Financial service decision-making method and device, electronic equipment and storage medium

Abstract

The invention discloses a financial service decision-making method and device, electronic equipment and a storage medium, and relates to the field of artificial intelligence or other related technical fields, and the method comprises the steps: receiving a financial service request of a user side, and carrying out the analysis of the financial service request, and obtaining the service information and the service type; calling N artificial intelligence agents based on the service type, and combining the N artificial intelligence agents to obtain an artificial intelligence decision body; inputting the service information into an artificial intelligence decision-making body, and performing service decision-making through the artificial intelligence decision-making body to obtain a service decision-making result; and obtaining a confidence value of the intelligent decision body, pushing the service decision result to the service terminal for correction when the confidence value is smaller than a preset confidence threshold, and executing the corrected service decision result. According to the method and the device, the technical problem of relatively low service decision accuracy caused by adopting a rule-based single artificial intelligence model to assist in service decision in related technologies is solved.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence or other related technical fields, and more specifically, to a financial business decision-making method and apparatus, electronic device and storage medium. Background Technology

[0002] In recent years, with the popularization of internet finance and the continuous innovation of financial technology, the volume of financial business has shown an explosive growth trend. This trend is not only reflected in the number of daily transactions, but also extends to multiple business areas such as customer consultation, product customization, loan applications, and risk management. Especially in the context of globalization and digitalization, financial business processing faces unprecedented challenges: a massive amount of data floods into financial systems every day, including transaction records, customer information, and market data. This data is not only enormous in quantity but also diverse in structure, including both structured and unstructured data, such as images, voice, and natural language text. For such a large-scale and complex dataset, traditional manual processing methods are far from sufficient to meet the requirements of timeliness and accuracy. As consumers' demands for financial services increase, financial institutions must be able to provide more personalized products and services. This means that it is necessary to deeply and quickly analyze user needs and provide corresponding financial services to users.

[0003] In related technologies, the use of a single rule-based AI model to assist business decision-making results in subsystems such as credit assessment and anomaly detection operating independently, leading to data silos and an inability to share cross-business knowledge. Intelligent interaction models only support text-based question-and-answer, lacking multimodal understanding of user needs, resulting in low accuracy in business decisions.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This invention provides a financial business decision-making method, apparatus, electronic device, and storage medium to at least address the technical problem in related technologies where the accuracy of business decisions is low when using a single rule-based artificial intelligence model for assistance.

[0006] According to one aspect of the present invention, a financial business decision-making method is provided, comprising: receiving a financial business request from a user terminal and parsing the financial business request to obtain business information and a business type; invoking N artificial intelligence agents based on the business type and combining the N artificial intelligence agents to obtain an artificial intelligence decision-making body, wherein N is a positive integer; inputting the business information into the artificial intelligence decision-making body, performing a business decision through the artificial intelligence decision-making body, and obtaining a business decision result; obtaining a confidence value of the artificial intelligence decision-making body, and if the confidence value is less than a preset confidence threshold, pushing the business decision result to a business terminal for correction, and executing the corrected business decision result.

[0007] Furthermore, before receiving financial service requests from users, the process includes: constructing AI agents for different business scenarios, wherein the AI ​​agents are configured with domain knowledge graphs and adaptive decision-making mechanisms; configuring business role tags for the AI ​​agents; and establishing a mapping relationship between the AI ​​agents and the business types based on the business role tags.

[0008] Furthermore, the step of invoking N AI agents based on the business type and combining the N AI agents to obtain an AI decision-making body includes: when the business type is a loan business, invoking an identity verification agent, a credit assessment agent, and a loan anomaly detection agent; and combining the identity verification agent, the credit assessment agent, and the loan anomaly detection agent to obtain the AI ​​decision-making body for the loan business.

[0009] Furthermore, the step of invoking N AI agents based on the business type and combining the N AI agents to obtain an AI decision-making body includes: when the business type is user asset management business, invoking a perception agent, a compliance review agent, and an asset allocation agent; and combining the perception agent, the compliance review agent, and the asset allocation agent to obtain the AI ​​decision-making body for the user asset management business.

[0010] Further, the steps of inputting the business information into the AI ​​decision-making body and making business decisions through the AI ​​decision-making body to obtain business decision results include: when the business type is a loan business, inputting the business information into the AI ​​decision-making body for the loan business; extracting the text data and voice data of the business information through the identity verification agent, and verifying the consistency of the content of the text data and the voice data to obtain a verification result; conducting a credit assessment on the target user through the credit assessment agent to obtain a credit assessment result; conducting anomaly detection on the loan business of the target user through the loan anomaly detection agent to obtain a loan anomaly detection result; and obtaining a loan approval result based on the verification result, the credit assessment result, and the loan anomaly detection result.

[0011] Furthermore, the steps of inputting the business information into the AI ​​decision-making body and making business decisions through the AI ​​decision-making body to obtain business decision results include: when the business type is user asset management business, obtaining a user profile of the target user; inputting the business information and the user profile into the AI ​​decision-making body of the user asset management business; monitoring fluctuation events through the perception agent and identifying non-compliant events through the compliance review agent; and making asset allocation decisions through the asset allocation agent by combining the user profile, the business information, the fluctuation events, and the non-compliant events to generate an asset allocation plan.

[0012] Furthermore, before obtaining the user profile of the target user, the method further includes: collecting multi-source data of the target user, wherein the multi-source data includes at least one of the following: financial interaction data and user behavior data; inputting the multi-source data into a feature extraction model, and extracting the sentiment tendency features and preference features of the target user through the feature extraction model, wherein the feature extraction model is a pre-trained machine learning model for feature extraction; and updating the user profile of the target user based on the sentiment tendency features and the preference features.

