Movatterモバイル変換


[0]ホーム

URL:


CN117634865B - Workflow creation method, device, equipment and storage medium - Google Patents

Workflow creation method, device, equipment and storage medium
Download PDF

Info

Publication number
CN117634865B
CN117634865BCN202410102089.6ACN202410102089ACN117634865BCN 117634865 BCN117634865 BCN 117634865BCN 202410102089 ACN202410102089 ACN 202410102089ACN 117634865 BCN117634865 BCN 117634865B
Authority
CN
China
Prior art keywords
workflow
creation
information
workflow creation
response information
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202410102089.6A
Other languages
Chinese (zh)
Other versions
CN117634865A (en
Inventor
崔智军
王磊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Aspire Technologies Shenzhen Ltd
Original Assignee
Aspire Technologies Shenzhen Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Aspire Technologies Shenzhen LtdfiledCriticalAspire Technologies Shenzhen Ltd
Priority to CN202410102089.6ApriorityCriticalpatent/CN117634865B/en
Publication of CN117634865ApublicationCriticalpatent/CN117634865A/en
Application grantedgrantedCritical
Publication of CN117634865BpublicationCriticalpatent/CN117634865B/en
Activelegal-statusCriticalCurrent
Anticipated expirationlegal-statusCritical

Links

Classifications

Landscapes

Abstract

Translated fromChinese

本发明公开了一种工作流创建方法、装置、设备及存储介质,涉及工作流管理技术领域,包括:获取目标用户对应的工作流创建信息;基于工作流创建信息、预设流程知识库以及预设流程知识图谱,确定第一应答信息以及工作流类型;将第一应答信息输入至业务流程问答模型,得到业务流程问答模型输出的第二应答信息;将第二应答信息以及工作流类型发送至工作流创建引擎,以供工作流创建引擎创建与工作流创建信息对应的工作流。本发明通过语聊机器人对工作流创建信息进行应答及确认,并通过工作流创建引擎的自动创建功能实现工作流的自动化生成和智能填写等,从而显著提高工作效率和准确性,同时减少了人为错误,有效降低运营成本,减轻用户操作负担。

The present invention discloses a workflow creation method, device, equipment and storage medium, and relates to the technical field of workflow management, including: obtaining workflow creation information corresponding to a target user; determining first response information and workflow type based on workflow creation information, a preset process knowledge base and a preset process knowledge graph; inputting the first response information into a business process question and answer model to obtain second response information output by the business process question and answer model; sending the second response information and workflow type to a workflow creation engine, so that the workflow creation engine can create a workflow corresponding to the workflow creation information. The present invention responds to and confirms the workflow creation information through a voice chat robot, and realizes automatic generation and intelligent filling of workflows through the automatic creation function of the workflow creation engine, thereby significantly improving work efficiency and accuracy, while reducing human errors, effectively reducing operating costs, and alleviating the user's operating burden.

