CN-121981749-A - Marketing content generation method based on multi-agent cooperation and related equipment

CN121981749ACN 121981749 ACN121981749 ACN 121981749ACN-121981749-A

Abstract

The application discloses a marketing content generation method based on multi-agent cooperation and related equipment, wherein the method comprises the steps of determining a target business scene according to content generation demand information and content generation reference information, and further obtaining a corresponding target scene data packet from a marketing scene library; and calling a target agent plug-in to execute the generating task according to the target agent list and the scheduling plan to obtain a stage content generating result and checking and accepting the stage content generating result to obtain a final content generating result. The embodiment of the application can realize full-flow automation from scene recognition, task disassembly, task arrangement, multi-agent execution generation task to content acceptance, and can effectively improve the efficiency, quality and stability of marketing content generation. The application can be widely applied to the technical field of the generation type artificial intelligence.

Inventors

LIU ZHEN
XIE DONGFANG
YAO JUNYI

Assignees

广东因赛品牌营销集团股份有限公司

Dates

Publication Date: 20260505
Application Date: 20251226

Claims (10)

1. A marketing content generation method based on multi-agent cooperation is characterized by comprising the following steps: acquiring content generation requirement information, content generation reference information, available material information and constraint information; Determining a target service scene according to the content generation demand information and the content generation reference information, and acquiring a corresponding target scene data packet from a preset marketing scene library according to the target service scene; determining a target agent list and a task list according to the target scene data packet; Generating a task directed acyclic graph according to the target scene data packet, the task list, the available material information and the constraint information; Generating an arrangement plan according to the task directed acyclic graph and the constraint information; Calling a plurality of target intelligent agent plug-ins to execute a generating task according to the target intelligent agent list and the arrangement plan, and obtaining a stage content generating result; and checking and accepting the generation result of the stage content according to the target scene data packet to obtain a final content generation result.
2. The method for generating marketing content based on multi-agent collaboration according to claim 1, wherein the content generation reference information includes user portrait information and history item information, and the determining a target business scenario according to the content generation requirement information and the content generation reference information comprises: Carrying out semantic alignment processing on the content generation requirement information and the content generation reference information to obtain a first semantic alignment result; Inputting the first semantic alignment result into a preset scene classifier for processing to obtain the target service scene; The target business scene is any one of a brand marketing scene, an effect marketing scene and an e-commerce marketing scene.
3. The marketing content generation method based on multi-agent collaboration according to claim 1, wherein the target scenario data package comprises a workflow template, and the determining a target agent list and a task list according to the target scenario data package comprises: determining agent demand information and task demand information according to the workflow template; matching from a preset agent library according to the agent demand information to obtain the target agent list; and generating the task list according to the task demand information.
4. The method for generating marketing content based on multi-agent collaboration according to claim 1, wherein the generating a task directed acyclic graph from the target scene data packet, the task list, the available material information, and the constraint information comprises: carrying out semantic alignment processing on the target scene data packet, the task list, the available material information and the constraint information to obtain a second semantic alignment result; Performing executable analysis and dependency analysis on the task list through a preset task decomposer to obtain a task connection relation analysis result and a plurality of task nodes; And generating the task directed acyclic graph according to the second semantic alignment result, the task connection relation analysis result and each task node.
5. The marketing content generation method based on multi-agent collaboration according to claim 1, wherein the constraint information includes age constraint information and resource constraint information, the generating a scheduling plan based on the task directed acyclic graph and the constraint information comprises: analyzing the parallelism of the task to be executed according to the task directed acyclic graph to obtain a parallelism analysis result; determining a task execution sequence of the task to be executed according to the aging constraint information, the resource constraint information and the parallelism analysis result; And generating the arrangement plan according to the task execution sequence.
6. The marketing content generation method based on multi-agent collaboration according to claim 1, wherein the step of calling a plurality of target agent plug-ins to execute generation tasks according to the target agent list and the scheduling plan to obtain a phase content generation result comprises the steps of: Confirming each target intelligent agent plug-in required for executing the generating task according to the target intelligent agent list, and binding each target intelligent agent to a pluggable intelligent agent bus; And scheduling each target intelligent agent plug-in to execute the generating task according to the scheduling plan through the pluggable intelligent agent bus to obtain the stage content generating result.
7. The marketing content generation method based on multi-agent collaboration according to claim 1, wherein the target scene data packet includes acceptance rule data, and the accepting the stage content generation result according to the target scene data packet to obtain a final content generation result includes: judging whether the generation result of the stage content meets the preset acceptance condition according to the acceptance rule data; when the generation result of the stage content does not meet the preset acceptance condition, returning to a check point in the execution process of the generation task, and regenerating the generation result of the stage content; And when the stage content generation result meets the preset acceptance condition, taking the stage content generation result as the final content generation result.
8. A marketing content generating device based on multi-agent collaboration, the device comprising: The information acquisition module is used for acquiring content generation requirement information, content generation reference information, available material information and constraint information; the target scene data packet acquisition module is used for determining a target service scene according to the content generation demand information and the content generation reference information and acquiring a corresponding target scene data packet from a preset marketing scene library according to the target service scene; The target intelligent agent list and task list generation module is used for determining a target intelligent agent list and a task list according to the target scene data packet; The task directed acyclic graph generation module is used for generating a task directed acyclic graph according to the target scene data packet, the task list, the available material information and the constraint information; The arrangement plan generation module is used for generating an arrangement plan according to the task directed acyclic graph and the constraint information; the multi-agent cooperation generating module is used for calling a plurality of target agent plug-ins to execute generating tasks according to the target agent list and the arrangement plan to obtain a stage content generating result; and the generation result acceptance module is used for accepting the generation result of the stage content according to the target scene data packet to obtain a final content generation result.
9. An electronic device comprising a memory storing a computer program and a processor implementing the method of any of claims 1 to 7 when the computer program is executed by the processor.
10. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the method of any one of claims 1 to 7.

