Search

CN-122020765-A - Product demand design auxiliary method, demand intelligent agent system and related devices

CN122020765ACN 122020765 ACN122020765 ACN 122020765ACN-122020765-A

Abstract

The application discloses a product demand design auxiliary method, a demand intelligent agent system and related devices, which relate to the technical field of product development and comprise the following steps: obtaining user interaction information, determining skill intention information corresponding to the user interaction information, planning tasks according to the skill intention information, and obtaining a plurality of planned task steps, wherein the plurality of task steps comprise target steps, the target steps are to obtain multi-source heterogeneous data related to the user interaction information from a tool platform preconfigured for a product demand design flow or other flows, the multi-source heterogeneous data is processed into a unified format, and the plurality of task steps are sequentially executed to obtain staged output data corresponding to specific sub-stages based on the multi-source heterogeneous data and the user interaction information. According to the application, multi-source heterogeneous data can be acquired from a plurality of tool platforms in the process of product demand design, so as to assist demand understanding, and the accuracy of staged output data is improved.

Inventors

  • Ye Manxia
  • HU RUI
  • HUANG YUCHEN

Assignees

  • 科大讯飞股份有限公司

Dates

Publication Date
20260512
Application Date
20260130

Claims (13)

  1. 1. A product demand design assistance method, comprising: Acquiring user interaction information, and determining skill intention information corresponding to the user interaction information, wherein the skill intention information is information describing skill actions of specific sub-stages in a product demand design flow, and the specific sub-stages are sub-stages pointed by the user interaction information; Task planning is carried out according to the skill intention information to obtain a plurality of task steps for planning, wherein the task steps comprise target steps, the target steps are to obtain multi-source heterogeneous data related to the user interaction information from a tool platform preconfigured for the product demand design flow or other flows, and the multi-source heterogeneous data are processed into a unified format; And sequentially executing the task steps to obtain the staged output data corresponding to the specific sub-stage based on the multi-source heterogeneous data and the user interaction information.
  2. 2. The product demand design assistance method of claim 1, wherein said processing said multi-source heterogeneous data into a unified format comprises: taking the data of the target type in the multi-source heterogeneous data as first data, and determining a target prompter of the first data; Analyzing and extracting the corresponding prompter data of the target prompter from the first data by using the configured large language model; And carrying out data cleaning on the prompter data and the second data, and rectifying the prompter data and the second data into data in a unified format, wherein the second data is other data except the first data in the multi-source heterogeneous data.
  3. 3. The product demand design assistance method of claim 1, wherein the determining skill intent information corresponding to the user interaction information comprises: if the user interaction information comprises the selected skills, generating skill intention information corresponding to the selected skills according to the user interaction information, and taking the skill intention information as the skill intention information corresponding to the user interaction information; And if the user interaction information does not comprise the selected skills, carrying out skill intention recognition according to the user interaction information to obtain skill intention information corresponding to the user interaction information.
  4. 4. The product demand design assistance method according to claim 3, wherein the performing skill intention recognition according to the user interaction information to obtain skill intention information corresponding to the user interaction information includes: extracting an entity from the user interaction information based on a preset named entity recognition strategy, extracting an entity relationship from the user interaction information based on a preset entity relationship extraction rule, and taking the extracted entity and entity relationship as target extraction information; Retrieving memory information related to the target extraction information from a pre-stored memory library, wherein the memory library stores the whole flow information of the product demand design process; and carrying out skill intention recognition in a preconfigured skill range according to the user interaction information and the memory information to obtain skill intention information corresponding to the user interaction information.
  5. 5. The product demand design assistance method according to claim 4, further comprising, after the skill intention recognition is performed based on the user interaction information and the memory information, obtaining skill intention information corresponding to the user interaction information: Acquiring user feedback information aiming at the skill intention information; Determining the confidence level of the skill intention information according to the user feedback information; If the confidence coefficient is larger than or equal to a preset confidence coefficient threshold value, executing the task planning according to the skill intention information; and if the confidence coefficient is smaller than the confidence coefficient threshold value, re-executing the skill intention information corresponding to the determined user interaction information, or outputting and displaying the skill intention information for manual auditing.
  6. 6. The product demand design assistance method according to claim 1, wherein the task planning according to the skill intention information, a plurality of task steps of planning are obtained, comprising: obtaining a preconfigured fine-grained intent type set comprising a plurality of fine-grained intent types, the fine-grained intent types being standardized task templates that can be mapped to specific operational flows; screening the fine grain intent type set for at least one fine grain intent type adapted by the skill intent information; and carrying out task planning according to the skill intention information and the at least one fine grain intention type to obtain the plurality of task steps, wherein the plurality of task steps comprise target steps and task steps which are in one-to-one correspondence with the at least one fine grain intention type.
  7. 7. The product demand design assistance method according to claim 1, further comprising: And updating the user interaction information, the staged output data and the service condition information of the tool platform to a memory bank, wherein the memory in the memory bank is also updated globally along with memory retention, and the memory retention is determined according to a time interval corresponding to the memory and a memory attenuation coefficient.
  8. 8. The product demand design assistance method according to claim 1, wherein the user interaction information includes natural language description instructions and demand design assistance data that a user explicitly inputs, the natural language description instructions being instructions for instructing generation of the staged production data or instructions for editing modification of the generated staged production data.
  9. 9. The product demand design assistance method of claim 1, wherein the staged production data includes a product demand document, a product prototype design drawing, and a visualized design assistance chart including a mind drawing and a timing diagram.
  10. 10. The demand agent system is characterized by comprising a one-stop workbench, a demand agent and a tool calling and adapting module; the one-stop workbench is used for responding to a user interaction event and sending user interaction information to the demand agent; The tool calling and adapting module is used for carrying out access configuration on a tool platform in the product demand design auxiliary method according to any one of claims 1-8, and calling the tool platform according to information of the access configuration so as to assist the demand agent to execute the product demand design auxiliary method according to any one of claims 1-8; the demand agent for executing the product demand design assisting method according to any one of claims 1 to 8.
  11. 11. A computer program product comprising computer readable instructions which, when run on an electronic device, cause the electronic device to implement the product demand design assistance method of any one of claims 1 to 9.
  12. 12. An electronic device comprising at least one processor and a memory coupled to the processor, wherein: the memory is used for storing a computer program; the processor is configured to execute the computer program to enable the electronic device to implement the product demand design assistance method according to any one of claims 1 to 9.
  13. 13. A computer storage medium carrying one or more computer programs which, when executed by an electronic device, enable the electronic device to implement the product demand design assistance method of any one of claims 1 to 9.