Description

Workflow creation method, device, equipment and storage medium
Technical Field
The present invention relates to the field of workflow management technologies, and in particular, to a workflow creation method, device, equipment, and storage medium.
Background
For a workflow system, whether an open source engine such as Activiti, flowable or Camunda is used, a business process such as creating a process and filling a form is usually required to be manually operated, when a process design is changed, a system administrator is also required to be relied on to manually process, an administrator is also required to draw a flow chart, a configuration form and the like, so that the process design efficiency is low, and operations such as filling errors can also occur in the manual process.
Meanwhile, various processes in the workflow system are complicated, most users are not clear about how to select the processes, so that a great deal of time is spent on training by developers and writing an operation manual to guide user bill of lading, the bill of lading efficiency is low, and the operation cost is greatly increased. The user can not fill out the field meaning because of unclear field meaning when the bill is presented, so that a large amount of time is required to communicate with the filling form, the answering and communication cost is high during the period, various unstructured related documents such as FAQ are available, the related documents are lack of iterative update due to the change of personnel during the time lapse, the user consulting effect is poor, the user needs to find corresponding professional operation and maintenance colleagues to communicate when encountering problems, and the communication cost is increased.
Disclosure of Invention
Based on the above, it is necessary to provide a workflow creation method, device, equipment and storage medium for realizing automatic generation and intelligent filling of a workflow, thereby remarkably improving the working efficiency and accuracy, reducing human errors, effectively reducing the operation cost, reducing the operation burden of a user and improving the user experience.
A workflow creation method applied to a chat robot, comprising:
acquiring workflow creation information corresponding to a target user;
Determining first response information and workflow types based on the workflow creation information, a preset flow knowledge base and a preset flow knowledge map;
inputting the first response information into a business process question-answer model to obtain second response information output by the business process question-answer model;
And sending the second response information and the workflow type to a workflow creation engine so that the workflow creation engine creates a workflow corresponding to the workflow creation information.
According to the workflow creation method provided by the invention, the determining of the first response information and the workflow type based on the workflow creation information, the preset flow knowledge base and the preset flow knowledge map comprises the following steps:
Text segmentation is carried out on the workflow creation information, and the segmented workflow creation information is vectorized to obtain a workflow creation vector;
Inquiring and matching the workflow creation vector with a preset flow knowledge base to obtain the workflow type and initial response information;
And carrying out structuring treatment on the initial response information, and matching the structured initial response information with a preset flow knowledge graph to obtain the first response information.
According to the workflow creation method provided by the invention, before determining the first response information and the workflow type, the workflow creation method further comprises the steps of:
Acquiring business process knowledge;
Information extraction and knowledge fusion are carried out on the business process knowledge to obtain workflow structured knowledge, and quality evaluation is carried out on the workflow structured knowledge;
and if the quality evaluation is passed, storing the workflow structured knowledge to the preset flow knowledge graph.
According to the workflow creation method provided by the invention, before the first response information is input into the business process question-answer model to obtain the second response information output by the business process question-answer model, the workflow creation method further comprises the following steps:
Obtaining a plurality of groups of flow demand samples, and dividing each flow demand sample into a model training set, a model verification set and a model test set;
And performing iterative training and model fine tuning on an initial business process question-answer model based on the model training set, the model verification set and the model test set to obtain the business process question-answer model.
According to the workflow creation method provided by the invention, the workflow creation information corresponding to the target user is obtained, and the workflow creation method comprises the following steps:
Acquiring workflow creation requirements input by a target user;
judging whether history creation information is matched or not based on the workflow creation requirement;
If yes, performing association combination on the matched history creation information and the workflow creation requirement to generate workflow creation information;
if not, generating the workflow creation information based on the workflow creation requirement.
A workflow creation method applied to a workflow creation engine, comprising:
receiving second response information sent by the chat robot and workflow type;
Determining a workflow creation module corresponding to the workflow based on the workflow type;
analyzing the second response information to obtain workflow entities, entity attributes and entity association conditions;
And inputting the workflow entity, the entity attribute and the entity association condition into the workflow creation module so as to enable the workflow creation module to create the workflow.
According to the workflow creation method provided by the present invention, after inputting the workflow entity, the entity attribute and the entity association condition into the workflow creation module, the workflow creation module creates the workflow, the workflow creation method further includes:
Acquiring the creation success time and the checking mode of the workflow;
generating creation success information based on the workflow type, the creation success time and the viewing mode;
Pushing the creation success information to a target user so as to remind the target user to check.
A workflow creation apparatus comprising:
The voice chat robot is used for acquiring workflow creation information input by a user, determining first response information and workflow type based on the workflow creation information and a preset flow knowledge graph, inputting the first response information into a business flow question-answer model, and obtaining second response information output by the business flow question-answer model;
The workflow creation engine is used for receiving second response information and workflow types sent by the chat robot, determining a workflow creation module corresponding to the workflow based on the workflow types, analyzing the second response information to obtain workflow entities, entity attributes and entity association conditions, and inputting the workflow entities, the entity attributes and the entity association conditions into the workflow creation module so as to enable the workflow creation module to create the workflow.
An electronic device comprising a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, the processor implementing the workflow creation method described above when executing the computer readable instructions.
One or more readable storage media storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform a workflow creation method as described above.
According to the workflow creation method, device, equipment and storage medium, the workflow creation information corresponding to the target user is obtained, and the first response information and the workflow type are determined based on the workflow creation information, the preset flow knowledge base and the preset flow knowledge map, and then the first response information is input into the business flow question-answer model to obtain the second response information output by the business flow question-answer model, and then the second response information and the workflow type are sent to the workflow creation engine to enable the workflow creation engine to create the workflow corresponding to the workflow creation information, so that automatic generation, intelligent filling and the like of the workflow are achieved, work efficiency and accuracy are remarkably improved, human errors are reduced, operation cost is effectively reduced, user operation burden is reduced, and user experience is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments of the present invention will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow diagram of a workflow creation method provided by the present invention;
FIG. 2 is a second flow chart of the workflow creation method according to the present invention;
FIG. 3 is a schematic diagram of a workflow creation apparatus provided by the present invention;
FIG. 4 is a flow chart of a partial construction of a preset flow knowledge base provided by the present invention;
FIG. 5 is a partial functional implementation diagram of a workflow creation system provided by the present invention;
FIG. 6 is a partial construction flow chart of a preset flow knowledge graph provided by the invention;
FIG. 7 is a diagram of an exemplary information structure provided by the present invention;
