core target indicator may be defined as the average daily number of job views; if it is an AI intelligent screening package, it may be the percentage of high-quality resume recommendations. These indicator values ​​extracted from the historical data before the service are established as the evaluation baseline and used as a benchmark for comparison. Subsequently, the agent extracts a window of the same duration as the baseline data statistical period from the second dynamic behavior data (i.e., the new data stream after the service is used) and extracts the exact same indicator sequence as the effect data. This step technically ensures that subsequent comparisons are conducted under the same measurement standard and time scale, avoiding evaluation bias caused by inconsistent time periods or different indicator definitions.

[0050] After data preparation, the AI ​​effect tracking agent doesn't simply calculate the arithmetic difference between the effect data and the baseline data. Instead, it performs more rigorous causal inference analysis or significance testing. Causal inference analysis is a class of advanced statistical or machine learning methods designed to identify the change in indicators directly caused by the intervention of using the target service—the net effect—after controlling for other potential influencing factors (such as seasonal fluctuations or overall market changes). Significance testing, on the other hand, is a statistical hypothesis test used to determine whether the observed change in indicators is large enough to be unlikely to be caused by random fluctuations alone, thus confirming its statistical significance. By applying these analytical methods, the system can go beyond superficial correlations and provide a quantitative and reliable estimate of the true, purely technical effects of the service.

[0051] Finally, the system enters the report synthesis and output stage. It analyzes the calculated net effect, the specific changes in various indicators, and the derived business results that are more easily understood by the business (such as the number of resumes submitted due to increased exposure). Then, following a pre-set template, the system automatically integrates this data, necessary textual explanations, and visualizations (such as trend comparison charts and effect bar charts) to generate a standardized, data-driven, structured, and visualized service effect report. This report itself is a standardized data document or interactive interface.

[0052] This application's embodiments construct a rigorous, automated, and interpretable effect evaluation and feedback loop. Through causal inference or significance testing, effect verification is elevated from simple comparison to quantitative analysis, significantly enhancing the objectivity and credibility of the evaluation results. Simultaneously, the generated structured effect report provides the system with high-quality feedback signals, which can be directly used to optimize the diagnostic and recommendation models, thereby driving the entire recommendation system to achieve continuous self-iteration based on empirical data, forming a complete technology enhancement loop from perception and recommendation to verification and optimization.

[0053] The implementation methods and advantages of the embodiments of this application have been described above through multiple examples. The specific processing procedures of the embodiments of this application are described in detail below with reference to specific examples.

[0054] XX Technology is an internet company that urgently needs to recruit 3 senior Java engineers within two weeks. HR Li Lei posted the job on the recruitment platform. Based on the recruitment service recommendation method of this application embodiment, the information to be obtained includes the static profile of the enterprise: industry, size, historical preferences; and the dynamic behavior of the enterprise on the recruitment platform: page views, search terms, and communication rate. Thus, in this specific embodiment, the information to be obtained includes: (1) Static information (enterprise information): XX Technology (internet industry, Series B financing, 200-500 employees, historically purchased communication membership packages). (2) Initial dynamic behavior (posted job information): posted a senior Java engineer job, the description emphasizes high concurrency and distributed systems. (3) Continuous dynamic behavior flow (first dynamic behavior data): daily page views, application volume, keywords actively searched by HR for candidates (such as microservice architect), average number of communication rounds with candidates, acceptance rate after interview invitations are sent, etc. Figure 3 As shown, the entire process includes the following steps: Step 1: Real-time Diagnosis and Bottleneck Identification (AI Diagnosis). Three days after the job posting, the AI ​​agent analysis revealed that the daily growth rate of job views plummeted from 50% to 5%, indicating an exposure bottleneck. Simultaneously, it detected HR frequently using keywords such as "advanced Java" and "architecture" to actively search, but among the candidates found, the response rate after initiating chat was low (<20%). Combining static information (Internet, Series B funding), the agent inferred that the company likely had high requirements for candidate quality and was in a hurry to recruit, suggesting that current communication efficiency might become the next bottleneck.