[0013] According to another aspect of the present invention, a financial business decision-making device is also provided, comprising: a receiving unit, configured to receive a financial business request from a user terminal and parse the financial business request to obtain business information and a business type; an invocation unit, configured to invoke N artificial intelligence agents based on the business type and combine the N artificial intelligence agents to obtain an artificial intelligence decision-making body, wherein N is a positive integer; a decision-making unit, configured to input the business information into the artificial intelligence decision-making body, perform business decision-making through the artificial intelligence decision-making body, and obtain a business decision result; and a correction unit, configured to obtain a confidence value of the artificial intelligence decision-making body, and if the confidence value is less than a preset confidence threshold, push the business decision result to a business terminal for correction and execute the corrected business decision result.

[0014] Furthermore, the financial business decision-making device further includes: a first construction module for constructing AI agents under different business scenarios, wherein the AI ​​agents are configured with domain knowledge graphs and adaptive decision-making mechanisms; a first configuration module for configuring business role tags for the AI ​​agents; and a first establishment module for establishing a mapping relationship between the AI ​​agents and the business types based on the business role tags.

[0015] Furthermore, the invocation unit includes: a first invocation module, used to invoke an identity verification agent, a credit assessment agent, and a loan anomaly detection agent when the business type is a loan business; and a first combination module, used to combine the identity verification agent, the credit assessment agent, and the loan anomaly detection agent to obtain an artificial intelligence decision-making body for the loan business.

[0016] Furthermore, the invocation unit also includes: a second invocation module, used to invoke the perception agent, compliance review agent, and asset allocation agent when the business type is user asset management business; and a second combination module, used to combine the perception agent, the compliance review agent, and the asset allocation agent to obtain the artificial intelligence decision-making body for the user asset management business.

[0017] Further, the decision-making unit includes: a first input module, used to input the business information into the AI ​​decision-making body of the loan business when the business type is a loan business; a first verification module, used to extract text data and voice data of the business information through the identity verification agent, and to perform consistency verification on the content of the text data and the voice data to obtain a verification result; a first evaluation module, used to perform credit evaluation on the target user through the credit evaluation agent to obtain a credit evaluation result; a first detection module, used to perform anomaly detection on the loan business of the target user through the loan anomaly detection agent to obtain a loan anomaly detection result; and a first acquisition module, used to obtain a loan approval result based on the verification result, the credit evaluation result, and the loan anomaly detection result.

[0018] Furthermore, the decision-making unit further includes: a second acquisition module, used to acquire a user profile of the target user when the business type is user asset management business; a second input module, used to input the business information and the user profile into the artificial intelligence decision-making body of the user asset management business; a first identification module, used to monitor fluctuation events through the perception agent and identify non-compliant events through the compliance review agent; and a first decision module, used to make asset allocation decisions and generate an asset allocation plan by combining the user profile, the business information, the fluctuation events, and the non-compliant events through the asset allocation agent.

[0019] Furthermore, the financial business decision-making device further includes: a first acquisition module, used to acquire multi-source data of the target user, wherein the multi-source data includes at least one of the following: financial interaction data and user behavior data; a first extraction module, used to input the multi-source data into a feature extraction model, and extract the target user's sentiment tendency features and preference features through the feature extraction model, wherein the feature extraction model is a pre-trained machine learning model for feature extraction; and a first update module, used to update the target user's user profile based on the sentiment tendency features and the preference features.

[0020] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform any of the above-described financial business decision-making methods.

[0021] According to another aspect of the present invention, an electronic device is also provided, including one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement any of the above-described financial business decision-making methods.

[0022] According to another aspect of the present invention, a computer program product is also provided, the computer program product comprising a computer program, wherein the computer program, when executed by a processor, implements any of the above-described financial business decision-making methods.

[0023] In this application, the following steps are performed: receiving financial business requests from users and parsing the requests to obtain business information and business type; calling N artificial intelligence agents based on the business type and combining the N agents to obtain an artificial intelligence decision-making body, where N is a positive integer; inputting the business information into the artificial intelligence decision-making body; making business decisions through the artificial intelligence decision-making body and obtaining the business decision result; finally obtaining the confidence value of the intelligent decision-making body; if the confidence value is less than a preset confidence threshold, pushing the business decision result to the business terminal for correction and executing the corrected business decision result.

[0024] This application delves into the financial service requests sent by users and dynamically combines multiple AI agents with different functions based on the analyzed service information to form an AI decision-making body. This multi-service AI agent-based intelligent decision-making breaks down the barriers between traditional task models, promotes cross-business knowledge sharing and application, solves the problem of data silos, and improves the accuracy of intelligent decision-making. Furthermore, it addresses the technical problem in related technologies where using a single rule-based AI model to assist in business decision-making results in low accuracy. Attached Figure Description

[0025] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0026] Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing financial business decision-making methods is shown.

[0027] Figure 2 This is a flowchart of an optional financial business decision-making method according to an embodiment of the present invention;

[0028] Figure 3 This is a schematic diagram of an optional financial business decision-making device according to an embodiment of the present invention;

[0029] Figure 4 This is a hardware structure block diagram of an electronic device (or mobile device) for executing a financial business decision-making method according to an embodiment of the present invention. Detailed Implementation

[0030] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0032] It should be noted that the financial business decision-making method and apparatus in this application can be used in the field of artificial intelligence to make decisions on financial business based on artificial intelligence, or in any field other than artificial intelligence to make decisions on financial business based on artificial intelligence. This application does not limit the application field of the financial business decision-making method and apparatus.

[0033] It should be noted that the information collected in this application (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for display, data used for analysis, etc.) are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of this data all comply with relevant laws, regulations, and standards, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding access points are provided for users to choose to authorize or refuse. For example, interfaces are set up between this system and relevant users or organizations, providing users with corresponding access points to choose to agree to or refuse automated decision-making results; if the user chooses to refuse, the process proceeds to the expert decision-making stage.

[0034] The following embodiments of the present invention can be applied to various financial business decision-making systems / applications / devices. The present invention, through adaptive intelligent decision-making based on multimodal artificial intelligence agents, can be applied to various scenarios such as credit approval, customer service, risk management, and personalized financial product recommendations in financial institutions. This addresses the shortcomings of traditional automated systems, such as data silos preventing cross-business knowledge application and the inability to process multimodal data, thereby improving the accuracy of intelligent decision-making.