FIG. 8 is a diagram of an example implementation of the function of job ticket generation provided by the present invention;
FIG. 9 is a diagram of an exemplary implementation of the function of form filling provided by the present invention;
FIG. 10 is a functional implementation example diagram of a flow design change provided by the present invention;
Fig. 11 is a schematic structural diagram of an electronic device provided by the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The terminology used in the one or more embodiments of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of the invention. As used in one or more embodiments of the invention, the singular forms "a," "an," "the," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used in one or more embodiments of the present invention refers to and encompasses any or all possible combinations of one or more of the associated listed items.
It should be understood that, although the terms first, second, etc. may be used in one or more embodiments of the invention to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, a first may also be referred to as a second, and similarly, a second may also be referred to as a first, without departing from the scope of one or more embodiments of the invention. Depending on the context of the user, the word "if" as used herein may be interpreted the method is "in the environment. Or" when.
Fig. 1 is a schematic flow chart of a workflow creation method provided by the invention. As shown in fig. 1, the workflow creation method is applied to a chat robot, and includes:
Step S11, workflow creation information corresponding to a target user is obtained;
it should be noted that, the target user refers to a user who needs to create a workflow, and the workflow creation information refers to creation information required by specific requirements and expectations of the target user in the process of creating the workflow, including, but not limited to, business process descriptions (such as steps, stages, related workflow entities, entity attributes, entity association situations, etc.), user interaction requirements (such as form elements, user authority management, etc.), history creation information, custom rules (such as approval conditions, flow branches, automated decisions, etc.), structured data such as flow monitoring and reporting, business knowledge questions and answers, and unstructured data.
Specifically, a workflow creation requirement input by a target user is acquired, and whether history creation information is matched is further judged based on the workflow creation requirement, so that if yes, the matched history creation information and the workflow creation requirement are combined in a correlation mode, and the workflow creation information is generated.
Step S12, determining first response information and workflow type based on the workflow creation information, a preset flow knowledge base and a preset flow knowledge map;
It should be noted that, the preset flow knowledge base is formed by data such as a BPM (Business Process Management ) related FAQ document (Frequently Asked Questions, common problem solution), a system stock flow, user information, history creation information and the like, and can realize data access in various modes such as SQL, excel, word and the like, enrich data sources, and refer to fig. 4, fig. 4 is a partial construction flow chart of the preset flow knowledge base provided by the invention, wherein a LANGCHAIN document loader can be utilized to perform operations such as file loading and the like. In addition, the preset flow knowledge base can be continuously updated along with the access of data so as to improve more accurate response generation capability.
It should be further noted that the preset flow knowledge graph refers to a preset knowledge graph storing various structured data such as basic attribute knowledge, associated knowledge, event knowledge, time sequence knowledge, resource knowledge, and the like, where the preset flow knowledge graph may be automatically updated in an iterative manner to obtain more accurate knowledge and generate more accurate and detailed response.
Additionally, the first response information is response information generated by the preset flow knowledge graph, the workflow creation information, and the like. The workflow types include, but are not limited to, fine-tuning of flow design, generation of work orders, filling of forms, and the like.
Specifically, text segmentation is performed on the workflow creation information, the segmented workflow creation information is vectorized to obtain a workflow creation vector, query matching is performed on the workflow creation vector and a preset flow knowledge base to obtain the workflow type and initial response information, and accordingly structuring processing is performed on the initial response information, and the structured initial response information is matched with a preset flow knowledge graph to obtain the first response information.
Step S13, inputting the first response information into a business process question-answer model to obtain second response information output by the business process question-answer model;
It should be noted that, the business process question-answer model is obtained by iterative training of process demand samples such as a model training set, a model verification set and a model test set. In addition, the second response information refers to response information generated by the business process question-answer model aiming at the first response information, and when the target user carries out business query through the chat robot, the second response information can be returned to the target user so as to realize user interaction and enhance system response capability.
And step S14, the second response information and the workflow type are sent to a workflow creation engine, so that the workflow creation engine creates a workflow corresponding to the workflow creation information.
The workflow creation engine refers to a logic solution engine for delivering routes, content levels, etc. according to different decision information such as workflow entities, entity attributes, entity association situations, etc. in the workflow creation information, and may be an actiti flow engine, etc., which is not limited herein. For example, a workflow of employee invitation bars is generated, and when a department director approves the invitation, the workflow creation engine generates a workflow of issuing the invitation bars to a company upper layer leader for confirmation approval according to business logic. The workflow creation engine includes various workflow type modules, such as a work order generation module, a form filling module, a flow design fine adjustment module, and the like, which are not limited herein.
It should be further noted that, the Workflow (Workflow) is a systematic and orderly business process for organizing, managing and executing specific tasks or business activities, and the Workflow may cover business processes in various organizations, and may be planned and managed from simple daily tasks to complex business processes by using the Workflow, and may be set according to practical situations, which will not be described in detail herein.
Specifically, the second response information and the information such as the workflow type are sent to the workflow creation engine, so that after the workflow creation engine receives the second response information and the information such as the workflow type, a workflow corresponding to the workflow creation information is created, and fig. 5 can be referred to fig. 5, where fig. 5 is a partial function implementation diagram of the workflow creation system provided by the present invention.
According to the embodiment of the invention, the workflow creation information corresponding to the target user is obtained, and the first response information and the workflow type are determined based on the workflow creation information, the preset flow knowledge base and the preset flow knowledge map, so that the first response information is input into the business flow question-answer model to obtain the second response information output by the business flow question-answer model, and the second response information and the workflow type are further sent to the workflow creation engine to enable the workflow creation engine to create the workflow corresponding to the workflow creation information, thereby realizing automatic generation, intelligent filling and the like of the workflow, remarkably improving the working efficiency and accuracy, simultaneously reducing human errors, effectively reducing the operation cost, reducing the operation burden of the user and improving the user experience.
In one embodiment of the present invention, the determining the first response information and the workflow type based on the workflow creation information, the preset flow knowledge base and the preset flow knowledge graph includes:
The method comprises the steps of carrying out text segmentation on workflow creation information, vectorizing the segmented workflow creation information to obtain workflow creation vectors, carrying out query matching on the workflow creation vectors and a preset flow knowledge base to obtain workflow types and initial response information, carrying out structuring processing on the initial response information, and carrying out matching on the structured initial response information and a preset flow knowledge graph to obtain the first response information.
It should be noted that, the workflow creation vector refers to a vector value obtained in the process of converting workflow creation information into a mathematical vector, so as to allow a computer system and the like to better understand and process data such as text. Additionally, the initial response information refers to response information generated according to the preset flow knowledge base.
