An agent-based service management method, device, medium and equipment

Abstract

The embodiment of the present specification discloses a service management method based on an agent, so that the second service party providing service business to the first service party no longer needs to manually manage its own business, which can significantly improve the efficiency of service management. Moreover, when the first service party needs to face many second service parties, the first agent can also obtain the state of the service provided by each second service party in a unified form, which is convenient for the first service party to manage each second service party.

Description

Technical Field

[0001] This specification relates to the field of computer technology, and in particular to a business management method, apparatus, medium and device based on intelligent agents. Background Technology

[0002] With the development of computer and internet technology, service providers offering various services to users are emerging in large numbers.

[0003] For some service providers, as their business expands, they also need services from other service providers to better serve users. For ease of distinction, the service provider offering services to users will be referred to as the first service provider, and the service provider offering services to the first service provider will be referred to as the second service provider.

[0004] For example, when the first service provider offers payment services to users, it often needs to install its payment acceptance equipment (such as POS machines, non-payment acceptance devices, etc.) in many merchant stores. In this case, the first service provider needs a second service provider to provide the installation service for the payment acceptance equipment.

[0005] Generally speaking, the second service provider needs to not only complete the deployment of payment acceptance equipment, but also maintain and operate these payment acceptance equipment, and regularly report relevant information on the deployment and operation and maintenance of the payment acceptance equipment to the first service provider. All of these fall under the management of the second service provider's own business.

[0006] For second-party service providers, managing the status of their deployed payment acceptance devices and obtaining timely status information is typically done manually, which is not only extremely difficult but also inefficient. Therefore, improving the efficiency of second-party service providers in managing their own operations is a pressing issue that needs to be addressed. Summary of the Invention

[0007] This specification provides an agent-based business management method, apparatus, storage medium, and electronic device to partially solve the problems existing in the prior art.

[0008] The embodiments in this specification adopt the following technical solutions: This specification provides a business management method based on intelligent agents, the method comprising: The server of the first service provider receives input information in natural language form sent by the second service provider; The input information is input into the first intelligent agent. Through the large language model (LLM) corresponding to the first intelligent agent, at least one target intent of the second service provider is determined from each preset business management intent based on the input information. For each target intent, the first intelligent agent generates the corresponding business management task. The business management task is input into the second intelligent agent corresponding to the target intent, and the business management task is executed by the second intelligent agent. Obtain the business management results obtained by each second intelligent agent after executing its respective received business management task, and display the business management results to the second service provider.

[0009] This specification provides a business management device based on an intelligent agent, the device comprising: The receiving module is used to receive input information in natural language form sent by the second service provider; The intent recognition module is used to input the input information into the first intelligent agent, and through the large language model (LLM) corresponding to the first intelligent agent, determine at least one target intent of the second service provider from among the preset business management intents based on the input information. The task generation module is used to generate a business management task corresponding to each target intent through the first intelligent agent. The task execution module is used to input the business management task into the second intelligent agent corresponding to the target intent, and execute the business management task through the second intelligent agent; The integrated presentation module obtains the business management results obtained by each second intelligent agent after executing its respective received business management tasks, and displays the business management results to the second service provider.

[0010] This specification provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned agent-based business management method.

[0011] This specification provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the above-described agent-based business management method.

[0012] The above-described at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects: This specification discloses an agent-based business management method. A first service provider offers a second service provider a first agent and a second agent to manage business operations. The first agent receives input information in natural language from the second service provider and identifies and determines each target intent of the second service provider from a set of preset business management intents. It then generates a corresponding business management task for each target intent and inputs the task into the responding second agent. The second agent executes its respective business management task, and the first agent obtains the business management results obtained by each second agent and displays them to the second service provider. This method eliminates the need for manual management of the business operations of the second service provider, significantly improving efficiency. Furthermore, when the first service provider faces multiple second service providers, the first agent provides a unified understanding of the service status of each second service provider, facilitating management by the first service provider. Attached Figure Description

[0013] The accompanying drawings, which are included to provide a further understanding of this specification and form part of this specification, illustrate exemplary embodiments and are used to explain this specification, but do not constitute an undue limitation thereof. In the drawings: Figure 1 A flowchart of a business management method based on intelligent agents provided in the embodiments of this specification; Figure 2 This is a schematic diagram of the logical structure of the business management system provided in the embodiments of this specification; Figure 3 A schematic diagram of an agent-based service management device provided in the embodiments of this specification; Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this specification. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this specification will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of them. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0015] The technical solutions provided in the various embodiments of this specification are described in detail below with reference to the accompanying drawings.