[0035] The present invention will now be described in detail with reference to various embodiments.

[0036] Example 1

[0037] According to an embodiment of the present invention, an embodiment of a financial business decision-making method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0038] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware block diagram of a computer terminal (or mobile device) for implementing financial business decision-making methods is shown. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more processors 102 (shown as 102a, 102b, ..., 102n in the figure) 102 (processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0039] It should be noted that the aforementioned one or more processors 102 and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10 (or mobile device). As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0040] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the financial business decision-making method in this embodiment. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned financial business decision-making method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0041] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0042] The display may be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0043] Under the aforementioned operating environment, this application provides the following: Figure 2 The financial business decision-making method shown is implemented by the financial business decision-making system.

[0044] Figure 2 This is a flowchart of an optional financial business decision-making method according to an embodiment of the present invention, such as... Figure 2 As shown, the method includes the following steps:

[0045] Furthermore, before receiving financial business requests from users, the process includes: constructing AI agents for different business scenarios, wherein the AI ​​agents are configured with domain knowledge graphs and adaptive decision-making mechanisms; configuring business role tags for the AI ​​agents; and establishing a mapping relationship between the AI ​​agents and business types based on the business role tags.

[0046] In one optional embodiment, this application introduces a dynamic combination AI agent mechanism. Multiple AI agents with different functions are pre-configured, and in actual business scenarios, these agents are dynamically combined according to business needs to construct an AI decision-making body. Specifically, firstly, corresponding AI agents (AI agents) are designed and trained for typical scenarios in financial business, such as credit approval, customer service, and risk management. Each AI agent is equipped with a domain knowledge graph, a network structure containing rich industry knowledge, compliance rules, historical cases, and product information, enabling the AI ​​agent to understand the business context and make knowledge-based decisions. Furthermore, the AI ​​agent integrates an adaptive decision-making mechanism based on deep learning and reinforcement learning technologies, allowing the AI ​​agent to continuously learn and optimize its decision-making strategies during actual operation to adapt to the ever-changing business environment and customer needs.

[0047] To differentiate the functions and expertise of each AI agent, a clear business role label was assigned to each AI agent. These labels may include credit reviewer, anti-fraud specialist, financial advisor, etc., ensuring that each AI agent can focus on a specific business area and improve the professionalism and relevance of decision-making.

[0048] Based on the business role tags of AI agents, a mapping relationship between AI agents and business types has been established. This means that whenever a specific type of business request is received, the system can quickly identify and invoke the most suitable combination of AI agents to form an AI decision-making body to execute the corresponding business decisions. For example, upon receiving a loan application, the system will automatically activate AI agents related to credit approval, such as identity verification agents and credit assessment agents. Through the collaborative work of different AI agents, a comprehensive loan eligibility review can be achieved.

[0049] Through the steps described above, each AI agent is assigned a specific business role label and equipped with a domain knowledge graph. When AI agents are dynamically combined, they not only automatically collaborate based on their labels but also share information from the knowledge graph, enabling cross-business domain knowledge and data flow and effectively solving the data silo problem in the financial sector.

[0050] Step S201: Receive financial service requests from the user terminal, parse the financial service requests, and obtain service information and service type.

[0051] In step S201 above, financial service requests from users are received. These requests can be of various types, such as online loan applications, wealth management consultations, and transaction anomaly inquiries, and typically include multimodal data such as spreadsheets, documents, voice messages, and video materials submitted by the user. For example, a user may submit a loan application through a mobile application, which includes a personal income verification document and a voice explanation of the loan's purpose.

[0052] Upon receiving a financial transaction request, the system triggers a data parsing process to identify and extract key information from the request. This step involves using advanced natural language processing, speech recognition, optical character recognition, and image processing technologies to perform in-depth analysis of text, speech, images, and time-series data. For example, optical character recognition technology is used to extract text content from documents; speech recognition and natural language processing technologies are used to understand the description of the loan purpose in the user's voice message, ensuring that the information is accurately understood and recorded by the system.

[0053] After data parsing, the system automatically determines the specific type of business request based on the extracted information, such as personal loan applications or wealth management, and further refines the business information, such as loan amount, repayment period, investment preferences, and risk tolerance. This process ensures that the system can accurately identify business needs, providing a basis for subsequent decision-making.

[0054] Step S202: Invoke N AI agents based on the business type, and combine the N AI agents to obtain an AI decision-making body.

[0055] In step S202 above, a set of pre-configured artificial intelligence agents (AI agents) is automatically selected based on the determined business type. Each AI agent has a specific business role label, such as credit reviewer, compliance officer, or financial advisor. These labels clarify the functional positioning and professional field of the AI ​​agent. For example, for a loan application, the system will call the identity verification agent, credit assessment agent, and anomaly detection agent, etc. The selected N AI agents (N is a positive integer representing the number of agents participating in the decision-making) are integrated to form a collaborative AI decision-making body. This decision-making body can combine the expertise and capabilities of each AI agent to comprehensively analyze complex multimodal business information. For example, the identity verification agent is responsible for verifying the customer's identity information, the credit assessment agent gives a score based on the customer's financial situation and credit history, and the anomaly detection agent checks whether there are any suspicious circumstances in the transaction. Through this combination, the decision-making body can handle more complex and diverse business needs.

[0056] In building the AI ​​decision-making body, the system also has the ability to dynamically adjust the roles of AI agents. This means that when a business scenario requires different expertise or processing methods, the decision-making body can reorganize its members and call upon AI agents that are more suitable for the current task, thereby ensuring the accuracy and effectiveness of the decision.

[0057] By dynamically invoking and combining AI agents, the system can quickly respond to different business types and scenarios, providing customized, high-quality decisions. This mechanism not only reduces human factors in the decision-making process and improves the objectivity and accuracy of decisions, but also effectively avoids information silos and promotes the sharing and application of cross-business knowledge.

[0058] Furthermore, the steps of invoking N AI agents based on the business type and combining the N AI agents to obtain the AI ​​decision-making body include: in the case of loan business, invoking the identity verification agent, credit assessment agent, and loan anomaly detection agent; and combining the identity verification agent, credit assessment agent, and loan anomaly detection agent to obtain the AI ​​decision-making body for the loan business.