Specifically, the workflow creation information is subjected to text segmentation, such as text segmentation into words or vocabulary units, and removal of stop words, which are words that frequently occur in text but do not usually carry specific information, such as "and", "the", and the like. And vectorizing the partitioned workflow creation information to obtain a workflow creation vector, wherein vectorization can be performed by adopting a Word Bag model (Bag of Words, boW), a TF-IDF (Term Frequency-Inverse Document Frequency), a Word embedding model (such as Word2Vec, gloVe, fastText) and the like, and the method is not limited herein.
Further, query matching is performed on the workflow creation vector and a preset flow knowledge base, wherein a faiss (Facebook AI SIMILARITY SEARCH) search method may be adopted, the workflow type and the initial response information are determined according to a matching result, and then the initial response information is structured, that is, the initial response information is converted into structured information, and referring to fig. 7, fig. 7 is an information structured example diagram provided by the present invention, and the structured initial response information is matched with a preset flow knowledge graph, that is, matched to a relevant entity node corresponding to the initial response information, so that the structured initial response information is filled into the relevant entity node to obtain the first response information.
According to the embodiment of the invention, the workflow creation information is subjected to text segmentation, the segmented workflow creation information is vectorized to obtain the workflow creation vector, the workflow creation vector is further subjected to query matching with the preset flow knowledge base to obtain the workflow type and the initial response information, the initial response information is subjected to structuring processing, the structured initial response information is matched with the preset flow knowledge graph to obtain the first response information, the user intention is effectively and accurately identified through the preset flow knowledge base and the preset flow knowledge graph, and accurate and actual response information is generated according to the intention, so that the response capability of the chat robot is effectively improved, the user experience and the work efficiency are improved, the operation cost is further reduced, and the operation burden of the user is lightened.
In an embodiment of the present invention, before determining the first response information and the workflow type based on the workflow creation information, the preset flow knowledge base and the preset flow knowledge map, the method further includes:
Acquiring business process knowledge, extracting information and fusing knowledge of the business process knowledge to obtain workflow structured knowledge, performing quality assessment on the workflow structured knowledge, and storing the workflow structured knowledge to the preset process knowledge graph if the quality assessment is passed.
It should be noted that, the conventional vector search-based process is that a file is loaded, a text is read, a text is segmented, a text is vectorized, a question is vectorized, top k pieces of text which are most similar to the question vector are matched in the text vector, the matched text is added to a prompt as a context and a question, and the answer is submitted to an LLM (Large Language Model ) to generate the answer. This approach is very effective when vector searching can produce relevant text blocks. However, when LLM requires information from multiple documents and even multiple blocks to generate an answer, a simple vector similarity search may not be accurate enough. Moreover, this approach also tends to miss interrelated document blocks, resulting in incomplete information retrieval. Meanwhile, therefore, in order to reduce inaccuracy caused by the semantic search based on embedding, a preset flow knowledge graph based on the BPM business flow field needs to be constructed.
It should be further noted that, the business process knowledge refers to various kinds of knowledge related to the BPM, including basic attribute knowledge, associated knowledge, event knowledge, time sequence knowledge, resource class knowledge, and the like. The workflow structured knowledge is obtained by carrying out information extraction, knowledge fusion and other construction operations on business process knowledge.
Specifically, business process knowledge is obtained, information extraction and knowledge fusion are further carried out on the business process knowledge to obtain workflow structured knowledge, wherein the information extraction refers to the extraction of workflow entities, entity attributes, entity association conditions and the like from the business process knowledge, such as a middle gate supervisor of a leave-out process and a company upper layer leader, and logic relations and the like between the two entities, on the basis of the information extraction, an ontology knowledge expression is formed, for example, the leave-out process is transferred to the company upper layer leader only after the leave-out process is approved by a department supervisor, the knowledge fusion refers to the integration of the business process knowledge to eliminate contradictions and ambiguities, such as that a certain entity may have various expressions, and a certain specific name also corresponds to the association among a plurality of different entities and ambiguities.
Further, the quality evaluation is performed on the workflow structured knowledge, wherein the quality evaluation can be performed through a system, the quality evaluation can also be performed through manual screening, and if the quality evaluation passes, the workflow structured knowledge is stored in the preset flow knowledge graph, and fig. 6 can be referred to fig. 6, and fig. 6 is a partial construction flow chart of the preset flow knowledge graph provided by the invention. Additionally, in order to update the knowledge graph and pursue more accurate response generation capability, constructing the preset flow knowledge graph is an iterative updating process, and each iteration includes four stages of information extraction, knowledge fusion, knowledge calculation (quality assessment) and knowledge storage.
In addition, the knowledge graph of the preset flow can also realize knowledge modeling, namely, a knowledge graph centering on the target user is established aiming at all workflow creation information and the like of the target user, so that more efficient and personalized accurate response information is generated, and the working efficiency of the target user is improved.
According to the embodiment of the invention, the business process knowledge is obtained, the information extraction and knowledge fusion are further carried out on the business process knowledge, the workflow structured knowledge is obtained, and the quality evaluation is carried out on the workflow structured knowledge, so that if the quality evaluation passes, the workflow structured knowledge is stored in the preset process knowledge graph, the inaccuracy caused by semantic search based on embedding is further reduced, the response accuracy is improved, and the service quality of higher quality is improved for users.
In one embodiment of the present invention, before the first response information is input to the business process question-answer model to obtain the second response information output by the business process question-answer model, the method further includes:
and performing iterative training and model fine tuning on an initial business process question-answer model based on the model training set, the model verification set and the model test set to obtain the business process question-answer model.
It should be noted that if a large model corpus of language models such as ChatGLM B model is directly adopted, because the corpus does not have a corresponding training set for BPM workflow, the answer result generated by ChatGLM B model and the like cannot reach the target accuracy, and therefore, retraining is needed to obtain the business process question-answer model, wherein the model fine Tuning can be performed by adopting a P-Tuning v2 method, so that the model fine Tuning can serve the target user better, and better service experience can be brought to the user.
It should be further noted that, the flow requirement samples refer to relevant question-answer data generated by the target user in the system operation and maintenance process, such as "what is a configuration class? sample labels corresponding to" what is a configuration class "are" generally mentioned requirements do not relate to development, such as non-development requirements of data, rights applications, etc., currently supported configuration class requirements are: guide business data, authority configuration, process node adjustment, etc.). In addition, the model training set is used for a model training process, the model verification set is used for evaluating the performance of the verification set after each training iteration so as to carry out model fine adjustment according to the evaluation result, and the model test set is used for evaluating the performance and the accuracy of the model.
Specifically, in an embodiment, a plurality of sets of process requirement samples are obtained, and each process requirement sample is divided into a model training set, a model verification set and a model test set, so that data preprocessing is performed on the model training set, the model verification set and the model test set, such as word segmentation is performed on an input text, stop words are removed, and the input text is converted into an input format required by a model.
