Agent or workflow task configuration automatic generation system, method and device

Abstract

This application provides an automatic task configuration generation system, method, and device for intelligent agents or workflows, comprising: a scenario analysis module, used to obtain business requirements and scenario features based on a task generation request, and to determine a target template based on the business requirements and scenario features; a configuration generation module, used to determine development task configuration information corresponding to the target template based on the business requirements and scenario features; a knowledge base management module, used to select a target knowledge base based on the business requirements and scenario features; and a task generation module, used to generate a target intelligent agent or target workflow based on the development task configuration information and the target knowledge base. This enables rapid construction of business processes and improves development efficiency and quality.

Description

Technical Field

[0001] This application belongs to the field of computer technology, and specifically relates to a system, method and device for automatically generating task configurations for intelligent agents or workflows. Background Technology

[0002] In current AI agent or workflow development processes, the implementation of business logic primarily relies on manual design and coding. This process typically involves a deep understanding of business requirements, as well as meticulous planning and coding of specific implementation details. High levels of human involvement are not only time-consuming and labor-intensive, but also prone to logical errors or inefficiencies in complex business scenarios. Summary of the Invention

[0003] This application provides a system, method, and device for automatically generating task configurations for intelligent agents or workflows, which addresses the problems of low efficiency and error-proneness in manually designing and implementing business logic during current development processes, enabling rapid construction of business processes and improving development efficiency and accuracy.

[0004] This application provides an automatic task configuration generation system for intelligent agents or workflows, including: The scenario analysis module is used to obtain business requirements and scenario characteristics based on the task generation request, and to determine the target template based on the business requirements and scenario characteristics; A configuration generation module is used to determine the development task configuration information corresponding to the target template based on the business requirements and the scenario characteristics. The knowledge base management module is used to select the target knowledge base based on the business requirements and the scenario characteristics. The task generation module is used to generate a target intelligent agent or a target workflow based on the development task configuration information and the target knowledge base.

[0005] According to the task configuration automatic generation system for intelligent agents or workflows provided in this application, the scenario analysis module includes a business requirement parsing submodule and a scenario feature extraction submodule. The requirement parsing submodule is used to convert the intelligent agent generation request into a structured task description, and generate the business requirement based on the structured task description and the task generation request. The scenario feature extraction submodule is used to determine the business scenario corresponding to the task generation request, and to determine the target scenario features corresponding to the business scenario. The target scenario features include any one or more of the following: process complexity, data processing requirements, knowledge base dependency, reasoning ability requirements, and interaction experience requirements.

[0006] According to the task configuration automatic generation system for intelligent agents or workflows provided in this application, the scene analysis module further includes a domain knowledge base generation submodule. This submodule is used to: identify professional concepts in training texts and establish a domain lexicon; generate a domain knowledge graph based on the lexicon, the domain knowledge graph including the concepts, attributes, and relationships between domain terms; obtain semantic associations and dependencies between different domain terms; and generate a domain knowledge base based on the semantic associations, dependencies, and the domain knowledge graph, the domain knowledge base being used to provide domain knowledge interpretation methods.

[0007] According to the task configuration automatic generation system for intelligent agents or workflows provided in this application, the development task configuration information includes workflow and auxiliary configuration information. The configuration generation module includes a workflow configuration generation submodule and an auxiliary configuration generation submodule. The workflow configuration generation submodule is used to generate the workflow according to the target business requirements and the target scenario characteristics. The auxiliary configuration generation submodule is used to generate auxiliary configuration information according to the target scenario characteristics. The auxiliary configuration information includes any one or more of the following: model parameter configuration information, prompt word template configuration information, knowledge base configuration information, database configuration information, and related plugin configuration information.

[0008] According to the task configuration automatic generation system for intelligent agents or workflows provided in this application, the knowledge base management module includes a knowledge base selection submodule; the knowledge base selection submodule is used to select a target knowledge base according to the business requirements and the scenario characteristics, and to perform quality assessment on the target knowledge base and / or generate a collaborative optimization strategy for multiple target knowledge bases.

[0009] According to the task configuration automatic generation system for intelligent agents or workflows provided in this application, the knowledge base management module further includes a retrieval strategy generation submodule and a knowledge base optimization submodule; the retrieval strategy generation submodule is used to determine the retrieval strategy of the target knowledge base and adjust the retrieval strategy and optimize retrieval parameters based on test results; the knowledge base optimization submodule is used to integrate multi-source knowledge to construct a knowledge graph, and to update the knowledge base and / or process conflicting data in the knowledge base.