[0059] Specifically, upon receiving a loan request, the system first invokes an identity verification agent. This agent is responsible for verifying all identity information submitted by the applicant, including but not limited to scanned copies of their ID card, personal information provided during voice calls, and related multimodal data. Through optical character recognition (OCR) technology and natural language processing, the identity verification agent can quickly and accurately identify and compare information to ensure the applicant's identity is authentic and trustworthy.

[0060] Next, the system will invoke a credit assessment agent to perform a credit score based on the applicant's financial statements, credit history, transaction records, and other data. It can not only process structured financial data but also use natural language processing models to understand unstructured information in multimodal data, such as the applicant's descriptions of income stability and financial status revealed in voice communication, thereby arriving at a more comprehensive and accurate credit assessment result.

[0061] Subsequently, a loan anomaly detection agent is invoked to monitor for potential fraud and non-compliance. By analyzing loan application patterns, applicants' past loan records, and internal and external loan-related data, this agent can promptly identify any unusual behavior, such as excessively high debt ratios or sudden changes in income, thus preventing the occurrence of bad loans.

[0062] Finally, the aforementioned AI agents are combined to form an AI decision-making body for loan processing. This body integrates the analytical results of each agent, uses built-in decision-making algorithms to assess loan risk, decide whether to approve the application, and determine specific conditions such as loan amount and interest rate. This combined approach fully leverages the expertise of each agent, ensuring comprehensive and accurate decision-making.

[0063] Furthermore, the steps of invoking N AI agents based on the business type and combining the N AI agents to obtain the AI ​​decision-making body include: in the case of user asset management business, invoking the perception agent, compliance review agent, and asset allocation agent; and combining the perception agent, compliance review agent, and asset allocation agent to obtain the AI ​​decision-making body for user asset management business.

[0064] Specifically, when handling asset management business, the system first invokes the Perception Agent to monitor macroeconomic indicators, market dynamics, and any volatile events that may affect investment management. It can utilize attention-based predictive models to analyze news, economic data, and social media sentiment in real time, assessing market risks and predicting potential investment opportunities.

[0065] Subsequently, a compliance audit agent is invoked. This agent is responsible for overseeing the recommended asset management plans to ensure they strictly comply with current financial regulations and internal policies. By combining a rules engine with a dynamically updated regulatory knowledge graph, it can accurately determine the compliance of the plans, avoiding potential risks and losses due to regulatory violations.

[0066] Based on market awareness and compliance review, an asset allocation agent is invoked to generate personalized asset allocation strategies. The agent uses reinforcement learning algorithms to optimize the portfolio based on the client's risk appetite, investment objectives, and market forecasts, ensuring that the investment strategy meets the client's individual needs while adapting to the ever-changing market environment and delivering competitive investment returns.

[0067] Finally, the market perception agent, compliance review agent, and asset allocation agent are combined to form an AI decision-making body for user asset management. This decision-making body can work collaboratively, comprehensively considering market conditions, regulations, and the client's individual needs to provide customized asset management advice. By dynamically adjusting the combination of AI agents, the system can flexibly respond to various asset management scenarios, ensuring the comprehensiveness and professionalism of its decisions.

[0068] Step S203: Input business information into the artificial intelligence decision-making body, and make business decisions through the artificial intelligence decision-making body to obtain business decision results.

[0069] In step S203 above, after the AI ​​agent is invoked and combined to form an AI decision-making body for a specific business type, the business information parsed from the financial business request is input into the AI ​​decision-making body, which then makes automated intelligent decisions. Before transmitting the business information to the decision-making body, the system integrates and standardizes the information to ensure that all data conforms to the input requirements of the decision-making body in terms of form and format. This information may include basic customer information, financial status, market trends, laws and regulations, etc., covering comprehensive data required for business processing.

[0070] Pre-processed business information is mapped to a pre-built decision-making body, which consists of AI agents with different business roles, optimized for specific business types. For example, in loan approval, the decision-making body would include an identity verification agent, a credit assessment agent, and a loan anomaly detection agent. Each agent within the decision-making body analyzes and evaluates the information based on its own knowledge base and decision-making algorithm, outputting preliminary decision suggestions or scores. The master agent (or integration agent) in the decision-making body is responsible for aggregating the analysis results of all agents and performing in-depth comprehensive analysis. It may use voting mechanisms, weighted averages, or more complex decision tree algorithms to determine the final business decision. For example, in a loan approval scenario, the master agent would integrate the results of identity verification, credit assessment, and anomaly detection to ultimately decide whether to approve the loan and the loan conditions. After completing the comprehensive analysis, the decision-making body outputs the business decision result.

[0071] Furthermore, the steps for inputting business information into the AI ​​decision-making body and obtaining business decision results include: in the case of loan business, inputting business information into the AI ​​decision-making body for loan business; extracting text and voice data of business information through an identity verification agent, and verifying the consistency of the content of the text and voice data to obtain verification results; conducting credit assessment on the target user through a credit assessment agent to obtain credit assessment results; conducting anomaly detection on the target user's loan business through a loan anomaly detection agent to obtain loan anomaly detection results; and obtaining loan approval results based on the verification results, credit assessment results, and loan anomaly detection results.

[0072] Specifically, for loan business decisions, firstly, all business information related to the loan application, including text data submitted by the customer (such as application forms, financial statements, etc.) and voice data (such as recorded phone calls explaining the purpose of the loan), is collected and formatted into standardized input that the decision-making body can understand. The identity verification agent uses advanced information extraction technologies, such as OCR (Optical Character Recognition) and ASR (Automatic Speech Recognition), to extract key information from the text and voice data. The agent compares the extracted text descriptions with the voice descriptions for semantic consistency, checking whether information such as the customer's name and loan amount are consistent across different modalities, and whether there are any contradictions or anomalous statements. A comprehensive credit assessment is conducted on the target user, including but not limited to analyzing the user's financial situation, credit history, transaction behavior, and market credit score. Patterns and historical data of loan applications are analyzed to identify any possible fraudulent behavior or non-compliance, such as unusual loan amount requests or inadequate repayment plans. Based on the consistency verification results, credit assessment results, and loan anomaly detection results, the decision-making body uses built-in decision-making algorithms (such as rule-based systems or machine learning models) to conduct a comprehensive analysis and ultimately decide whether to approve the loan and the specific conditions of the loan, such as interest rate and repayment period.