Further, a pre-trained initial business process question-answer model is loaded, wherein the initial business process question-answer model comprises a vocabulary table, weight parameters and the like, the initial business process question-answer model can be ChatGLM B model and the like, and additional trainable parameters are added to an input layer of the initial business process question-answer model and used for representing the Prompt parameters so as to construct an initial business process question-answer model based on a P-Tuning v2 method.
Furthermore, the model training set is input to the initial business process question-answer model to obtain a predicted value output by the initial business process question-answer model, and further, based on the predicted value and a sample label corresponding to the process demand sample, a model loss value is obtained by calculating an entropy loss function, wherein in the embodiment, the loss function can be set according to actual demands, and the method is not particularly limited herein. After the model loss value is obtained through calculation, the training process is finished, model parameters in the initial business process question-answering model are updated by using an error back propagation algorithm, an SGD optimizer and the like, and then the next training is carried out. After each training iteration, the performance of the model validation set is evaluated and model adjustments are made as needed.
Further, in the model adjustment process, judging whether the updated initial business process question-answer model meets a preset training ending condition according to the model test set, if so, taking the updated initial business process question-answer model as the business process question-answer model, and if not, continuing training the model, wherein the preset training ending condition comprises loss convergence, reaching a maximum iteration number threshold and the like.
Further, after model training is completed, the training test set is used for evaluating the performance and accuracy of the business process question-answer model.
According to the embodiment of the invention, a plurality of groups of flow demand samples are obtained, and each flow demand sample is divided into a model training set, a model verification set and a model test set, so that an initial business flow question-answer model is subjected to iterative training based on the model training set, the model verification set and the model test set, and the business flow question-answer model is obtained, so that the robustness of the model is improved, more accurate and more specialized professional field response is generated, and high-quality service is improved for users, and user experience is improved.
In one embodiment of the present invention, the obtaining workflow creation information corresponding to the target user includes:
The method comprises the steps of obtaining workflow creation requirements input by a target user, judging whether history creation information is matched based on the workflow creation requirements, if so, performing association combination on the matched history creation information and the workflow creation requirements to generate the workflow creation information, and if not, generating the workflow creation information based on the workflow creation requirements.
It should be noted that, the workflow creation requirement refers to requirement information of a target user for creating a workflow, which may be obtained by performing multiple rounds of conversations with the target user by the chat robot, for example, when the target user wants to generate a workflow, the chat robot needs to determine whether to generate the workflow according to a history creation record and whether to fill out a default bill according to the form through multiple rounds of conversations with the target user so as to obtain information required for generating the workflow from the target user, and when the target user only needs to generate a whole notice, the chat robot can also directly understand the intention of the target user through one round of conversations, so that the acquisition mode of the workflow creation requirement depends on the requirement of the target user, which is not limited, and can be set according to practical situations.
It should be further noted that, in order to further improve the efficiency of creating the workflow, the chat robot may determine the real requirement of the target user by extracting the history creation information of the target user, where the history creation information refers to the record information of the workflow created by the target user before, for example, the form generation history information, the work form extraction history information, the flow design modification history information, and the like, and further includes structured and unstructured data such as form data and flow design nodes, which may be set according to the real requirement of the target user, and is not limited herein.
Specifically, the workflow creation requirement input by the target user is obtained, and then whether the history creation information is matched is judged based on the workflow creation requirement, for example, semantic analysis is performed on the workflow creation requirement, and whether the history creation information is matched is judged by directly initiating inquiry (for example, whether a bill is carried out according to the history bill record of 2024, 1 month and 8. And if so, carrying out association combination on the matched history creation information, the workflow creation requirement and other information to generate the workflow creation information.
In addition, if the history creation information does not need to be matched, the workflow creation information can be generated directly according to the workflow creation requirement, for example, data information and the like required for extracting a form can be obtained directly through multiple rounds of dialogue between the chat robot and the target user, and further the data information is analyzed and integrated to generate the workflow creation information, and the workflow creation information can be set according to actual conditions without limitation.
According to the embodiment of the invention, the workflow creation requirement input by the target user is acquired, and whether the history creation information is matched is judged based on the workflow creation requirement, so that if yes, the matched history creation information and the workflow creation requirement are combined in a correlated way to generate the workflow creation information, otherwise, the workflow creation information is generated based on the workflow creation requirement, and further, the conversation is carried out between the chat robot and the target user, so that the real requirement of the target user is clarified, the workflow creation efficiency is effectively improved, efficient portable service is provided for the user, the operation cost is effectively reduced, and the operation load of the user is reduced.
Fig. 2 is a schematic flow chart of a workflow creation method provided by the invention. As shown in fig. 2, the workflow creation method is applied to a workflow creation engine, and includes:
Step S21, receiving second response information sent by the chat robot and workflow type;
It should be noted that, the chat robot is an AI chat robot, and may generate corresponding response information for the problem of the target user, and return to the target user, so as to implement user interaction, make sure the user intention, and so on.
Step S22, determining a workflow creation module corresponding to the workflow based on the workflow type;
The workflow creation module refers to a creation module corresponding to a workflow type, and includes a work order generation module, a form filling module, a flow design fine adjustment module, and the like, which are not limited herein.
Specifically, comparing and matching the workflow type with a preset workflow creation module library, and determining a workflow creation module corresponding to the workflow.
Step S23, analyzing the second response information to obtain workflow entities, entity attributes and entity association conditions;
It should be noted that, the workflow entity includes a name of the workflow, an executor, filling data, etc., the entity attribute includes an identity of the executor, etc., and the entity association condition includes an execution sequence, etc., and reference may be made to fig. 7.
Specifically, the analysis may be performed according to the corresponding relevant entity node (i.e., the knowledge graph) in the second response information, and may also be performed by means of semantic segmentation, etc., to obtain information such as a workflow entity, an entity attribute, and an entity association condition, which is not limited herein.
And step S24, inputting the workflow entity, the entity attribute and the entity association condition into the workflow creation module so as to enable the workflow creation module to create the workflow.
It should be noted that, the creation manner of the workflow creation module depends on the workflow type and is not consistent, and each module has a corresponding creation method, and reference may be made to fig. 8, 9 and 10, where the creation manner is not limited, and may be set according to practical situations. Meanwhile, the independent creation module is beneficial to improving the working efficiency, and the creation of a plurality of workflows can be performed simultaneously.