[0016] Figure 1A flowchart of a business management method based on intelligent agents, provided for embodiments of this specification, specifically includes the following steps: S100: The server of the first service provider receives input information in natural language form sent by the second service provider.

[0017] In the embodiments described in this specification, the first service provider may be a service provider that directly provides services to users, while the second service provider is a service provider that provides services to the first service provider. For example, the first service provider may be a service provider that directly provides payment services to users, while the second service provider may be a service provider that provides payment acceptance equipment installation services to the first service provider.

[0018] For ease of explanation, the following will only use the roles and functions of the first and second service providers in the above example for illustration.

[0019] Although the first service provider is the client of the second service provider, in order to improve the efficiency of the second service provider in managing the business it provides to the first service provider, and because the second service provider reports its own business management information to the first service provider in different forms when the first service provider is dealing with many second service providers, this is not conducive to the first service provider's management of all its payment acceptance devices, nor is it conducive to its management of the various second service providers.

[0020] Therefore, in the embodiments of this specification, the first service provider provides the second service provider with an intelligent agent for the second service provider to perform business management, and each second service provider manages its own business uniformly through this intelligent agent.

[0021] In this specification, the intelligent agents provided by the first service provider include two types: first intelligent agents and second intelligent agents. The first intelligent agent is used to identify the business management intent of the second service provider, generate at least one business management task based on the business management intent, and then send each business management task to the appropriate second intelligent agent, which then executes its respective received business management task.

[0022] Specifically, the second service provider can first log in to the server using the login account used by the second service provider to log in to the server through the application (APP) corresponding to the server of the first service provider or through a browser. The second service provider can then send a login request to the server of the first service provider, carrying the aforementioned login account and login verification information (such as login password, verification code, or facial recognition).

[0023] In response to the login request, the server of the first service provider can verify the login account and login verification information. If the verification is successful, the second service provider is allowed to log in to the server using the login account; otherwise, the second service provider is denied access to the server using the login account.

[0024] After the second service provider successfully logs into the server using the login account, it can send input information in natural language form to the server based on the login account, such as "How was the health of the device yesterday?"

[0025] S101: Input the input information into the first intelligent agent, and through the large language model (LLM) corresponding to the first intelligent agent, determine at least one target intent of the second service provider from among the preset business management intents based on the input information.

[0026] After receiving input information in natural language form from the second service provider, the server can input the input information into the first agent, so that the LLM corresponding to the first agent can identify the semantics of the input information, and determine at least one target intent of the second service provider from among the preset business management intents based on the semantics.

[0027] S102: For each target intent, the first intelligent agent generates the corresponding business management task.

[0028] After identifying the target intents of the second service provider, the LLM corresponding to the first intelligent agent can, for each target intent, first determine the task type of the business management task corresponding to the target intent, including but not limited to: querying the status of all deployed payment acceptance devices, querying the status of payment acceptance devices deployed in a specific store, querying the health of all or a specific payment acceptance device, querying the maintenance work orders of payment acceptance devices that are being maintained, and querying the revenue generated by all or a specific payment acceptance device, etc.

[0029] Once the task type corresponding to the target intent is determined, the LLM corresponding to the first agent can further extract the structured task parameters required to execute the task based on the task type and the aforementioned natural language input information. These parameters include, but are not limited to, time, personnel associated with each payment acceptance device, region, and required indicators for each payment acceptance device. After extracting these task parameters from the input information, the LLM can organize them in JSON format to create structured task parameters. This is because the first agent's role is only to identify the second service provider's business management intent and generate the corresponding business management task accordingly. The downstream second agent is responsible for executing the business management task. Therefore, the business management task needs to include the input parameters used to execute it, i.e., these task parameters. Since the downstream second agent needs to call relevant tools when executing the business management task, these task parameters need to be organized into structured task parameters.

[0030] Finally, the LLM can generate a business management task corresponding to the target intent based on the task type determined above and the obtained structured task parameters. This business management task includes both a task description and structured task parameters.

[0031] S103: Input the business management task into the second intelligent agent corresponding to the target intent, and execute the business management task through the second intelligent agent.

[0032] After the first intelligent agent generates the business management tasks corresponding to each target intent through the above step S102, it can input the business management task into the second intelligent agent corresponding to the target intent for each business management task, so that the business management task can be executed by the corresponding second intelligent agent.