Description

Marketing content generation method based on multi-agent cooperation and related equipment Technical Field The application relates to the technical field of generation type artificial intelligence, in particular to a marketing content generation method based on multi-agent cooperation and related equipment. Background With the development of generated artificial intelligence, automated content generation technology based on large models and agents has been widely applied to the marketing field. In the prior art, workflow templates and tool chains are respectively configured according to marketing scenes of different types such as brand propaganda, effect propaganda and electronic commerce popularization, corresponding workflow templates and agents participating in generating tasks are selected after manual understanding of requirements, task steps are manually adjusted, when different types of service requirements are met, the current service scene is required to be manually judged again, the steps are adjusted, appropriate tools are selected to participate in execution, the multiplexing degree among different projects is low, the efficiency is low, the flow configuration and the agent invoking lack of unified standards, quality inspection and acceptance are dependent on manual inspection, the influence of subjective factors is large, and the quality and style of generated content are unstable. In summary, the technical problems in the related art are to be improved. Disclosure of Invention The embodiment of the application mainly aims to provide a marketing content generation method based on multi-agent cooperation and related equipment, which can realize full-flow automation from scene recognition, task disassembly, task arrangement, multi-agent execution generation task to content acceptance, and can effectively improve the efficiency, quality and stability of marketing content generation. In order to achieve the above object, an aspect of an embodiment of the present application provides a marketing content generation method based on multi-agent collaboration, the method including: acquiring content generation requirement information, content generation reference information, available material information and constraint information; Determining a target service scene according to the content generation demand information and the content generation reference information, and acquiring a corresponding target scene data packet from a preset marketing scene library according to the target service scene; determining a target agent list and a task list according to the target scene data packet; Generating a task directed acyclic graph according to the target scene data packet, the task list, the available material information and the constraint information; Generating an arrangement plan according to the task directed acyclic graph and the constraint information; Calling a plurality of target intelligent agent plug-ins to execute a generating task according to the target intelligent agent list and the arrangement plan, and obtaining a stage content generating result; and checking and accepting the generation result of the stage content according to the target scene data packet to obtain a final content generation result. In some embodiments, the content generation reference information includes user portrait information and historical item information, and the determining a target business scenario according to the content generation requirement information and the content generation reference information includes: Carrying out semantic alignment processing on the content generation requirement information and the content generation reference information to obtain a first semantic alignment result; Inputting the first semantic alignment result into a preset scene classifier for processing to obtain the target service scene; The target business scene is any one of a brand marketing scene, an effect marketing scene and an e-commerce marketing scene. In some embodiments, the target scenario data package includes a workflow template, and the determining a target agent list and a task list according to the target scenario data package includes: determining agent demand information and task demand information according to the workflow template; matching from a preset agent library according to the agent demand information to obtain the target agent list; and generating the task list according to the task demand information. In some embodiments, the generating a task directed acyclic graph according to the target scene data packet, the task list, the available material information, and the constraint information includes: carrying out semantic alignment processing on the target scene data packet, the task list, the available material information and the constraint information to obtain a second semantic alignment result; Performing executable analysis and dependency analysis on the task list through a preset task d