According to the embodiment of the invention, the workflow creation module corresponding to the workflow is determined by receiving the second response information and the workflow type sent by the chat robot and based on the workflow type, and further the second response information is analyzed to obtain the workflow entity, the entity attribute and the entity association condition, so that the workflow entity, the entity attribute and the entity association condition are input into the workflow creation module to be used for creating the workflow by the workflow creation module, and further the automatic generation, intelligent filling and the like of the workflow are realized by the workflow creation engine, thereby remarkably improving the working efficiency and the accuracy, simultaneously reducing the human error, further effectively reducing the operation cost, relieving the user operation burden and improving the user experience.
In one embodiment of the present invention, after the inputting the workflow entity, the entity attribute, and the entity association condition into the workflow creation module, the workflow creation module creates the workflow, the method further includes:
the method comprises the steps of obtaining creation success time and a viewing mode of the workflow, generating creation success information based on the workflow type, the creation success time and the viewing mode, pushing the creation success information to a target user to remind the target user to view.
It should be noted that the creation success information refers to reminder information generated when the workflow is created successfully, including, but not limited to, a workflow type, a creation success time, a workflow screenshot, a viewing manner, and the like.
Specifically, the creation success time and the viewing mode of the workflow are obtained, and then the information such as the workflow type, the creation success time and the viewing mode are combined in a correlated manner to generate creation success information, so that the creation success information is pushed to a target user to remind the target user to view, wherein the information can be pushed in modes such as a short message, an APP notification, an applet reminder and an electronic mailbox, and the information is not limited and can be set according to actual conditions.
According to the embodiment of the invention, the creation success time and the checking mode of the workflow are obtained, and the creation success information is generated based on the workflow type, the creation success time and the checking mode, so that the creation success information is pushed to the target user to remind the target user to check, thereby improving the working efficiency, enabling the target user to check the workflow with the creation success timely and efficiently, and improving the user experience.
It should be understood that the sequence number of each step in the foregoing embodiment does not mean that the execution sequence of each process should be determined by the function and the internal logic, and should not limit the implementation process of the embodiment of the present invention.
In an embodiment, a workflow creation device is provided, which corresponds to the workflow creation method in the above embodiment one by one. As shown in fig. 3, the workflow creation apparatus includes a chat robot 31 and a workflow creation engine 32. The functional modules are described in detail as follows:
The chat robot 31 is used for acquiring workflow creation information input by a user, determining first response information and workflow type based on the workflow creation information and a preset flow knowledge graph, inputting the first response information into a business flow question-answer model, and obtaining second response information output by the business flow question-answer model;
the workflow creation engine 32 is configured to receive second response information and a workflow type sent by the chat robot, determine a workflow creation module corresponding to the workflow based on the workflow type, parse the second response information to obtain a workflow entity, an entity attribute and an entity association condition, and input the workflow entity, the entity attribute and the entity association condition into the workflow creation module to be used by the workflow creation module to create the workflow.
The workflow creation apparatus is further for:
Text segmentation is carried out on the workflow creation information, and the segmented workflow creation information is vectorized to obtain a workflow creation vector;
Inquiring and matching the workflow creation vector with a preset flow knowledge base to obtain the workflow type and initial response information;
And carrying out structuring treatment on the initial response information, and matching the structured initial response information with a preset flow knowledge graph to obtain the first response information.
The workflow creation apparatus is further for:
Acquiring business process knowledge;
Information extraction and knowledge fusion are carried out on the business process knowledge to obtain workflow structured knowledge, and quality evaluation is carried out on the workflow structured knowledge;
and if the quality evaluation is passed, storing the workflow structured knowledge to the preset flow knowledge graph.
The workflow creation apparatus is further for:
Obtaining a plurality of groups of flow demand samples, and dividing each flow demand sample into a model training set, a model verification set and a model test set;
And performing iterative training and model fine tuning on an initial business process question-answer model based on the model training set, the model verification set and the model test set to obtain the business process question-answer model.
The workflow creation apparatus is further for:
Acquiring workflow creation requirements input by a target user;
judging whether history creation information is matched or not based on the workflow creation requirement;
If yes, performing association combination on the matched history creation information and the workflow creation requirement to generate workflow creation information;
if not, generating the workflow creation information based on the workflow creation requirement.
The workflow creation apparatus is further for:
Acquiring the creation success time and the checking mode of the workflow;
generating creation success information based on the workflow type, the creation success time and the viewing mode;
Pushing the creation success information to a target user so as to remind the target user to check.
For specific limitations on the workflow creation means, reference may be made to the above limitations on the workflow creation method, and no further description is given here. The respective modules in the above-described workflow creation apparatus may be implemented in whole or in part by software, hardware, and a combination thereof. The above modules may be embedded in hardware or independent of a processor in the electronic device, or may be stored in software in a memory in the electronic device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, an electronic device is provided, which may be a server. The electronic device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device includes a readable storage medium, an internal memory. The readable storage medium stores an operating system, computer readable instructions, and a database. The internal memory provides an environment for the execution of an operating system and computer-readable instructions in a readable storage medium. The database of the electronic device is used for storing data related to the workflow creation method. The network interface of the electronic device is used for communicating with an external terminal through a network connection. The computer readable instructions when executed by a processor implement a workflow creation method. The readable storage medium provided by the present embodiment includes a nonvolatile readable storage medium and a volatile readable storage medium.
In one embodiment, an electronic device is provided, which may be a terminal device, and an internal structure diagram thereof may be as shown in fig. 11. The electronic device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device includes a readable storage medium. The readable storage medium stores computer readable instructions. The network interface of the electronic device is used for communicating with an external terminal through a network connection. The computer readable instructions when executed by a processor implement a workflow creation method. The readable storage medium provided by the present embodiment includes a nonvolatile readable storage medium and a volatile readable storage medium.
In one embodiment, an electronic device is provided that includes a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, when executing the computer readable instructions, implementing the steps of a workflow creation method as described above.
In an embodiment, a readable storage medium is provided, the readable storage medium storing computer readable instructions which, when executed by a processor, implement the workflow creation method steps as described above. Those skilled in the art will appreciate that implementing all or part of the above described embodiment methods may be accomplished by instructing the associated hardware by computer readable instructions stored in a non-volatile readable storage medium or a volatile readable storage medium, which when executed may comprise the above described embodiment methods. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions.
The foregoing embodiments are merely for illustrating the technical solution of the present invention, but not for limiting the same, and although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solution described in the foregoing embodiments may be modified or substituted for some of the technical features thereof, and that these modifications or substitutions should not depart from the spirit and scope of the technical solution of the embodiments of the present invention and should be included in the protection scope of the present invention.