[0010] The automatic task configuration generation system for intelligent agents or workflows provided in this application further includes a quality assurance module. This quality assurance module comprises a configuration verification submodule, a performance evaluation submodule, and an automatic correction submodule. The configuration verification submodule checks the validity of configuration parameters, the integrity of dependencies between configuration items, the security of configuration content, and / or the consistency between multiple configuration items based on the development task configuration information. The performance evaluation submodule evaluates the response time, resource utilization, and / or load capacity of the target intelligent agent or target workflow based on the test results. The automatic correction submodule generates error diagnosis reports and corresponding correction schemes based on the test results, and adjusts configuration parameters based on the performance evaluation results.

[0011] According to the task configuration automatic generation system for intelligent agents or workflows provided in this application, the intelligent agent or workflow generation system further includes a monitoring module; the monitoring module is used to manage the version of the development task configuration information, collect the performance indicators and abnormal information of the target intelligent agent or the target workflow, record the logs of the target intelligent agent or the target workflow, and / or analyze user feedback to determine user needs.

[0012] This application also provides a method for automatically generating task configurations for intelligent agents or workflows, including: Obtain business requirements and scenario characteristics based on the task generation request; Determine the target template based on the business requirements and the scenario characteristics; Determine the development task configuration information corresponding to the target template based on the business requirements and the scenario characteristics; Select the target knowledge base based on the business requirements and scenario characteristics; Generate a target intelligent agent or target workflow based on the development task configuration information and the target knowledge base.

[0013] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement an automatic task configuration generation method for any of the above-described intelligent agents or workflows.

[0014] This application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements an automatic task configuration generation method for any of the intelligent agents or workflows described above.

[0015] This application also provides a computer program product, including a computer program that, when executed by a processor, implements an automatic task configuration generation method for any of the intelligent agents or workflows described above.

[0016] This application provides a system, method, and device for automatically generating task configurations for intelligent agents or workflows. It acquires business requirements and scenario characteristics through a scenario analysis module and determines a target template. Then, a configuration generation module generates development task configuration information corresponding to the target template, and a knowledge base management module selects a target knowledge base. Finally, based on the development configuration information and the target knowledge base, it generates a target intelligent agent or target workflow. This enables rapid construction of business processes and improves the efficiency and accuracy of automatically generating task configurations for intelligent agents or workflows. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the structure of the automatic task configuration generation system for intelligent agents or workflows provided in this application; Figure 2 This is a schematic diagram of the configuration interface of the automatic task configuration generation system for intelligent agents or workflows provided in this application. Figure 3 This is one of the flowcharts illustrating the automatic task configuration generation method for intelligent agents or workflows provided in this application; Figure 4 This is the second flowchart illustrating the automatic task configuration generation method for intelligent agents or workflows provided in this application; Figure 5 This is a schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0020] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0021] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0022] Currently, the development of intelligent agent acquisition workflows still relies mainly on manual development, which results in low development efficiency, high error rate, and waste of human resources.

[0023] To address the aforementioned issues, embodiments of this application provide an automatic task configuration generation system, apparatus, and device for intelligent agents or workflows. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0024] Please see Figure 1 , Figure 1 This is a schematic diagram of the structure of the automatic task configuration generation system for intelligent agents or workflows provided in this application. The automatic task configuration generation system 10 for intelligent agents or workflows includes the following:

[0025] The scenario analysis module 101 is used to obtain business requirements and scenario features based on the task generation request, and to determine the target template based on the business requirements and scenario features.

[0026] The configuration generation module 102 is used to determine the development task configuration information corresponding to the target template based on the business requirements and the scenario characteristics.

[0027] The knowledge base management module 103 is used to select a target knowledge base based on the business requirements and the scenario characteristics.

[0028] The task generation module 104 is used to generate a target intelligent agent or a target workflow based on the development task configuration information and the target knowledge base.

[0029] As can be seen, this embodiment obtains business requirements and scenario characteristics through the scenario analysis module and determines the target template. Then, it generates the development task configuration information corresponding to the target template through the configuration generation module and selects the target knowledge base through the knowledge base management module. Finally, it generates the target intelligent agent or target workflow based on the development configuration information and the target knowledge base. This enables the rapid construction of business processes and improves the efficiency and accuracy of automatic generation of task configurations for intelligent agents or workflows.

[0030] In one possible embodiment, the scene analysis module 101 includes a business requirement parsing submodule and a scene feature extraction submodule.

