CN-122018874-A - Intelligent assistant creation method, system and storage medium

CN122018874ACN 122018874 ACN122018874 ACN 122018874ACN-122018874-A

Abstract

The application discloses an intelligent assistant creation method, an intelligent assistant creation system and a storage medium, which are used for efficiently and conveniently creating an intelligent assistant fitting user demands in a full-flow automatic mode and can adapt to diversification and dynamic evolution of the user demands. The intelligent assistant creation method includes the steps of responding to a first task requirement of a user instruction to create an intelligent assistant, obtaining a first flow Cheng Moban matched with the first task requirement from a template library, generating a first flow executed by the created intelligent assistant based on the first flow Cheng Moban, matching at least one first agent component for a first subtask in the first flow based on description information of each agent component in an agent component library, assembling at least one first agent component corresponding to the first subtask to obtain a first agent corresponding to the first subtask, and creating interaction logic between the first agents corresponding to the first subtask based on a logic relationship between the first subtasks to obtain the first intelligent assistant.

Inventors

WANG NA
YANG JIN
YAO RONG
CUI TONG
GUO KEQIANG
GONG YI
ZHU JINGYU
TANG YANJUAN
CHEN DONGJIE

Assignees

中移数智科技有限公司
中国移动通信集团设计院有限公司
中国移动通信集团有限公司

Dates

Publication Date: 20260512
Application Date: 20251210

Claims (10)

1. A method for creating an intelligent assistant, comprising: In response to a user indication to create a first task requirement of the intelligent assistant, obtaining a first flow Cheng Moban from a template library that matches the first task requirement, and generating a first flow based on the first flow Cheng Moban that is performed by the created intelligent assistant, the first flow including a plurality of first subtasks and logical relationships between the plurality of first subtasks; Matching at least one first agent component for a first subtask in the first flow based on the description information of each agent component in the agent component library; And assembling at least one first agent component corresponding to the first subtask in the first flow to obtain a first agent corresponding to the first subtask in the first flow, and creating interaction logic between the first agents corresponding to the first subtask in the first flow based on the logic relationship to obtain a first intelligent assistant.
2. The method of claim 1, wherein the first task requirement is obtained by: receiving dialogue data input by the user on a dialogue interface; performing intention recognition on the dialogue data to obtain a first intention; In response to the first intent being to create an intelligent assistant, the first task requirement is generated based on the first intent.
3. The method of claim 2, wherein the generating the first task demand based on the first intent comprises: Responding to the dialogue data not containing the first user characteristics of the user, acquiring second user characteristics of the user from a user characteristic library, and expanding the first intention based on the second user characteristics to obtain the first task requirement; Responding to the dialogue data to contain a first user characteristic of the user, expanding the first intention based on the first user characteristic to obtain the first task requirement, and writing the first user characteristic into the user characteristic library to update the second user characteristic.
4. The method of claim 1, wherein retrieving a first stream Cheng Moban from the template library that matches the first task requirement comprises: Decomposing the first task requirement into a plurality of second subtasks; Matching the plurality of second subtasks with the subtasks in each flow template in the template library to obtain the matching degree of the first task requirement and each flow template; And determining a first flow Cheng Moban matched with the first task demand in the template library based on the matching degree of the first task demand and each flow template.
5. The method of claim 1, wherein the description information of the agent component includes functional information of the agent component; The matching of at least one first agent component for a first subtask in the first flow based on the description information of each agent component in the agent component library includes: for each first subtask, matching task information of the first subtask with function information of each intelligent body component in the intelligent body component library to obtain matching degree between each intelligent body component and the first subtask; And acquiring at least one intelligent agent component with the matching degree meeting the preset matching condition between the intelligent agent component library and the first subtask from the intelligent agent component library, and taking the at least one intelligent agent component as at least one first intelligent agent component corresponding to the first subtask.
6. The method according to claim 1, wherein the method further comprises: Testing the first intelligent assistant based on the first flow to obtain a test result; And optimizing the first intelligent assistant based on the test result.
7. The method of claim 6, wherein the optimizing the first intelligent assistant based on the test results comprises: Presenting the first intelligent assistant in response to the test result being a test pass; In response to receiving negative feedback operation of the user on the first intelligent assistant, presenting first prompt information, wherein the first prompt information is used for indicating to propose an optimization suggestion for the first intelligent assistant; And in response to receiving the first optimization suggestion information input by the user, updating the first task requirement based on the first optimization suggestion, and repeating the steps of acquiring a first flow template matched with the first task requirement from a template library to present the created first intelligent assistant based on the updated first task requirement until a positive feedback operation of the user on the created first intelligent assistant is received.
8. The method of claim 6, wherein the testing the first intelligent assistant based on the first flow results in a test result, comprising: For each first subtask in the first flow, responding to the fact that the first subtask has a corresponding first agent, and indicating the first agent to execute the first subtask to obtain an execution result of the first subtask; The first intelligent assistant is instructed to execute the first flow, and an execution result of the first flow is obtained; And determining a test result of the first intelligent assistant based on the execution result of the first subtask and the execution result of the first flow.
9. The intelligent assistant creation system is characterized by comprising a demand processing intelligent agent subsystem, an inference planning intelligent agent subsystem, an assembly and collaboration subsystem and a data pool, wherein a template library and an intelligent agent assembly library are arranged in the data pool; The demand processing intelligent agent subsystem is used for acquiring a first task demand of a user for indicating to create an intelligent assistant; The reasoning planning agent subsystem is used for responding to the first task requirement, acquiring a first flow Cheng Moban matched with the first task requirement from the template library, and generating a first flow executed by the created intelligent assistant based on the first flow Cheng Moban, wherein the first flow comprises a plurality of first subtasks and logic relations among the plurality of first subtasks; The assembly and collaboration subsystem is used for matching at least one first agent component for a first subtask in the first flow based on the description information of each agent component in the agent component library, assembling at least one first agent component corresponding to the first subtask in the first flow to obtain a first agent corresponding to the first subtask in the first flow, and creating interaction logic between the first agents corresponding to the first subtask in the first flow based on the logic relationship to obtain a first intelligent assistant.
10. A computer readable storage medium, characterized in that instructions in the storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the intelligent assistant creation method of any of claims 1 to 8.