[0073] Furthermore, the steps for inputting business information into the AI ​​decision-making body and obtaining business decision results through the AI ​​decision-making body include: in the case of user asset management business, obtaining user profiles of target users; inputting business information and user profiles into the AI ​​decision-making body for user asset management business; monitoring fluctuation events through a perception agent and identifying non-compliant events through a compliance audit agent; and generating asset allocation plans by combining user profiles, business information, fluctuation events, and non-compliant events through an asset allocation agent.

[0074] For user asset management decisions, the system first collects relevant information about target users from multiple channels, including historical transaction records, investment preferences, and risk tolerance. This information collectively constitutes a user profile. Leveraging deep learning technology and natural language processing algorithms, the system can extract key user characteristics from unstructured data (such as emails or voice recordings of customer communications). The collected business information (e.g., market data, economic indicators) and user profile information are then uniformly formatted and input into a pre-built AI-powered decision-making body for user asset management.

[0075] The perception agent utilizes advanced natural language processing technology and event recognition algorithms to extract and monitor volatile events that may affect asset value from data sources such as news, social media, and economic reports, including interest rate adjustments and significant stock market fluctuations. The compliance audit agent, based on the latest financial regulations and internal policies, uses rule engines and knowledge graph technology to identify any potential non-compliance events, ensuring all decisions and operations comply with relevant rules. The asset allocation agent combines user profiles, business information, volatility events, and non-compliance events, using reinforcement learning algorithms and historical data backtesting to dynamically adjust asset portfolio strategies and generate asset allocation plans that align with user risk preferences and market expectations.

[0076] Furthermore, before obtaining the user profile of the target user, the process includes: collecting multi-source data of the target user, wherein the multi-source data includes at least one of the following: financial interaction data and user behavior data; inputting the multi-source data into a feature extraction model, and extracting the target user's sentiment tendency features and preference features through the feature extraction model, wherein the feature extraction model is a pre-trained machine learning model for feature extraction; and updating the target user's user profile based on the sentiment tendency features and preference features.

[0077] Specifically, to more accurately update the user profiles of target users, the system first collects multi-source data from multiple channels, including but not limited to financial interaction data (such as transaction records and credit history) and user behavior data (such as application usage habits and web browsing history). The collection of multi-source data covers various aspects of user financial activities, providing a rich source of information for subsequent analysis. The collected multi-source data is then input into a pre-trained feature extraction model. This model, designed based on machine learning algorithms, is used to extract user sentiment and preference features from the raw data. The feature extraction model can utilize natural language processing technology to analyze users' text and voice data to identify their emotional tendencies (e.g., whether their trading attitude is positive or negative); simultaneously, it infers their preference features (e.g., which type of financial products or services they prefer) through user behavior data. Based on the extracted sentiment and preference features, the system automatically updates the user profiles of target users in the database. The user profile is a virtual description that integrates user personal information, behavioral habits, preferences, and emotional state, enabling the intelligent agent to more accurately identify and respond to user needs.

[0078] Step S204: Obtain the confidence value of the intelligent decision-making body. If the confidence value is less than the preset confidence threshold, push the business decision result to the business terminal for correction and execute the corrected business decision result.

[0079] In step S204 above, after the intelligent decision-making body completes the business decision, the system automatically calculates the confidence level of this decision. The confidence level reflects the degree of certainty the decision-making body has about its output result, and is a quantitative indicator derived by analyzing various factors such as the output of each agent in the decision-making process, the stability of model predictions, and the coherence of decision logic. The calculated confidence level is then compared with a preset confidence threshold. The preset threshold is an empirically set numerical standard used to define the minimum confidence level required for the decision-making body to make an independent decision without additional human review. If the confidence level of the intelligent decision-making body is lower than the preset threshold, it indicates that the decision-making body cannot be fully confident in the correctness or completeness of its decision. In this case, the system will push the business decision result to the business terminal for relevant business personnel to review and correct. Through the review at the business terminal, expert knowledge and judgment can be combined to correct any omissions or errors in the automatic decision, improving the accuracy and applicability of the decision. After receiving the decision result, business personnel can review it based on their professional knowledge and experience and make corrections if necessary. The corrected content will be re-entered into the system, and the intelligent decision-making body will execute the corrected business decision result. Ensure that the final business decision is the optimal solution for human-machine collaboration.

[0080] Through the above steps, financial business requests from users are received and parsed to obtain business information and business type. Based on the business type, N artificial intelligence agents are invoked and combined to obtain an artificial intelligence decision-making body, where N is a positive integer. Then, the business information is input into the artificial intelligence decision-making body, which makes business decisions and obtains the business decision results. Finally, the confidence value of the intelligent decision-making body is obtained. If the confidence value is less than a preset confidence threshold, the business decision results are pushed to the business terminal for correction, and the corrected business decision results are executed.

[0081] In this embodiment, the financial service requests sent by users are analyzed in depth, and multiple AI agents with different functions are dynamically combined based on the analyzed service information to form an AI decision-making body. Intelligent decision-making through multi-service AI agents breaks down the barriers between traditional task models, promotes the sharing and application of cross-service knowledge, solves the problem of data silos, and improves the accuracy of intelligent decision-making. This also addresses the technical problem in related technologies where using a single rule-based AI model to assist in business decision-making results in low accuracy.

[0082] The following describes in detail another optional implementation method.

[0083] This invention provides an adaptive decision-making system for financial business based on a multimodal AI agent, applicable to multiple business scenarios such as credit approval, customer service, risk management, and personalized financial product recommendation.