[0055] Step 2: Service Mapping and Contextualized Recommendation (AI Recommendation). Addressing exposure bottlenecks, the most direct service matched from the service knowledge graph is the Job Refresh Card (function: refreshes jobs to the top of the list, regaining significant exposure). Addressing potential communication efficiency and quality bottlenecks, a precise resume exposure package (including resume placement benefits) and an AI intelligent screening package are matched (automatically filters and prioritizes resumes with the highest matching degree). Based on the urgency of enterprise recruitment, the above services are dynamically combined into a rapid Java engineer recruitment package, along with estimated effects based on historical data (e.g., an estimated 80% increase in exposure and a 50% increase in screening efficiency).

[0056] Step 3: Interactive Execution and Paid Conversion (Interaction). When HR logs into the platform or views the job data, the system non-intrusively generates a contextualized prompt in the job management dashboard: "We have detected a slowdown in the growth rate of your Senior Java Engineer job posting, indicating a bottleneck in exposure. Based on your efficient search behavior, we have customized a rapid recruitment package for you. Click to view details." After clicking, Li Lei saw the package details, price, and estimated results.

[0057] Step 4: Performance Tracking and Value Demonstration (Performance Tracking). HR purchased the package. The AI ​​agent began tracking: pageview data within 24 hours of using the job refresh card; and the quality of resumes recommended to HR after AI intelligent filtering (measured by subsequent communication rate and interview rate). 24 hours later, the system automatically pushed a message and a mini-report to HR: Service Performance Update: Your rapid recruitment package has taken effect: 1) After job refresh, pageviews increased by 150% on the same day, with 8 new applications; 2) Of the 3 resumes prioritized for you by AI filtering, the communication response rate reached 67%, and 1 candidate has been scheduled for an interview.

[0058] This application's embodiments construct a data-driven, scalable precision marketing and value-added service system. In detail, contextualized prompts greatly improve user experience and marketing acceptance, as well as create perceived service value, thereby increasing user conversion rates towards paid services.

[0059] Correspondingly, this application also provides a recruitment service recommendation device 100 based on multiple AI agents, for reference. Figure 4 ,include: Module 110 is used to acquire the first dynamic behavioral data of the target company and the target job postings it has published; Analysis module 120 analyzes the first dynamic behavior data through an AI diagnostic agent to determine that the recruitment business of the target enterprise has entered a target node. The target node includes at least one of the following: exposure reduction node, screening efficiency reduction node, and communication response rate reduction node. The matching module 130 is used to match at least one target recruitment value-added service corresponding to the target node from the pre-set service knowledge graph through the AI ​​recommendation agent, and generate scenario-based service recommendation information based on the enterprise information of the target enterprise and the target recruitment value-added service. The recommendation module 140 is used to recommend service recommendation information to target companies and obtain second dynamic behavioral data of target companies and target positions after the target companies use the target recruitment value-added services. The comparison module 150 is used to compare the first dynamic behavior data and the second dynamic behavior data through the AI ​​effect tracking agent, and generate a service effect report based on the comparison results.

[0060] Based on at least one of the above embodiments, the electronic device in the embodiments of this application may be a user terminal device, a server, other computing devices, or a cloud server. Figure 5 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. The electronic device may include a processor 601 and a memory 602 storing computer program instructions. The processor 601 reads and executes the computer program instructions stored in the memory 602 to implement the process or function of any of the methods in the above embodiments.

[0061] Specifically, processor 601 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application. Memory 602 may include a mass storage device for data or instructions. For example, memory 602 may be at least one of the following: a hard disk drive (HDD), read-only memory (ROM), random access memory (RAM), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, universal serial bus (USB) drive, or other physical / tangible memory storage device. Alternatively, memory 602 may include removable or non-removable (or fixed) media. Furthermore, memory 602 may be internal or external to the integrated gateway disaster recovery device. Memory 602 may be non-volatile solid-state memory. In other words, typically memory 602 includes a tangible (non-transitory) computer-readable storage medium (such as a memory device) encoded with computer-executable instructions, and when the software is executed (e.g., by one or more processors), it can perform the operations described in the methods of the embodiments of this application.