Claims (9)

The voice chat robot is used for acquiring workflow creation information corresponding to a target user, determining first response information and workflow types based on the workflow creation information, a preset flow knowledge base and a preset flow knowledge map, inputting the first response information into a business flow question-answer model to obtain second response information output by the business flow question-answer model, sending the second response information and the workflow types to a workflow creation engine for the workflow creation engine to create a workflow corresponding to the workflow creation information, performing text segmentation on the workflow creation information, vectorizing the segmented workflow creation information to obtain a workflow creation vector, inquiring and matching the workflow creation vector with the preset flow knowledge base to obtain the workflow types and initial response information, performing structural processing on the initial response information, matching the initial response information after the structural processing with the preset flow knowledge map to obtain relevant entity nodes, and filling the initial response information into the relevant entity nodes to obtain the first response information;
CN202410102089.6A2024-01-252024-01-25 Workflow creation method, device, equipment and storage mediumActiveCN117634865B (en)

Priority Applications (1)

Application NumberPriority DateFiling DateTitle
CN202410102089.6ACN117634865B (en)2024-01-252024-01-25 Workflow creation method, device, equipment and storage medium

Applications Claiming Priority (1)

Application NumberPriority DateFiling DateTitle
CN202410102089.6ACN117634865B (en)2024-01-252024-01-25 Workflow creation method, device, equipment and storage medium