[0084] First, multimodal perception fusion is introduced to construct a heterogeneous data processing pipeline, simultaneously parsing text (application forms), voice (telephone recordings), images (scanned documents), and time-series data (transaction logs). A cross-modal alignment algorithm is used to verify the consistency of user data from multiple sources.

[0085] Simultaneously, a reconfigurable agent cluster is designed, with each agent containing: business role labels (credit reviewer, compliance officer, financial advisor, etc.), a domain knowledge graph (financial regulatory rules, product library, risk case library), and an adaptive decision-making module (a policy generator based on reinforcement learning). Agent combinations can be automatically activated based on business type. For example, in a loan application scenario, an identity verification agent, a credit assessment agent, and an anomaly detection agent can be combined to construct an AI decision-making body for loan business.

[0086] Finally, human-machine collaborative incremental learning is configured. When the confidence level of the AI ​​decision is lower than the threshold, a progressive verification mechanism is triggered, an interpretability report is pushed to experts, and the decision results are corrected and fed back to the online learning module to update the Agent's decision strategy library.

[0087] For example, in credit approval, the customer submits a loan application (including an income certificate PDF and a voice explanation of the loan purpose); the identity verification agent uses OCR to extract the text and performs semantic consistency checks on the voice content (such as whether the income amount is consistent); the credit assessment agent performs a credit assessment, and the decision agent outputs the credit decision result.

[0088] For user asset management business, multi-source data collection and preprocessing are carried out on users. Multi-source data includes, but is not limited to: unstructured data: user emails / chat logs (text), user manager communication recordings (voice), video interview records (visual); behavioral data: financial APP operation logs (click stream, page dwell time), financial product browsing / transaction history (time series data); market data: gold price fluctuations, foreign exchange rates.

[0089] Subsequently, sentiment characteristics (anxiety / optimism index) and investment preference characteristics (such as "risk aversion" and "stability") in the email text were extracted using financial domain features, thereby constructing a three-dimensional knowledge graph of customer-asset-risk. A real-time trigger mechanism was then used to activate the agent's decision-making process.

[0090] The perception agent identifies real-time market volatility events, the compliance review agent ensures that the decision-making results comply with asset management rules, and the strategy generation agent outputs asset allocation portfolios.

[0091] This invention provides an adaptive intelligent decision-making approach based on multimodal AI agents, applicable to various scenarios such as credit approval, customer service, risk management, and personalized financial product recommendations in financial institutions. This addresses the shortcomings of traditional automated systems, such as data silos preventing cross-business knowledge application and the inability to process multimodal data, thereby improving the accuracy of intelligent decision-making.

[0092] The following is a detailed description with reference to another embodiment.

[0093] Example 2

[0094] The financial business decision-making device provided in this embodiment includes multiple implementation units, each of which corresponds to a specific implementation step in the above embodiment one. The specific implementation method and beneficial effects can be referred to the foregoing method embodiment, and will not be repeated here.

[0095] Figure 3 This is a schematic diagram of an optional financial business decision-making device according to an embodiment of the present invention, such as... Figure 3 As shown, the financial business decision-making device may include: a receiving unit 31, a calling unit 32, a decision-making unit 33, and a correction unit 34, wherein,

[0096] The receiving unit 31 is used to receive financial business requests from the user terminal, and to parse the financial business requests to obtain business information and business type.

[0097] Calling unit 32 is used to call N artificial intelligence agents based on the business type, and combine the N artificial intelligence agents to obtain an artificial intelligence decision-making body, where N is a positive integer;

[0098] Decision unit 33 is used to input business information into the artificial intelligence decision-making body, and to make business decisions through the artificial intelligence decision-making body to obtain business decision results;

[0099] The correction unit 34 is used to obtain the confidence value of the intelligent decision-making body, and when the confidence value is less than the preset confidence threshold, push the business decision result to the business terminal for correction and execute the corrected business decision result.

[0100] The aforementioned financial business decision-making device receives financial business requests from the user terminal through receiving unit 31, parses the requests to obtain business information and business type; calls N artificial intelligence agents based on the business type through calling unit 32, and combines the N agents to obtain an artificial intelligence decision-making body, where N is a positive integer; inputs the business information into the artificial intelligence decision-making body through decision-making unit 33, performs business decisions through the artificial intelligence decision-making body, and obtains the business decision result; obtains the confidence value of the intelligent decision-making body through correction unit 34, and if the confidence value is less than a preset confidence threshold, pushes the business decision result to the business terminal for correction, and executes the corrected business decision result.

[0101] In this embodiment, the financial service requests sent by users are analyzed in depth, and multiple AI agents with different functions are dynamically combined based on the analyzed service information to form an AI decision-making body. Intelligent decision-making through multi-service AI agents breaks down the barriers between traditional task models, promotes the sharing and application of cross-service knowledge, solves the problem of data silos, and improves the accuracy of intelligent decision-making. This also addresses the technical problem in related technologies where using a single rule-based AI model to assist in business decision-making results in low accuracy.

[0102] Furthermore, the financial business decision-making device also includes: a first construction module for constructing AI agents for different business scenarios, wherein the AI ​​agents are configured with domain knowledge graphs and adaptive decision-making mechanisms; a first configuration module for configuring business role labels for the AI ​​agents; and a first establishment module for establishing a mapping relationship between the AI ​​agents and business types based on the business role labels.

[0103] Furthermore, the calling unit includes: a first calling module, used to call the identity verification agent, credit assessment agent, and loan anomaly detection agent when the business type is loan business; and a first combining module, used to combine the identity verification agent, credit assessment agent, and loan anomaly detection agent to obtain the artificial intelligence decision-making body for loan business.

[0104] Furthermore, the calling unit also includes: a second calling module, used to call the perception agent, compliance review agent, and asset allocation agent when the business type is user asset management business; and a second combination module, used to combine the perception agent, compliance review agent, and asset allocation agent to obtain the artificial intelligence decision-making body for user asset management business.