[0062] As an example, Figure 5The illustrated electronic device may also include a communication interface 603 and a bus 610. The processor 601, memory 602, and communication interface 603 are connected via bus 610 and communicate with each other. Bus 610 may include hardware, software, or both, and may couple components of an online data traffic metering device together. The bus may include at least one of the following: Accelerated Graphics Port (AGP) or other graphics bus, Enhanced Industry Standard Architecture (EISA) bus, Front Side Bus (FSB), HyperTransport (HT) Interconnect, Industry Standard Architecture (ISA) bus, Infinite Bandwidth Interconnect, Low Pin Count (LPC) bus, memory bus, Microchannel Architecture (MCA) bus, Peripheral Component Interconnect (PCI) bus, PCI-Express (PCI-X) bus, Serial Advanced Technology Attachment (SATA) bus, Video Electronics Standards Association Local (VLB) bus, or other suitable bus. Bus 610 may include one or more buses. Although specific buses are described or shown in the embodiments of this application, any suitable bus or interconnection method is contemplated in the embodiments of this application.

[0063] In conjunction with the methods in the above embodiments, this application also provides a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the process or function of any of the methods in the above embodiments.

[0064] This application also provides a computer program product that stores computer program instructions, which, when executed by a processor, implement the process or function of any of the methods described above.

[0065] The flowcharts and / or block diagrams of methods, terminals, systems, and computer program products according to embodiments of this application have been exemplarily described above, and related aspects have been described. It should be understood that each block or combination thereof in the flowcharts and / or block diagrams may be implemented by computer program instructions, by dedicated hardware performing a specified function or action, or by a combination of dedicated hardware and computer instructions. For example, these computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to form a machine such that these instructions, executed via such processor, enable the implementation of the function / action specified in each block or combination thereof in the flowcharts and / or block diagrams. Such a processor may be a general-purpose processor, a dedicated processor, a special-purpose application processor, or a field-programmable logic circuit.

[0066] The functional blocks shown in the structural block diagrams of this application can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc.; when implemented in software, they are programs or code segments used to perform the required tasks. Programs or code segments can be stored in memory or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. Code segments can be downloaded via computer networks such as the Internet or intranets.

[0067] It should be noted that this application is not limited to the specific configurations and processes described above or shown in the figures. The above descriptions are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the described systems, devices, terminals, modules, or units can be referred to the corresponding processes in the method embodiments, and need not be repeated here. It should be understood that the scope of protection of this application is not limited thereto. Any equivalent modifications or substitutions that can be conceived by those skilled in the art within the scope of the technology disclosed in this application should be covered within the scope of protection of this patent application.

Claims

1. A recruitment service recommendation method based on multiple AI agents, characterized in that, include: Obtain first-level dynamic behavioral data on the target company and the target job postings it has published; By analyzing the first dynamic behavior data through an AI diagnostic agent, it is determined that the recruitment business of the target enterprise has entered the target node. The target node includes at least one of the following: exposure reduction node, screening efficiency reduction node, and communication response rate reduction node. The AI ​​recommendation agent matches at least one target recruitment value-added service corresponding to the target node from a pre-set service knowledge graph, and generates scenario-based service recommendation information based on the enterprise information of the target enterprise and the target recruitment value-added service. The service recommendation information is recommended to the target enterprise, and after the target enterprise uses the target recruitment value-added service, the second dynamic behavioral data of the target enterprise and the target position is obtained; The AI-powered performance tracking agent compares the first dynamic behavior data with the second dynamic behavior data and generates a service performance report based on the comparison results.

2. The method according to claim 1, characterized in that, The AI ​​diagnostic agent includes a recruitment process health time series model, which can determine whether the recruitment process has entered the target node by analyzing the changing trend of the first dynamic behavior data over time and comparing it with a preset threshold. The step of analyzing the first dynamic behavior data through an AI diagnostic agent to determine whether the target company's recruitment business has entered the target node includes: When the recruitment process health time series model identifies that the first dynamic data meets the preset threshold, it determines that the recruitment business has entered the target node. The first dynamic behavioral data includes at least one of the following: the target job's pageview growth rate, resume submission conversion rate, enterprise's active search keyword set, and communication response rate. Specifically, if the pageview growth rate is lower than the first threshold, the recruitment business is determined to have entered a node of reduced exposure; if the resume submission conversion rate is lower than the second threshold, the recruitment business is determined to have entered a node of reduced screening efficiency; and if the communication response rate is lower than the third threshold, the recruitment business is determined to have entered a node of reduced communication response rate.