[0031] The requirement parsing submodule is used to convert the agent generation request into a structured task description, and generate the business requirements based on the structured task description and the task generation request.

[0032] In practical implementation, when generating business requirements, a large model can be used for text understanding and structured extraction. Firstly, pre-defined prompt templates can guide the large model to extract core information such as key business objectives, functional requirements, and performance metrics from the task generation request. Simultaneously, the understanding of business requirements can be enhanced by leveraging professional terminology and expressions from different industry sectors within a domain knowledge graph. During structured extraction, unstructured text can be converted into a structured representation containing dimensions such as objectives, constraints, performance metrics, and data requirements based on a structured requirement representation model. This results in a structured requirement understanding report that includes key information extraction, dependency analysis, and risk assessment.

[0033] In practical implementation, when generating business requirements, semantic understanding and context analysis can be used to identify gaps in information, logical contradictions, or ambiguities in the description of the task generation request. These gaps can then be supplemented through multiple interactions with the user, guiding them to refine the requirement information. Based on this complete information, a comprehensive structured requirement understanding report is generated, and the business requirements are determined based on the report's content. Furthermore, suggestions for improving the requirement description can be provided based on historical experience.

[0034] The scenario feature extraction submodule is used to determine the business scenario corresponding to the task generation request, and to determine the target scenario features corresponding to the business scenario. The target scenario features include any one or more of the following: process complexity, data processing requirements, knowledge base dependency, reasoning ability requirements, and interaction experience requirements.

[0035] In practice, when assessing process complexity, we can first identify key steps in the business process through semantic understanding, construct a dependency graph between steps, and then analyze the number of nodes, nesting depth, and number of branch paths. Simultaneously, based on dependency analysis, we can discover parallelizable task paths, and score the process complexity based on factors such as the number of steps, branches, and parallelism.

[0036] In practice, data processing requirements analysis may include analyzing the format and structural characteristics of input data, assessing the scale and timeliness requirements of data processing, evaluating data integrity and accuracy, identifying data cleaning and transformation requirements, analyzing data sensitivity and corresponding necessary security measures, analyzing the requirements for data persistence and rapid retrieval, and / or analyzing data exchange requirements with other systems.

[0037] In practice, when performing knowledge base dependency analysis, it can include analyzing the scope and depth of professional knowledge required for the task, analyzing the required frequency of knowledge updates, identifying the requirements for knowledge relevance and consistency, analyzing the efficiency and accuracy requirements for knowledge retrieval, analyzing the need for reasoning and decision-making based on knowledge, and / or analyzing the need for dynamic expansion and evolution of the knowledge base.

[0038] In practice, the evaluation of reasoning ability can include analyzing the logical reasoning complexity of the task, the processing requirements for analyzing fuzzy information and probabilistic reasoning, analyzing scenarios that require multiple rounds of reasoning to reach a conclusion, analyzing the requirements for reasoning speed and response time, analyzing the requirements for interpreting the reasoning process and results, and / or evaluating the requirements for reasoning accuracy.

[0039] In practice, interactive experience requirements analysis includes analyzing the frequency and patterns of user interactions, assessing users' expectations for system response speed, analyzing the complexity of the interactive interface, analyzing users' personalized customization needs, and / or analyzing continuous dialogue and context maintenance needs.

[0040] In its specific implementation, the scenario analysis module 101 further includes a template matching submodule, which selects a target template from existing templates that matches the current business requirements and scenario characteristics. When performing template matching, a single template can be selected, or a combination of templates can be determined based on requirements. Determining the template combination may include evaluating the interface compatibility and functional complementarity between templates to analyze the possibility of combining multiple templates. It may also include evaluating the performance impact of combining multiple templates, including the resource consumption and response time of the combined solution. Furthermore, it may include evaluating the workload and technical difficulty of implementing the combined solution, and predicting the implementation effect of the combined solution. Based on the above, an optimal combined solution suggestion is generated through a multi-objective optimization algorithm, along with the implementation strategy for that combined solution. Several alternative combined solutions can also be generated.

[0041] As can be seen, in this embodiment, based on multi-dimensional business requirement analysis and scenario feature extraction, a deeper understanding of development tasks can be achieved, improving the accuracy of subsequent configuration information determination and increasing development efficiency.

[0042] In one possible embodiment, the scene analysis module 101 further includes a domain knowledge base generation submodule, which is used to: identify professional concepts in the training text and establish a domain lexicon; generate a domain knowledge graph based on the domain lexicon, wherein the domain knowledge graph includes the concepts, attributes, and relationships between domain words; obtain semantic associations and dependencies between different domain words; and generate a domain knowledge base based on the semantic associations, dependencies, and the domain knowledge graph, wherein the domain knowledge base is used to provide a way to explain domain knowledge.