[0033] Specifically, the LLM corresponding to the first intelligent agent can query the agent identity information of each second intelligent agent. This agent identity information can be the AgentCard of the second intelligent agent, which contains both the agent identifier and the capability description information. The LLM corresponding to the first intelligent agent can then actively search for and discover available second intelligent agents through each AgentCard, determine the capability description information of each second intelligent agent, and then, based on the capability description information, select the second intelligent agent whose task type or task description matches the target intent's business management task—that is, the second intelligent agent suitable for executing the business management task. Finally, based on the A2A protocol, the business management task is sent to the second intelligent agent via a JSON-RPC message, enabling the second intelligent agent to execute the business management task.

[0034] When the second intelligent agent executes the received business management task, it can specifically select the target tool required to execute the business management task from among the available tools through the LLM corresponding to the second intelligent agent, based on the task type or task description of the business management task and the capability description information of each available tool pre-configured for the second intelligent agent.

[0035] Specifically, the LLM corresponding to the second agent can first determine the skill protocols corresponding to each capability pre-configured for the second agent. A skill protocol for a capability includes a Markdown document containing the capability description information, the tool identifiers of available tools supporting that capability on the MCP server, and the invocation method for calling those tools through the MCP server. Then, the LLM corresponding to the second agent can determine the capabilities required to execute the business management task based on the capability description information contained in the skill protocols for each capability. It can then determine the available tools corresponding to the tool identifiers contained in the skill protocols for each capability; these available tools are the target tools required to execute the business management task. Finally, the second agent can invoke these target tools according to the invocation methods for calling these target tools through the MCP server contained in the skill protocols for each capability, and pass the structured task parameters contained in the business management task to the target tools, enabling the target tools to query business data based on these structured task parameters.

[0036] The target tools described in this specification include, but are not limited to: tools for querying the status data of payment acceptance devices, tools for querying merchant information of stores where payment acceptance devices are located, tools for querying work orders for maintaining payment acceptance devices, and tools for calculating the revenue generated by each payment acceptance device.

[0037] These target tools allow querying the second service provider's business data on the first service provider's server. This business data can include equipment deployment data that the second service provider has uploaded to the first service provider's server (such as the equipment identifiers of the deployed payment acceptance devices, the regions or stores where they are deployed, and the maintenance personnel information corresponding to each payment acceptance device), as well as the operational status data reported by the second service provider's deployed payment acceptance devices to the first service provider's server during operation.

[0038] After the target tool retrieves the required business data from the business data, the LLM corresponding to the second agent can execute the business management task based on this business data and obtain the execution result. Specifically, the LLM corresponding to the second agent can at least use a statistical method corresponding to the task type of the business management task to statistically analyze the business data retrieved by the target tool and obtain the statistical result as the business management result obtained from executing the business management task.

[0039] For example, statistics can be compiled on the number of payment acceptance devices that have been deployed, the number of payment acceptance devices that have been deployed and linked to merchants' payment accounts, the number of payment acceptance devices that have been deployed and are operational, the number of payment acceptance devices that have been deployed and are currently performing payment acceptance services, the number of payment acceptance devices that have been deployed and are under maintenance, etc. It may also include the indicators of payment acceptance devices that need to be statistically analyzed in the business management task, such as the change in the number of currently deployed and operational payment acceptance devices compared to a specified historical period, or the change in any of the above-mentioned statistical quantities compared to a specified historical period.

[0040] S104: Obtain the business management results obtained by each second intelligent agent after executing its respective received business management task, and display the business management results to the second service provider.

[0041] After obtaining the business management result corresponding to the received business management task through step S103, the second intelligent agent can return the business management result to the first intelligent agent (still based on the A2A protocol). The first intelligent agent can then integrate the business management results returned by each second intelligent agent through its corresponding LLM to obtain the final management result, and finally display the final management result to the second service provider. The integration described in this specification includes, but is not limited to: deduplicating the business management results returned by each second intelligent agent, and adding the business management results returned by each second intelligent agent to a preset template.

[0042] Of course, the first intelligent agent can save the input information entered by the second service provider in step S100, and the final management result obtained and displayed to the second service provider in step S104, as a historical session corresponding to the login account used by the second service provider to log in to the server of the first intelligent agent. This allows the second service provider to continue sending the next round of input information based on the historical session, so as to continue through Figure 1 The method shown manages its own business.