[0105] Furthermore, the decision-making unit includes: a first input module, used to input business information into the AI ​​decision-making body for loan business when the business type is loan business; a first verification module, used to extract text and voice data of business information through an identity verification agent, and to verify the consistency of the content of the text and voice data to obtain verification results; a first evaluation module, used to conduct credit evaluation of the target user through a credit evaluation agent to obtain credit evaluation results; a first detection module, used to conduct anomaly detection of the target user's loan business through a loan anomaly detection agent to obtain loan anomaly detection results; and a first acquisition module, used to obtain loan approval results based on the verification results, credit evaluation results, and loan anomaly detection results.

[0106] Furthermore, the decision-making unit also includes: a second acquisition module, used to acquire user profiles of target users when the business type is user asset management business; a second input module, used to input business information and user profiles into the artificial intelligence decision-making body of user asset management business; a first identification module, used to monitor fluctuation events through a perception agent and identify non-compliant events through a compliance review agent; and a first decision module, used to make asset allocation decisions and generate asset allocation plans by combining user profiles, business information, fluctuation events, and non-compliant events through an asset allocation agent.

[0107] Furthermore, the financial business decision-making device also includes: a first acquisition module for acquiring multi-source data of the target user, wherein the multi-source data includes at least one of the following: financial interaction data and user behavior data; a first extraction module for inputting the multi-source data into a feature extraction model to extract the target user's sentiment tendency features and preference features through the feature extraction model, wherein the feature extraction model is a pre-trained machine learning model for feature extraction; and a first update module for updating the target user's user profile based on the sentiment tendency features and preference features.

[0108] It should be noted that the receiving unit 31, calling unit 32, decision-making unit 33, and correction unit 34 mentioned above correspond to steps S201 to S204 in Embodiment 1. The instances and application scenarios implemented by the above units and corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1. It should be noted that the above modules or units can be hardware or software components stored in memory (e.g., memory 104) and processed by one or more processors (e.g., processors 102a, 102b, ..., 102n). The above modules or units can also run as part of the device in the computer terminal 10 provided in Embodiment 1.

[0109] The invention will now be described in conjunction with another alternative embodiment.

[0110] Example 3

[0111] The present invention can also provide an electronic device. Figure 4 This is a hardware structure block diagram of an electronic device (or mobile device) for executing a financial business decision-making method according to an embodiment of the present invention, such as... Figure 4 As shown, the electronic device may include: one or more ( Figure 4 (Only one is shown) processor 502, memory 504, memory controller, and peripheral interface, wherein the peripheral interface is connected to the radio frequency module, audio module and display.

[0112] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the methods and apparatus in the embodiments of this application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby implementing the above-described methods. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0113] The processor can access information and applications stored in the memory via a transmission device to perform the following steps: receiving financial business requests from the user terminal and parsing the requests to obtain business information and business type; invoking N artificial intelligence agents based on the business type and combining the N agents to obtain an artificial intelligence decision-making body, where N is a positive integer; inputting the business information into the artificial intelligence decision-making body, making business decisions through the body, and obtaining the decision result; obtaining the confidence value of the decision-making body, and if the confidence value is less than a preset threshold, pushing the decision result to the business terminal for correction and executing the corrected decision result.

[0114] The processor can access information and applications stored in the memory via a transmission device to perform the following steps: constructing AI agents for different business scenarios, wherein the AI ​​agents are configured with domain knowledge graphs and adaptive decision-making mechanisms; configuring business role labels for the AI ​​agents; and establishing a mapping relationship between the AI ​​agents and business types based on the business role labels.

[0115] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: in the case of loan business, call the identity verification agent, credit assessment agent and loan anomaly detection agent; combine the identity verification agent, credit assessment agent and loan anomaly detection agent to obtain the artificial intelligence decision-making body for loan business.

[0116] The processor can invoke information and applications stored in the memory through the transmission device to perform the following steps: in the case of user asset management business, invoke the perception agent, compliance review agent and asset allocation agent; combine the perception agent, compliance review agent and asset allocation agent to obtain the artificial intelligence decision-making body for user asset management business.

[0117] The processor can access information and applications stored in the memory via a transmission device to execute the following steps: In the case of a loan business, input business information into the AI ​​decision-making body for the loan business; extract text and voice data of the business information through an identity verification agent, and verify the consistency of the content of the text and voice data to obtain verification results; conduct a credit assessment on the target user through a credit assessment agent to obtain credit assessment results; conduct anomaly detection on the target user's loan business through a loan anomaly detection agent to obtain loan anomaly detection results; and obtain loan approval results based on the verification results, credit assessment results, and loan anomaly detection results.

[0118] The processor can access information and applications stored in the memory via a transmission device to perform the following steps: In the case of a user asset management business, obtain a user profile of the target user; input the business information and user profile into the artificial intelligence decision-making body of the user asset management business; monitor fluctuation events through a perception agent and identify non-compliant events through a compliance audit agent; and make asset allocation decisions and generate an asset allocation plan by combining the user profile, business information, fluctuation events, and non-compliant events through an asset allocation agent.

[0119] The processor can invoke information and applications stored in the memory via a transmission device to perform the following steps: collecting multi-source data of the target user, wherein the multi-source data includes at least one of the following: financial interaction data and user behavior data; inputting the multi-source data into a feature extraction model, extracting the target user's sentiment tendency features and preference features through the feature extraction model, wherein the feature extraction model is a pre-trained machine learning model for feature extraction; and updating the target user's user profile based on the sentiment tendency features and preference features.

[0120] This invention provides a financial business decision-making scheme. It deeply analyzes user-sent financial business requests and dynamically combines multiple AI agents with different functions based on the analyzed business information to form an AI decision-making body. This multi-business AI agent-based intelligent decision-making breaks down the barriers between traditional task models, promotes cross-business knowledge sharing and application, solves the problem of data silos, and improves the accuracy of intelligent decision-making. Furthermore, it addresses the technical problem in related technologies where using a single rule-based AI model for business decision-making results in low accuracy.