[0043] In business requirements analysis, user task development needs can be analyzed based on the professional terms and concept explanations included in the domain knowledge base. Therefore, a domain graph can be constructed using a domain knowledge base generation submodule to generate the domain knowledge base. Professional concepts in the domain lexicon include, but are not limited to, professional terms, industry probabilities, and business entities. When obtaining semantic relationships and dependencies between different domain terms, analysis can be performed based on the knowledge transfer capabilities of a large model. After generating the domain knowledge base, new domain knowledge can be continuously collected and integrated, dynamically updating and optimizing the domain knowledge base.

[0044] As can be seen, in this embodiment, establishing a domain knowledge base allows for the analysis of development requests based on professional content within the domain knowledge base, which can improve the accuracy of business requirement analysis.

[0045] In one possible embodiment, the development task configuration information includes workflow and auxiliary configuration information, and the configuration generation module 102 includes a workflow configuration generation submodule and an auxiliary configuration generation submodule.

[0046] The workflow configuration generation submodule is used to generate the workflow based on the target business requirements and the target scenario characteristics. Specifically, different target templates may include different development task configuration information. For example... Figure 2 The development configuration information displayed in the template includes workflow and auxiliary configuration information. Figure 2 The auxiliary configuration information includes model parameter configuration information, that is... Figure 2 The "AI configuration" shown, as well as the knowledge base configuration information, are... Figure 2 The "knowledge base association" shown, as well as the database configuration information, are... Figure 2The "database association" shown, along with the related plugin configuration information, is... Figure 2 The "plugin" shown, and the prompt word template configuration information, namely Figure 2 The “opening remarks” shown.

[0047] The workflow generation process includes the following steps: First, node selection is performed based on the target business requirements and target scenario characteristics. When selecting nodes, an appropriate node type can be chosen based on the task description and context. This type includes data nodes, function nodes, interaction nodes, and control nodes. Then, based on the node usage model from historical intelligent agents or workflow generation cases, the currently selected and configured nodes are optimized, and auxiliary nodes are added, such as necessary auxiliary nodes for data preprocessing and format conversion. After selecting and configuring the nodes to obtain the node combination, the rationality of the node combination can be evaluated, checking the compatibility between nodes and the feasibility of data flow, etc. Based on the evaluation results, the node configuration parameters are adjusted to obtain the final node configuration scheme.

[0048] Then, based on the selected nodes, dependencies between nodes are constructed. This includes analyzing node input and output characteristics, constructing a data flow graph, designing a reasonable execution order and control structure based on business logic to construct control flow relationships, identifying node paths without dependencies, designing reasonable parallel execution schemes, eliminating circular dependencies between nodes in the execution scheme, evaluating the degree of coupling between nodes, and optimizing the overall architecture.

[0049] After establishing dependencies, data flow mapping is required. This involves analyzing the data characteristics transmitted by nodes to ensure data type matching, adding necessary data transformation nodes to ensure smooth data flow, optimizing the data transmission scheme to reduce unnecessary data transformations, monitoring the data flow process to promptly detect and handle anomalies, designing appropriate caching mechanisms for frequently used data, and selecting suitable compression algorithms for updating data characteristics.

[0050] The workflow is derived from node selection, dependency construction, and data flow mapping. Furthermore, after obtaining the workflow, its branching logic can be optimized. This optimization includes optimizing conditional logic, optimizing branch execution paths, adding exception handling branches, and merging branches with similar functions. Branch execution can also be monitored to continuously optimize and adjust branch selection strategies. Additionally, workflow performance can be optimized, such as analyzing node execution efficiency, resource allocation, caching strategies, concurrency control, and load balancing, and developing targeted optimization plans.

[0051] The auxiliary configuration generation submodule is used to generate auxiliary configuration information based on the target scene features. The auxiliary configuration information includes any one or more of the following: model parameter configuration information, prompt word template configuration information, knowledge base configuration information, database configuration information, and related plugin configuration information.

[0052] In practice, model parameter configuration information is used to optimize model parameters, such as optimizing the configuration of sampling temperature parameters, task information volume and system resource window size, model quantization parameters, and computing resource allocation.

