Product demand management method and system based on large model agent

Abstract

The invention relates to the technical field of artificial intelligence, in particular to a product demand management method and system based on a large model agent, and the method comprises the following steps: multi-source demand information collection and injection, demand intelligent analysis and structuring, demand collaboration and refinement, demand priority intelligent evaluation, demand tracking and influence analysis, and closed-loop feedback and self-evolution. The method has the beneficial effects that a large amount of repeated labor (such as information extraction, document writing and state tracking) is automatically processed, so that a product manager can focus on high-value decision; and through standardization processing of the LLM intelligent agent, demand ambiguity is reduced, and a unified demand language is established.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, specifically to a product demand management method and system based on a large-scale intelligent agent model. Background Technology

[0002] With the accelerated shift from "Internet+" to "AI+", enterprises are deepening their digital transformation, and artificial intelligence technology has gradually become a key driver for empowering enterprise development and enhancing competitiveness. Against this backdrop, enterprises are increasingly demanding efficient and intelligent management tools.

[0003] However, currently, enterprises still generally manage product requirements by product dimension. This traditional model lacks a unified channel for collecting user needs and efficient automated distribution tools, resulting in high management costs while failing to improve work efficiency. Faced with increasingly diverse and personalized user needs, traditional management methods are unable to meet the rapidly changing demands of the market.

[0004] In this context, how to accurately collect, efficiently process, and rationally distribute product demands through intelligent means has become a crucial issue that enterprises urgently need to address. This not only relates to an enterprise's ability to respond quickly to market changes but also directly impacts its product's market competitiveness and customer satisfaction. Therefore, building intelligent and systematic product demand management methods and systems has become a critical task that enterprises cannot ignore during their digital transformation. Summary of the Invention

[0005] The purpose of this invention is to provide a product demand management method and system based on a large model intelligent agent, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a product demand management method based on a large model intelligent agent, comprising the following steps: Multi-source demand information collection and injection steps: Collect raw demand information from multiple heterogeneous data sources and inject it into the demand data pool; The intelligent parsing and structuring steps for demand are as follows: The large language model agent parses the original demand information and stores the parsing results in a structured form in the demand knowledge graph; Requirements coordination and refinement steps: The large language model intelligent agent, based on the requirements knowledge graph, assists in requirements coordination and refinement; Intelligent evaluation steps for demand priority: The large language model agent constructs a multi-dimensional evaluation model, automatically calculates and outputs the priority score and ranking of demands; Requirements tracking and impact analysis steps: When the status changes or requirements change in the development process, the large language model agent automatically updates the requirement status and analyzes the potential impact. Closed-loop feedback and self-evolution steps: Collect feedback data after the requirements are launched to optimize the evaluation model of the large language model agent.

[0007] Preferably, the large language model agent identifies and extracts demand entities, classifies demands, divides demand granularity, quantifies demand sentiment, and identifies demand conflicts and repetitions.

[0008] Preferably, the nodes of the demand knowledge graph include user roles, demand functions, business value, business scenarios, and product modules.

[0009] Preferably, the evaluation dimensions in the demand priority intelligent assessment step include business value, technical implementation cost, user impact, strategic alignment, and risk.

[0010] Preferably, large-scale intelligent agents are given the perspective of product experts, technical experts, or project managers to perform tasks.

[0011] A product demand management system based on a large-scale intelligent agent, employing the method described in claim 5, includes: The multi-source acquisition module is used to collect raw demand information from multiple heterogeneous data sources; A demand data pool is used to store the original demand information; The core module of the LLM intelligent agent includes a large language model, which is used to perform tasks such as parsing, coordinating, evaluating and tracking requirements; The requirement knowledge graph module is used to store and manage structured requirement information and their relationships in a graph structure. Priority evaluation engine, used to calculate priority scores for requirements; Workflow integration module for integration with external project management tools, code hosting platforms, and communication tools; The Feedback and Evolution module is used to collect feedback data and optimize the LLM agent.

[0012] Preferably, the workflow integration module is integrated with Jira and GitLab tools via API interfaces.

[0013] Preferably, the feedback and evolution module optimizes the model parameters or prompt word strategy of the LLM agent through online learning or periodic fine-tuning.

[0014] Compared with the prior art, the beneficial effects of the present invention are: The product requirement management method and system proposed in this invention, based on large model intelligent agents, automates a large amount of repetitive work (such as information extraction, document writing, and status tracking), enabling product managers to focus on high-value decisions. Through standardized processing by LLM intelligent agents, requirement ambiguity is reduced, and a unified requirement language is established. A requirement knowledge graph visually displays the requirement context and influencing relationships, achieving end-to-end traceability. Quantitative and interpretable prioritization is provided, reducing subjective judgment bias and making resource allocation more rational. Through a closed-loop feedback loop, the system can continuously learn from historical decision results, constantly optimizing its evaluation and decision-making capabilities. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the present invention clear and complete, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only some, not all, embodiments of the present invention, and are merely illustrative of the embodiments of the present invention. They are not intended to limit the embodiments of the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Example 1: This invention provides a technical solution: a product demand management method based on a large-scale intelligent agent, comprising the following steps: Step 1 (Collection and Injection of User Requirement Information): Collect raw requirement information from multiple heterogeneous data sources in real time or periodically; the data sources include user feedback platforms, internal collaboration documents, meeting minutes systems, code repositories, and work order systems; use connectors to inject the raw requirement information into the requirement data pool.