Description

Product demand design auxiliary method, demand intelligent agent system and related devices Technical Field The application relates to the technical field of product development, in particular to a product demand design auxiliary method, a demand intelligent agent system and a related device. Background In the original product demand design flow, a product manager and a demand analyst perform demand analysis according to communication information with clients to obtain a primary demand analysis result, prototype design is performed based on the primary demand analysis result, a product demand document (Product Requirements Document, PRD) is written, and materials such as a product prototype design drawing, the PRD and the like are delivered to research staff to perform demand disassembly and demand quality assessment. The whole process is highly dependent on manual operation, the labor cost is high, and the product design efficiency is low. In order to improve efficiency, at present, some single-point tools can be used for carrying out artificial intelligence assistance on part of sub-stages, such as using a project management tool to manage preliminary demand analysis results produced in a demand analysis sub-stage, using cloud documents to write PRDs, using tools such as document writing assistants and the like to carry out color rendering, expanding writing and summarizing on the written PRDs, using professional design software to draw product prototype design drawings and the like. However, to use individual single point tools, related data needs to be manually input into the single point tool in a copy, paste, and transfer manner, which is inefficient and prone to errors or lost context. Disclosure of Invention In view of the above problems, the present application provides a product demand design assisting method, a demand intelligent agent system and a related device, so as to solve the problems of low efficiency and easy error or context loss caused by manual copying, pasting and transferring of related data in each sub-stage in the prior art. The specific scheme is as follows: the first aspect of the present application provides a product demand design assisting method, including: Acquiring user interaction information, and determining skill intention information corresponding to the user interaction information, wherein the skill intention information is information describing skill actions of specific sub-stages in a product demand design flow, and the specific sub-stages are sub-stages pointed by the user interaction information; Task planning is carried out according to the skill intention information to obtain a plurality of task steps for planning, wherein the task steps comprise target steps, the target steps are to obtain multi-source heterogeneous data related to the user interaction information from a tool platform preconfigured for the product demand design flow or other flows, and the multi-source heterogeneous data are processed into a unified format; And sequentially executing the task steps to obtain the staged output data corresponding to the specific sub-stage based on the multi-source heterogeneous data and the user interaction information. In one possible implementation, the processing the multi-source heterogeneous data into a unified format includes: taking the data of the target type in the multi-source heterogeneous data as first data, and determining a target prompter of the first data; Analyzing and extracting the corresponding prompter data of the target prompter from the first data by using the configured large language model; And carrying out data cleaning on the prompter data and the second data, and rectifying the prompter data and the second data into data in a unified format, wherein the second data is other data except the first data in the multi-source heterogeneous data. In one possible implementation, the determining skill intent information corresponding to the user interaction information includes: if the user interaction information comprises the selected skills, generating skill intention information corresponding to the selected skills according to the user interaction information, and taking the skill intention information as the skill intention information corresponding to the user interaction information; And if the user interaction information does not comprise the selected skills, carrying out skill intention recognition according to the user interaction information to obtain skill intention information corresponding to the user interaction information. In one possible implementation, the performing skill intention recognition according to the user interaction information to obtain skill intention information corresponding to the user interaction information includes: extracting an entity from the user interaction information based on a preset named entity recognition strategy, extracting an entity relationship from the user interaction information base