[0053] The prompt template configuration information is used to optimize prompt templates. For example, it includes configuring targeted prompts based on the characteristics of business scenarios, converting business rules and quality requirements into prompt constraints, optimizing the arrangement of prompt components, designing prompt templates for multi-turn dialogues, and adjusting the prompt style according to user characteristics and preferences.

[0054] The knowledge base configuration information includes setting a knowledge base relevance threshold to dynamically adjust the search matching threshold, as well as parameters for optimizing search strategies, configuring dynamic optimization weights, and configuring recall ratios.

[0055] The relevant plugin configuration information includes determining the plugins required for the task, generating the plugin call chain, identifying the dependencies between plugins, and configurations such as plugin call frequency, exception handling mechanism, and timeout policy.

[0056] As can be seen, in this embodiment, automatically generating workflow and auxiliary configuration information can improve the efficiency and accuracy of task development.

[0057] In one possible embodiment, the knowledge base management module 103 includes a knowledge base selection submodule; the knowledge base selection submodule is used to select a target knowledge base according to the business requirements and the scenario characteristics, and to perform quality assessment on the target knowledge base and / or generate a collaborative optimization strategy for multiple target knowledge bases.

[0058] The process of rotating the target knowledge base includes first analyzing the matching degree between the knowledge base content and the current business needs and scenario characteristics through semantic understanding, calculating the relevance score of the knowledge base, evaluating the coverage of the knowledge base in the current domain, and evaluating the timeliness of the knowledge base. Based on the relevance score, coverage, and timeliness, the target knowledge base most relevant to the current business needs and scenario characteristics is selected.

[0059] When assessing the quality of a knowledge base, the evaluation includes analyzing the accuracy and completeness of the knowledge, assessing its structure and standardization, and checking the consistency of knowledge within the base. A comprehensive assessment of the knowledge base's quality is then conducted. Collaborative optimization refers to establishing a collaborative working mechanism for multiple knowledge bases, clarifying the division of labor and cooperation methods for each base, identifying and resolving overlaps and inconsistencies between different knowledge bases, and determining the knowledge retrieval sequence for each base based on knowledge quality and relevance.

[0060] As can be seen, in this embodiment, selecting the target knowledge base through the knowledge base selection submodule can improve task development efficiency.

[0061] In one possible embodiment, the knowledge base management module 103 further includes a retrieval strategy generation submodule and a knowledge base optimization submodule.

[0062] The retrieval strategy generation submodule is used to determine the retrieval strategy for the target knowledge base and to adjust the retrieval strategy and optimize retrieval parameters based on test results.

[0063] In practice, appropriate retrieval algorithms, such as vector retrieval, full-text retrieval, or hybrid retrieval, can be selected based on the type and structural characteristics of the knowledge content. Furthermore, after determining the retrieval strategy, performance metrics can be collected based on test results to adjust the strategy and find a balance between recall and precision, thereby improving retrieval performance.

[0064] The knowledge base optimization submodule is used to integrate multi-source knowledge to construct a knowledge graph, update the knowledge base, and / or process conflicting data in the knowledge base.

[0065] The multi-source knowledge integration process includes data cleaning, standardization, and conflict resolution, resulting in a knowledge integration workflow. It also involves format conversion of multi-source knowledge to achieve standardization and consistency. During integration, natural language processing (NLP) is used to analyze the semantics of the knowledge, ensuring semantic accuracy and coherence. Based on the integrated knowledge, a multi-level knowledge relationship network is established, constructing a knowledge graph. This knowledge graph supports queries involving knowledge reasoning and association analysis.

[0066] When processing conflicting data, semantic analysis and rule engines can be used to identify contradictions in the knowledge base. The severity and resolution priority of these contradictions can be assessed based on factors such as timeliness, credibility, and frequency of use. Then, based on preset rules and historical processing experience, the contradictions are resolved according to priority and severity. Simultaneously, the knowledge base optimization submodule can also dynamically update and manage the version of the knowledge base.

[0067] As can be seen, in this embodiment, the knowledge base management module can also optimize the retrieval strategy and the knowledge base, so that the target task body generated based on the knowledge base can be more accurate.

[0068] In one possible embodiment, the intelligent agent or workflow generation system 10 further includes a quality assurance module, which includes a configuration verification submodule, a performance evaluation submodule, and an automatic correction submodule.

[0069] The configuration verification submodule is used to check the validity of configuration parameters, the integrity of dependencies between configuration items, the security of configuration content, and / or the consistency between multiple configuration items based on the development task configuration information.