3. The method according to claim 1 or 2, characterized in that, The step of matching at least one target recruitment value-added service corresponding to the target node from a pre-set service knowledge graph includes: Based on the type of target node into which the recruitment business enters, query the mapping relationship between the services and target node types recorded in the service knowledge graph; Based on the mapping relationship, at least one candidate recruitment value-added service associated with the target node type is retrieved; Based on the enterprise information of the target enterprise, at least one target recruitment value-added service that matches the enterprise information is selected from the at least one candidate recruitment value-added service; the enterprise information includes at least one of the following: the industry to which the enterprise belongs, the size of the enterprise, and historical service purchase records.

4. The method according to claim 1, characterized in that, The step of generating scenario-based service recommendation information based on the target company's enterprise information and the target recruitment value-added services includes: Based on the first dynamic behavior data, the recruitment urgency of the target company is determined; The at least one target recruitment value-added service obtained through matching is combined with the recruitment urgency using combination logic to generate a customized service package solution; Generate explanatory text for the service package solution, which includes at least a description of the target node, the core pain points that the service package solution can solve, and the performance improvement indicators estimated based on the enterprise's historical data or industry benchmark data. The service package scheme, its corresponding explanatory text, and selection entry point are encapsulated into the scenario-based service recommendation information.

5. The method according to claim 1, characterized in that, The step of recommending the service recommendation information to the target enterprise includes: The service recommendation information is embedded and presented in a non-intrusive manner into the recruitment business function interface that the target company is currently operating, so as to guide the target company to complete the service selection.

6. The method according to claim 1, characterized in that, The step of comparing the first dynamic behavior data and the second dynamic behavior data using an AI effect tracking agent, and generating a service effect report based on the comparison results, includes: The AI ​​effect tracking agent extracts the service-related target indicators from the first dynamic behavior data as baseline data, and extracts the corresponding indicators within the same time window from the second dynamic behavior data as effect data. The AI ​​effect tracking agent performs causal inference analysis or difference significance test on the baseline data and the effect data to quantify the net effect of the use of the target recruitment value-added service on each target indicator. The AI ​​effect tracking agent integrates the net effect, the magnitude of indicator changes, and the achieved business results to generate a structured and visualized service effect report. The business results include at least one of the following: the number of resumes submitted due to service use, the improved communication response rate, or the increased number of interview invitations.

7. A recruitment service recommendation device based on multiple AI agents, characterized in that, include: The acquisition module is used to acquire the first dynamic behavioral data of the target company and the target job postings it has published; The analysis module analyzes the first dynamic behavior data through an AI diagnostic agent to determine the target company's recruitment business entering the target node. The target node includes at least one of the following: exposure reduction node, screening efficiency reduction node, and communication response rate reduction node. The matching module is used to match at least one target recruitment value-added service corresponding to the target node from a pre-set service knowledge graph through an AI recommendation agent, and generate scenario-based service recommendation information based on the enterprise information of the target enterprise and the target recruitment value-added service. The recommendation module is used to recommend the service recommendation information to the target enterprise, and to obtain the second dynamic behavior data of the target enterprise and the target position after the target enterprise uses the target recruitment value-added service; The comparison module is used to compare the first dynamic behavior data and the second dynamic behavior data through an AI effect tracking agent, and generate a service effect report based on the comparison results.

8. An electronic device, characterized in that, The electronic device is a terminal device or a server. The electronic device includes a processor and a memory storing computer program instructions. When the electronic device executes the computer program instructions, it implements the method as described in any one of claims 1-6.

9. A computer program product, characterized in that, It includes computer program instructions that, when executed, implement the method as described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that, It stores computer program instructions that, when executed, implement the method as described in any one of claims 1-6.

Images