[0018] Step 2 (Intelligent Demand Parsing and Structured Processing): The large language model agent parses the raw demand information in the demand data pool; the parsing includes: identifying and extracting demand entities, classifying demands, dividing demand granularity, quantifying demand sentiment, and identifying demand conflicts and repetitions; the parsed results are stored in a structured form in a demand knowledge graph; the nodes of the demand knowledge graph include, but are not limited to, user roles, demand functions, business value, business scenarios, and product modules.

[0019] Step 3 (Requirement Collaboration and Refinement): The large language model intelligent agent assists product managers in requirement collaboration and refinement based on the requirement knowledge graph; including: automatically generating standardized requirement description documents (user stories / PRDs), initiating clarification dialogues with relevant personnel for ambiguous requirements, and conducting preliminary workload estimation and feasibility analysis of requirements based on historical data.

[0020] Step 4 (Intelligent Assessment of Requirement Priority): The large language model agent constructs a multi-dimensional assessment model; the dimensions include business value, technical implementation cost, user impact, strategic alignment, and risk; based on historical data and real-time market data in the requirement knowledge graph, the priority score and ranking of each requirement are automatically calculated, and an interpretable reasoning chain is provided for the ranking results.

[0021] Step 5 (Requirements Tracking and Impact Analysis): When the status of a task changes or code is committed during the development process, the large language model agent automatically updates the status of the corresponding requirement in the requirement knowledge graph; when a requirement change request occurs, the agent automatically analyzes the potential impact of the change on upstream and downstream related requirements, tasks, and code modules, and generates an impact analysis report.

[0022] Step 6 (Closed-loop feedback and self-evolution): Collect user feedback and system operation data after the requirements are launched, and inject them into the requirements data pool as new original requirements information; the large language model agent analyzes the difference between the actual value and expected value of the requirements, and uses it to optimize the future requirements assessment model, forming a closed-loop self-evolution system.

[0023] Example 2, based on Example 1, proposes a product requirement management system based on a large-scale intelligent agent, including: Multi-source acquisition module: Configured to collect raw demand information from multiple heterogeneous data sources in real time or periodically; Demand Data Pool: Configured to store the collected raw demand information; LLM Agent Core Module: As the core of the system, it includes at least one large language model and its prompt word engineering framework; the agent is given multiple role perspectives such as product expert, technical expert, and project manager, and is used to perform tasks such as requirement analysis, coordination, evaluation, and tracking. Demand Knowledge Graph Module: Configured for storing and managing structured demand information, entities, and their relationships in a graph structure after being parsed by the intelligent agent; Priority evaluation engine: It is configured with configurable evaluation dimensions and weight models, works in collaboration with the LLM agent, and calculates the priority score of the requirements. Workflow Integration Module: Configured for integration with external project management tools, code hosting platforms, and communication tools to achieve status synchronization and message delivery; Feedback and Evolution Module: Configured to collect online feedback data and use this data to fine-tune and optimize the evaluation model of the LLM agent.

[0024] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A product demand management method based on a large-scale intelligent agent, characterized in that: Includes the following steps: Multi-source demand information collection and injection steps: Collect raw demand information from multiple heterogeneous data sources and inject it into the demand data pool; The intelligent parsing and structuring steps for demand are as follows: The large language model agent parses the original demand information and stores the parsing results in a structured form in the demand knowledge graph; Requirements coordination and refinement steps: The large language model intelligent agent, based on the requirements knowledge graph, assists in requirements coordination and refinement; Intelligent evaluation steps for demand priority: The large language model agent constructs a multi-dimensional evaluation model, automatically calculates and outputs the priority score and ranking of demands; Requirements tracking and impact analysis steps: When the status changes or requirements change in the development process, the large language model agent automatically updates the requirement status and analyzes the potential impact. Closed-loop feedback and self-evolution steps: Collect feedback data after the requirements are launched to optimize the evaluation model of the large language model agent.

2. The product demand management method based on a large model intelligent agent according to claim 1, characterized in that: The large language model agent identifies and extracts demand entities, classifies demands, divides demand granularity, quantifies demand sentiment, and identifies demand conflicts and repetitions.

3. The product demand management method based on a large model intelligent agent according to claim 2, characterized in that: The nodes of the demand knowledge graph include user roles, demand functions, business value, business scenarios, and product modules.

4. The product demand management method based on a large model intelligent agent according to claim 1, characterized in that: The evaluation dimensions in the intelligent assessment of demand priorities include business value, technical implementation cost, user impact, strategic alignment, and risk.

5. The product demand management method based on a large-scale intelligent agent according to claim 1, characterized in that: Large-scale intelligent agents are given the perspective of product experts, technical experts, or project managers to perform tasks.

6. A product demand management system based on a large-scale intelligent agent, employing the method described in claim 5, characterized in that: include: The multi-source acquisition module is used to collect raw demand information from multiple heterogeneous data sources; A demand data pool is used to store the original demand information; The core module of the LLM intelligent agent includes a large language model, which is used to perform tasks such as parsing, coordinating, evaluating and tracking requirements; The requirement knowledge graph module is used to store and manage structured requirement information and their relationships in a graph structure. Priority evaluation engine, used to calculate priority scores for requirements; Workflow integration module for integration with external project management tools, code hosting platforms, and communication tools; The Feedback and Evolution module is used to collect feedback data and optimize the LLM agent.

7. A product demand management system based on a large-scale intelligent agent according to claim 6, characterized in that: The workflow integration module integrates with Jira and GitLab tools via API interfaces.

8. A product demand management system based on a large-scale intelligent agent according to claim 7, characterized in that: The feedback and evolution module optimizes the model parameters or prompt word strategy of the LLM agent through online learning or periodic fine-tuning.

Images