[0070] In practice, parameter validity verification includes checking whether the data type, value range, and format specifications of configuration parameters meet the requirements. Dependency integrity checks include verifying the availability of necessary components and services to prevent operational anomalies caused by missing components. Security checks include identifying security vulnerabilities, such as risks related to permission settings and sensitive information leakage. Consistency checks between configuration items include verifying the logical relationships between multiple configuration items to ensure overall consistency and rationality of the configuration.

[0071] The performance evaluation submodule is used to evaluate the response time, resource utilization, and / or load capacity of the target agent or the target workflow based on the test results of the target agent or the target workflow.

[0072] In practice, resource utilization assessment includes evaluating resources such as CPU, memory, and network bandwidth. Load capacity assessment evaluates the system's performance under different load conditions to determine optimal operating parameters. It can also predict and locate resource bottlenecks and provide optimization suggestions.

[0073] The automatic correction submodule is used to generate an error diagnosis report and corresponding correction scheme based on the test results, and to adjust configuration parameters based on the performance evaluation results.

[0074] In practice, when generating error diagnosis reports, the error type and root cause can be identified based on system logs, error stacks, and runtime status. Then, combined with system configuration, operating environment, and historical data, a complete error scenario profile is constructed. Based on this profile, a structured error diagnosis report is generated, including error location, scope of impact, and severity. Furthermore, corrective solutions corresponding to the errors in the error diagnosis report can be generated based on historical processing experience and practices. Additionally, error patterns and risk factors can be analyzed to provide suggestions for system optimization and predictive measures.

[0075] The adjustment of configuration parameters includes adjusting key configuration parameters based on performance indicators, and adjusting relevant configuration information such as task scheduling, cache measurement, concurrency control, and allocation of computing and storage resources based on load conditions and resource utilization.

[0076] As can be seen, in this embodiment, after configuration, performance evaluation testing can enable the system to reach its optimal operating state and improve the quality of the generated target intelligent agent or target workflow.

[0077] In one possible embodiment, the agent or workflow generation system 10 further includes a monitoring module; the monitoring module is used to manage the version of the development task configuration information, collect the performance indicators and anomaly information of the target agent or the target workflow, record the logs of the target agent or the target workflow, and / or analyze user feedback to determine user needs.

[0078] As can be seen, in this embodiment, monitoring the version and anomalies of the target intelligent agent or target workflow based on the monitoring module can improve the reliability and stability of operation.

[0079] Please see Figure 3 This application provides a method for automatically generating task configurations for intelligent agents or workflows, including the following steps.

[0080] S301: Obtain business requirements and scenario characteristics based on the task generation request.

[0081] Specifically, business requirements and scenario characteristics can be obtained through a scenario analysis module. In practice, obtaining business requirements may include: performing text analysis on the agent-generated request to obtain preliminary business requirements; performing requirement completeness analysis based on the preliminary business requirements and the agent-generated request to obtain requirements to be mined; engaging in dialogue with the user based on the requirements to be mined to obtain supplementary requirement information; and obtaining the final business requirements based on the supplementary requirement information and the preliminary business requirements.

[0082] S302, Determine the target template based on the business requirements and the scenario characteristics.

[0083] Specifically, the target template can be determined based on the template matching submodule within the scenario analysis module. Different templates can correspond to different development task configuration information.

[0084] S303, determine the development task configuration information corresponding to the target template based on the business requirements and the scenario characteristics.

[0085] Specifically, the development task configuration information can be determined based on the configuration generation module. This development task configuration information may include workflow and auxiliary task configuration information.

[0086] S304. Select the target knowledge base based on the business requirements and the scenario characteristics.

[0087] S305, Generate a target intelligent agent or a target workflow based on the development task configuration information and the target knowledge base.

[0088] As can be seen, in this embodiment, a target template is determined by acquiring business requirements and scenario characteristics. Then, the development task configuration information corresponding to the target template is determined, and a target knowledge base is selected based on business requirements and scenario characteristics. Finally, a target intelligent agent or target workflow is generated based on the development task configuration information and the target knowledge base. This enables rapid construction of business processes and improves the efficiency and accuracy of automatic generation of task configurations for intelligent agents or workflows.

[0089] Please see Figure 4 Taking the generation of an intelligent agent as an example, this paper details the generation process of the target intelligent agent in this application. First, in the requirement input stage, the system receives and parses the user's business requirements. The system can intelligently parse the requirement document provided by the user using natural language processing technology, extracting core elements such as key business information, technical constraints, and performance requirements. Simultaneously, by combining a historical case library and a domain knowledge graph, the system can accurately identify the characteristic attributes of the target application scenario, including dimensions such as business domain, interaction mode, and processing logic, providing a foundation for subsequent template matching.