[0121] Those skilled in the art will understand that Figure 4 The structure shown is for illustrative purposes only. Electronic devices can also be smartphones, tablets, handheld computers, mobile internet devices (MIDs), PADs, and other terminal devices. Figure 4 This does not limit the structure of the aforementioned electronic device. For example, electronic devices may also include components that are more... Figure 4 The more or fewer components shown (such as network interfaces, display devices, etc.), or having the same Figure 4 The different configurations shown.

[0122] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0123] The invention will now be described in conjunction with another alternative embodiment.

[0124] Example 4

[0125] This invention also provides a computer-readable storage medium. Optionally, in this invention, the computer-readable storage medium can be used to store the program code executed by the financial business decision-making method provided in Embodiment 1.

[0126] Optionally, in this embodiment of the invention, the storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0127] This invention also provides a computer program product, which, when executed on a data processing device, is suitable for performing the steps of a financial business decision-making method: receiving a financial business request from a user terminal, parsing the financial business request to obtain business information and business type; invoking N artificial intelligence agents based on the business type, and combining the N artificial intelligence agents to obtain an artificial intelligence decision-making body, where N is a positive integer; inputting the business information into the artificial intelligence decision-making body, making a business decision through the artificial intelligence decision-making body, and obtaining a business decision result; obtaining the confidence value of the intelligent decision-making body, and if the confidence value is less than a preset confidence threshold, pushing the business decision result to the business terminal for correction, and executing the corrected business decision result.

[0128] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0129] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0130] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0131] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0132] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0133] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0134] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method of financial services decision making, characterized by, The method comprises the following steps: receiving a financial service request of a user terminal and parsing the financial service request to obtain service information and a service type; calling N artificial intelligence agents based on the service type and combining the N artificial intelligence agents to obtain an artificial intelligence decision body, wherein N is a positive integer; inputting the service information into the artificial intelligence decision body to make a service decision through the artificial intelligence decision body and obtaining a service decision result; obtaining a confidence value of the intelligent decision body, and in the case that the confidence value is less than a preset confidence threshold, pushing the service decision result to a service terminal for correction and executing the corrected service decision result.

2. The method of claim 1, wherein, Before receiving the financial service request of the user terminal, the method further comprises the following steps: constructing artificial intelligence agents under different business scenarios, wherein the artificial intelligence agents are configured with a domain knowledge graph and a self-adaptive decision mechanism; configuring a business role label for the artificial intelligence agents; establishing a mapping relationship between the artificial intelligence agents and the service type based on the business role label.

3. The method of claim 2, wherein, The step of calling N artificial intelligence agents based on the service type and combining the N artificial intelligence agents to obtain an artificial intelligence decision body comprises the following steps: in the case that the service type is a loan service, calling an identity verification agent, a credit evaluation agent and a loan anomaly detection agent; combining the identity verification agent, the credit evaluation agent and the loan anomaly detection agent to obtain an artificial intelligence decision body for the loan service.

4. The method of claim 2, wherein, The step of calling N artificial intelligence agents based on the service type and combining the N artificial intelligence agents to obtain an artificial intelligence decision body comprises the following steps: in the case that the service type is a user asset management service, calling a perception agent, a compliance audit agent and an asset allocation agent; combining the perception agent, the compliance audit agent and the asset allocation agent to obtain an artificial intelligence decision body for the user asset management service.

5. The method of claim 3, wherein, The step of inputting the service information into the artificial intelligence decision body to make a service decision through the artificial intelligence decision body and obtaining a service decision result comprises the following steps: in the case that the service type is a loan service, inputting the service information into the artificial intelligence decision body for the loan service; extracting text data and voice data of the service information through the identity verification agent and performing consistency verification on the content of the text data and the voice data to obtain a verification result; performing credit evaluation on a target user through the credit evaluation agent to obtain a credit evaluation result; performing anomaly detection on the loan service of the target user through the loan anomaly detection agent to obtain a loan anomaly detection result; obtaining a loan approval result based on the verification result, the credit evaluation result and the loan anomaly detection result.

6. The method of claim 4, wherein, The step of inputting the service information into the artificial intelligence decision body to make a service decision through the artificial intelligence decision body and obtaining a service decision result comprises the following steps: in the case that the service type is a user asset management service, obtaining a user portrait of a target user; inputting the business information and the user portrait into an artificial intelligence decision body of the user asset management business; monitoring fluctuation events through the perception agent and identifying non-compliance events through the compliance audit agent; making asset allocation decisions through the asset allocation agent in combination with the user portrait, the business information, the fluctuation events, and the non-compliance events to generate an asset allocation scheme.

7. The method of claim 6, wherein, Before obtaining the user portrait of the target user, further comprising: collecting multi-source data of the target user, wherein the multi-source data comprises at least one of the following: financial interaction data, user behavior data; inputting the multi-source data into a feature extraction model to extract sentiment tendency features and preference features of the target user through the feature extraction model, wherein the feature extraction model is a pre-trained machine learning model for feature extraction; updating the user portrait of the target user based on the sentiment tendency features and the preference features.

8. A financial service decision apparatus characterized by comprising: comprising: a receiving unit configured to receive a financial service request of a user terminal, and parse the financial service request to obtain business information and a business type; a calling unit configured to call N artificial intelligence agents based on the business type, and combine the N artificial intelligence agents to obtain an artificial intelligence decision body, wherein N is a positive integer; a decision unit configured to input the business information into the artificial intelligence decision body, and make a business decision through the artificial intelligence decision body to obtain a business decision result; a correction unit configured to obtain a confidence value of the intelligent decision body, and in a case where the confidence value is less than a preset confidence threshold, push the business decision result to a business terminal for correction, and execute the corrected business decision result.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium comprises a stored computer program, wherein the computer program controls the device in which the computer-readable storage medium is located to execute the financial service decision method of any one of claims 1 to 7 when the computer program is running.

10. An electronic device, comprising: comprise one or more processors and memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the financial service decision method of any one of claims 1 to 7.

11. A computer program product, characterised in that, The computer program product comprises a computer program, wherein the computer program is executed by a processor to implement the financial service decision method of any one of claims 1 to 7.

Images