Description

Intelligent assistant creation method, system and storage medium Technical Field The application relates to the technical field of artificial intelligence, in particular to an intelligent assistant creation method, an intelligent assistant creation system and a storage medium. Background With the rapid development of artificial intelligence technology, creating intelligent assistants based on large model technology to assist business processing has become an important path for significantly improving business efficiency. However, in view of diversification and dynamic evolution of user demands, there is still a need for a perfect solution how to quickly and conveniently create an intelligent assistant that meets the user demands. Disclosure of Invention The embodiment of the application aims to provide an intelligent assistant creation method, an intelligent assistant creation system and a storage medium, which are used for efficiently and conveniently creating an intelligent assistant fitting user demands in a full-flow automatic mode and can adapt to diversification and dynamic evolution of the user demands. In order to achieve the above object, the embodiment of the present application adopts the following technical scheme: in a first aspect, an embodiment of the present application provides a method for creating an intelligent assistant, including: In response to a user indication to create a first task requirement of the intelligent assistant, obtaining a first flow Cheng Moban from a template library that matches the first task requirement, and generating a first flow based on the first flow Cheng Moban that is performed by the created intelligent assistant, the first flow including a plurality of first subtasks and logical relationships between the plurality of first subtasks; Matching at least one first agent component for a first subtask in the first flow based on the description information of each agent component in the agent component library; And assembling at least one first agent component corresponding to the first subtask in the first flow to obtain a first agent corresponding to the first subtask in the first flow, and creating interaction logic between the first agents corresponding to the first subtask in the first flow based on the logic relationship to obtain a first intelligent assistant. In a second aspect, an embodiment of the present application provides an intelligent assistant creation system, including: The system comprises a demand processing agent subsystem, an inference planning agent subsystem, an assembly and collaboration subsystem and a data pool, wherein a template library and an agent component library are arranged in the data pool; The demand processing intelligent agent subsystem is used for acquiring a first task demand of a user for indicating to create an intelligent assistant; The reasoning planning agent subsystem is used for responding to the first task requirement, acquiring a first flow Cheng Moban matched with the first task requirement from the template library, and generating a first flow executed by the created intelligent assistant based on the first flow Cheng Moban, wherein the first flow comprises a plurality of first subtasks and logic relations among the plurality of first subtasks; The assembly and collaboration subsystem is used for matching at least one first agent component for a first subtask in the first flow based on the description information of each agent component in the agent component library, assembling at least one first agent component corresponding to the first subtask in the first flow to obtain a first agent corresponding to the first subtask in the first flow, and creating interaction logic between the first agents corresponding to the first subtask in the first flow based on the logic relationship to obtain a first intelligent assistant. In a third aspect, embodiments of the present application provide a computer readable storage medium, which when executed by a processor of an electronic device, enables the electronic device to perform the intelligent assistant creation method as provided in the first aspect. The above at least one technical scheme adopted by the embodiment of the application can achieve the following beneficial effects: The method comprises the steps of integrating a large number of process templates through a process template library, acquiring matched first flows Cheng Moban from the template library according to task requirements in the creation process, generating a first process executed by the created intelligent assistant based on the first flows, reducing blindness in the process of generating the flows, ensuring that the generated first process fully fits task requirements, integrating a large number of intelligent body components through an intelligent body component library, matching at least one first intelligent body component for first subtasks in the first flow based on description information of