[0090] Then, in the template selection phase, based on the scenario features extracted in the requirements phase, the system can deeply analyze the technical characteristics, business models, and application requirements of the target scenario. Through a multi-dimensional feature matching algorithm, it selects the most suitable target template from a pre-built template library. The system can also intelligently generate targeted configuration suggestions, including system parameters, processing strategies, and optimization solutions, by combining historical application performance data, ensuring that the template can meet user needs to the greatest extent possible.

[0091] Then, the configuration generation phase begins. After selecting the target template, the system will automatically generate a complete configuration scheme. First, based on the scenario-based testing requirements generated by the intelligent test cases, the core workflow configuration is generated, defining the task processing flow, state transition rules, and triggering conditions. Then, based on the results of error diagnosis and analysis, the dialogue strategy, knowledge retrieval method, and response mode of the intelligent assistant are configured. Simultaneously, considering the requirements of adaptive optimization, appropriate knowledge bases and update strategies are configured. Finally, the functional configuration of necessary plugins is completed, including detailed parameter settings for modules such as data processing and external interface integration, and optimizations are made based on performance test results.

[0092] Then comes the optimization and verification phase. The generated configuration scheme needs to undergo comprehensive verification and optimization. The system first automatically verifies the legality and completeness of the configuration to ensure that all necessary parameters are set correctly. Then, performance testing is conducted to evaluate the system's performance under various load conditions, including metrics such as response time, resource consumption, and concurrent processing capabilities. Based on the test results, the system can automatically adjust and optimize relevant parameters and intelligently correct the configuration when necessary to ensure the system reaches its optimal operating state.

[0093] Finally, the deployment and release phase begins. After optimization, the system enters the deployment preparation phase. First, a configuration version management mechanism is established to record all configuration changes and support version tracking. An incremental update strategy is designed to achieve smooth upgrades and dynamic updates of configurations. Simultaneously, a robust rollback mechanism can be built to quickly restore to a stable version in the event of anomalies. Finally, the system can automatically generate a comprehensive monitoring solution, including performance indicator monitoring, anomaly detection, and alerting policies, to ensure the reliability and stability of system operation.

[0094] The automatic task configuration generation system for intelligent agents or workflows provided in this application is based on the above generation process, which can improve the efficiency and quality of automatic task configuration generation for intelligent agents or workflows, reduce manual configuration costs, and improve the availability and maintainability of intelligent agents or workflows.

[0095] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of the electronic device provided in this application. For example... Figure 5 As shown, the electronic device may include a processor 510, a communications interface 520, a memory 530, and a communication bus 540, wherein the processor 510, communications interface 520, and memory 530 communicate with each other via the communication bus 540. The processor 510 can invoke logical instructions in the memory 530 to execute an automatic task configuration generation method for an intelligent agent or workflow. This method includes: obtaining business requirements and scenario features based on a task generation request; determining a target template based on the business requirements and scenario features; determining development task configuration information corresponding to the target template based on the business requirements and scenario features; selecting a target knowledge base based on the business requirements and scenario features; and generating a target intelligent agent or target workflow based on the development task configuration information and the target knowledge base.

[0096] Furthermore, the logical instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0097] On the other hand, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements an automatic task configuration generation method for intelligent agents or workflows provided by the above methods. The method includes: obtaining business requirements and scenario features based on a task generation request; determining a target template based on the business requirements and scenario features; determining development task configuration information corresponding to the target template based on the business requirements and scenario features; selecting a target knowledge base based on the business requirements and scenario features; and generating a target intelligent agent or target workflow based on the development task configuration information and the target knowledge base.

[0098] In another aspect, this application also provides a computer program product, including a computer program that, when executed by a processor, implements an automatic task configuration generation method for any of the intelligent agents or workflows described above. This method includes: obtaining business requirements and scenario features based on a task generation request; determining a target template based on the business requirements and scenario features; determining development task configuration information corresponding to the target template based on the business requirements and scenario features; selecting a target knowledge base based on the business requirements and scenario features; and generating a target intelligent agent or target workflow based on the development task configuration information and the target knowledge base.

[0099] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0100] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0101] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automatic task configuration generation system for intelligent agents or workflows, characterized in that, include: The scenario analysis module is used to obtain business requirements and scenario characteristics based on the task generation request, and to determine the target template based on the business requirements and scenario characteristics; A configuration generation module is used to determine the development task configuration information corresponding to the target template based on the business requirements and the scenario characteristics. The knowledge base management module is used to select the target knowledge base based on the business requirements and the scenario characteristics. The task generation module is used to generate a target intelligent agent or a target workflow based on the development task configuration information and the target knowledge base.