Publications (2)

Publication NumberPublication Date
CN117634865A CN117634865A (en)2024-03-01
CN117634865Btrue CN117634865B (en)2024-12-06

Family

ID=90032440

Family Applications (1)

Application NumberTitlePriority DateFiling Date
CN202410102089.6AActiveCN117634865B (en)2024-01-252024-01-25 Workflow creation method, device, equipment and storage medium

Country Status (1)

CountryLink
CN (1)CN117634865B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
CN118941251A (en)*2024-10-142024-11-12深圳市智慧城市科技发展集团有限公司 Business process modeling method, device and storage medium
CN120634484A (en)*2025-08-082025-09-12北京飞书科技有限公司Model-based workflow creation method, device, equipment, medium and product

Citations (2)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
CN112507092A (en)*2020-12-032021-03-16北京智齿博创科技有限公司Implementation method for completing work order intelligent filling function based on AI text robot
CN115481229A (en)*2022-09-222022-12-16上海浦东发展银行股份有限公司Method and device for pushing answer call, electronic equipment and storage medium

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
CN103942220A (en)*2013-01-232014-07-23成都勤智数码科技股份有限公司Method used for intelligently linking work orders with knowledge of knowledge base and suitable for IT operation and maintenance system
US10474745B1 (en)*2016-04-272019-11-12Google LlcSystems and methods for a knowledge-based form creation platform
CN111753075B (en)*2020-08-122024-08-23腾讯科技(深圳)有限公司Method and device for creating customer service robot question-answer data and computer equipment
CN111966787B (en)*2020-08-182024-09-06上海海洋大学Knowledge graph-based construction method of intelligent fishery question-answering robot
CN112036736B (en)*2020-08-282023-08-01杭州指令集智能科技有限公司Workflow creation method and device
WO2023004805A1 (en)*2021-07-302023-02-02西门子股份公司Workflow modeling implementation system and method, and storage medium
CN114417010A (en)*2021-12-312022-04-29第四范式(北京)技术有限公司Knowledge graph construction method and device for real-time workflow and storage medium
CN116521653A (en)*2022-01-202023-08-01合肥美的电冰箱有限公司Food material question-answering method and system based on knowledge graph
CN115292022B (en)*2022-09-292023-01-20泰豪软件股份有限公司Workflow engine system, implementation method, storage medium and computer equipment
CN116992092A (en)*2023-07-312023-11-03远光软件股份有限公司Method and device for establishing flow model, storage medium and terminal equipment
CN117235228A (en)*2023-09-252023-12-15上海东普信息科技有限公司Customer service question-answer interaction method, device, equipment and storage medium

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication numberPriority datePublication dateAssigneeTitle
CN112507092A (en)*2020-12-032021-03-16北京智齿博创科技有限公司Implementation method for completing work order intelligent filling function based on AI text robot
CN115481229A (en)*2022-09-222022-12-16上海浦东发展银行股份有限公司Method and device for pushing answer call, electronic equipment and storage medium

Also Published As

Publication numberPublication date
CN117634865A (en)2024-03-01

Similar Documents

PublicationPublication DateTitle
CN117634865B (en) Workflow creation method, device, equipment and storage medium
US11005786B2 (en)Knowledge-driven dialog support conversation system
EP3557439A1 (en)Deep learning techniques based multi-purpose conversational agents for processing natural language queries
WO2021138020A1 (en)Systems and methods for artificial intelligence enhancements in automated conversations
CN112036842B (en)Intelligent matching device for scientific and technological service
CN113988071A (en) An intelligent dialogue method and device based on financial knowledge graph, and electronic equipment
US20140108103A1 (en)Systems and methods to control work progress for content transformation based on natural language processing and/or machine learning
CN116737129B (en)Supply chain control tower generation type large language model and construction method thereof
AU2012201529A1 (en)MVT optimization of business process modeling and management
Gupta et al.Reducing user input requests to improve IT support ticket resolution process
CN117151647B (en) A method, device and medium for auxiliary audit of documents
CN120011187B (en) Multi-scene and multi-base large model engine system
US11922352B1 (en)System and method for risk tracking
KR20240053911A (en)Method and system for AI collaboration service based on source code automatic generation system
CN114722729A (en)Automatic cutter recommendation method and device, terminal and storage medium
US11314488B2 (en)Methods and systems for automated screen display generation and configuration
CN113761193A (en) Log classification method, apparatus, computer equipment and storage medium
US11941500B2 (en)System for engagement of human agents for decision-making in a dynamically changing environment
Wang et al.SchemaAgent: A Multi-Agents Framework for Generating Relational Database Schema
US20250217341A1 (en)Metadata determination and storage method
GeerligsInformation Extraction from Contracts Using Large Language Models
CN119722349B (en)SaaS service management method, system, computer equipment and storage medium
US12379947B2 (en)Method for dynamic AI supported graph-analytics self learning templates
CN118673135A (en)Output control method and device for target object description information
CN120655226A (en) Method, device, equipment and medium for generating approval chain based on large model

Legal Events

DateCodeTitleDescription
PB01Publication
PB01Publication
SE01Entry into force of request for substantive examination
SE01Entry into force of request for substantive examination
GR01Patent grant
GR01Patent grant

[8]ページ先頭

©2009-2025 Movatter.jp