2. The automatic task configuration generation system for intelligent agents or workflows according to claim 1, characterized in that, The scenario analysis module includes a business requirement parsing submodule and a scenario feature extraction submodule; The requirement parsing submodule is used to convert the agent generation request into a structured task description, and generate the business requirements based on the structured task description and the task generation request. The scene feature extraction submodule is used to determine the business scene corresponding to the task generation request, and to determine the target scene features corresponding to the business scene. The target scene features include any one or more of the following: Process complexity, data processing requirements, knowledge base dependence, reasoning ability requirements, and interactive experience requirements.

3. The automatic task configuration generation system for intelligent agents or workflows according to claim 2, characterized in that, The scenario analysis module further includes a domain knowledge base generation submodule, which is used for: Identify specialized concepts in the training text and build a domain-specific vocabulary; A domain knowledge graph is generated based on the domain vocabulary, and the domain knowledge graph includes the concepts, attributes, and relationships between the domain terms. Obtain semantic relationships and dependencies between words in different domains; A domain knowledge base is generated based on the semantic relationships, the dependencies, and the domain knowledge graph. The domain knowledge base is used to provide a way to interpret domain knowledge.

4. The automatic task configuration generation system for intelligent agents or workflows according to claim 2, characterized in that, The development task configuration information includes workflow and auxiliary configuration information, and the configuration generation module includes a workflow configuration generation submodule and an auxiliary configuration generation submodule. The workflow configuration generation submodule is used to generate the workflow based on the target business requirements and the target scenario characteristics. The auxiliary configuration generation submodule is used to generate auxiliary configuration information based on the target scene features, wherein the auxiliary configuration information includes any one or more of the following: Model parameter configuration information, prompt word template configuration information, knowledge base configuration information, database configuration information, and related plugin configuration information.

5. The automatic task configuration generation system for intelligent agents or workflows according to claim 2, characterized in that, The knowledge base management module includes a knowledge base selection sub-module; The knowledge base selection submodule is used to select a target knowledge base based on the business requirements and the scenario characteristics, and to perform quality assessment on the target knowledge base and / or generate a collaborative optimization strategy for multiple target knowledge bases.

6. The automatic task configuration generation system for intelligent agents or workflows according to claim 5, characterized in that, The knowledge base management module also includes a retrieval strategy generation submodule and a knowledge base optimization submodule; The retrieval strategy generation submodule is used to determine the retrieval strategy for the target knowledge base and to adjust the retrieval strategy and optimize retrieval parameters based on test results; The knowledge base optimization submodule is used to integrate multi-source knowledge to construct a knowledge graph, update the knowledge base, and / or process conflicting data in the knowledge base.

7. The automatic task configuration generation system for intelligent agents or workflows according to claim 1, characterized in that, The intelligent agent or workflow generation system also includes a quality assurance module, which includes a configuration verification submodule, a performance evaluation submodule, and an automatic correction submodule. The configuration verification submodule is used to check the validity of configuration parameters, the integrity of dependencies between configuration items, the security of configuration content, and / or the consistency between multiple configuration items based on the development task configuration information. The performance evaluation submodule is used to evaluate the response time, resource utilization, and / or load capacity of the target agent or the target workflow based on the test results of the target agent or the target workflow. The automatic correction submodule is used to generate an error diagnosis report and corresponding correction scheme based on the test results, and to adjust configuration parameters based on the performance evaluation results.

8. The automatic task configuration generation system for intelligent agents or workflows according to claim 1, characterized in that, The intelligent agent or workflow generation system also includes a monitoring module; The monitoring module is used to manage the version of the development task configuration information, collect the performance indicators and anomaly information of the target agent or the target workflow, record the logs of the target agent or the target workflow, and / or analyze user feedback to determine user needs.

9. A method for automatically generating task configurations for intelligent agents or workflows, characterized in that, include: Obtain business requirements and scenario characteristics based on the task generation request; Determine the target template based on the business requirements and the scenario characteristics; Determine the development task configuration information corresponding to the target template based on the business requirements and the scenario characteristics; Select the target knowledge base based on the business requirements and scenario characteristics; Generate a target intelligent agent or target workflow based on the development task configuration information and the target knowledge base.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the automatic task configuration generation method for intelligent agents or workflows as described in claim 9.

Images