[0043] Furthermore, the second service provider may have more than one login account that can be used to log in to the first service provider's server. The first or second service provider can pre-configure the task execution permissions for each login account in the first service provider's server. Figure 1In step S102, after the server completes a business management task corresponding to a target intent of the second service provider through the first intelligent agent, before inputting the business management task to the second intelligent agent in step S103, it can query the task execution permissions of the login account currently used by the second service provider to log in to the server. Based on the queried task execution permissions and the task type or task description of the generated business management task, it determines whether the login account has the permission to execute the business management task. If so, the business management task can be input to the second intelligent agent in step S103 so that the second intelligent agent can execute the business management task. Otherwise, the input of the business management task to the second intelligent agent is refused.

[0044] Since the second service provider may have more than one objective, the first intelligent agent may also generate more than one business management task. The first intelligent agent can use the above-mentioned permission verification to execute only the business management tasks for which it has the task execution permission for the second service provider, and refuse to execute business management tasks for which it does not have the task execution permission for the second service provider.

[0045] The system logic structure diagram consisting of the first intelligent agent, the second intelligent agent, and the various available tools is as follows: Figure 2 As shown. Figure 2 This is a schematic diagram of the logical structure of the business management system provided in the embodiments of this specification. In the user interaction layer, the second service provider can log in to the server using a login account through the application corresponding to the server of the first service provider or through a webpage in a browser. After logging in, it sends input information in natural language form to the first intelligent agent, which generates corresponding business management tasks according to the target intents corresponding to the input information. Then, for each business management task, the permission verification module determines whether the logged-in account has the task execution permission to execute the business management task. After the permission verification is passed, the business management task is input to the corresponding second intelligent agent in the vertical intelligent agent layer. The second intelligent agent calls the target tool in the data service layer, queries business data in the data source layer through the target tool, and returns the queried business data to the second intelligent agent. The second intelligent agent executes the business management task according to the business data to obtain the business management result, and then returns the business management result to the first intelligent agent. The first intelligent agent summarizes and integrates the business management results returned by each second intelligent agent to obtain the final management result, displays it to the second service provider, and stores the input information of the second service provider and the final management result in the history session of the user interaction layer.

[0046] The above is an example of a business management method based on intelligent agents provided in this specification. Based on the same idea, this specification also provides corresponding devices, storage media and electronic devices.

[0047] Figure 3This is a schematic diagram of a business management device based on an intelligent agent provided in an embodiment of this specification. The device can be applied to the server side of a first service provider, and includes: The receiving module 300 is used to receive input information in natural language form sent by the second service provider; The intent recognition module 301 is used to input the input information into the first intelligent agent, and through the large language model (LLM) corresponding to the first intelligent agent, determine at least one target intent of the second service provider from among the preset business management intents based on the input information. Task generation module 302 is used to generate a business management task corresponding to each target intent through the first intelligent agent; Task execution module 303 is used to input the business management task into the second intelligent agent corresponding to the target intent, and execute the business management task through the second intelligent agent; The integration and presentation module 304 obtains the business management results obtained by each second intelligent agent after executing its respective received business management tasks, and displays the business management results to the second service provider.

[0048] Optionally, the receiving module 300 is specifically configured to: receive a login request sent by the second service provider, the login request carrying the login account and login verification information used by the second service provider to log in to the server; in response to the login request, after verifying the login account and the login verification information, allow the second service provider to log in to the server through the login account; and receive input information in natural language form sent by the second service provider based on the login account.

[0049] Optionally, the task generation module 302 is specifically configured to, through the LLM corresponding to the first intelligent agent, perform the following: determine the task type of the business management task corresponding to the target intent based on the target intent; extract the structured task parameters required for the task type from the input information based on the task type and the input information in natural language form; and generate the business management task corresponding to the target intent based on the task type and the structured task parameters.

[0050] Optionally, the task execution module 303 is specifically configured to: query the agent identity information of each second agent through the LLM corresponding to the first agent; determine the capability description information of each second agent based on the agent identity information of each second agent; select a second agent among the second agents whose task type matches the business management task corresponding to the target intent based on the capability description information of each second agent; and input the business management task into the selected second agent.

[0051] Optionally, the task execution module 303 is further configured to query the task execution permissions of the login account before inputting the business management task into the second intelligent agent corresponding to the target intent; and determine, based on the task execution permissions, that the login account has the permission to execute the business management task.

[0052] Optionally, the task execution module 303 is specifically configured to, through the LLM corresponding to the second agent, perform the following: based on the task type of the input business management task and the capability description information of each available tool pre-configured for the second agent, select the target tool required to execute the business management task from among the available tools; invoke the target tool through the second agent and pass the structured task parameters contained in the business management task to the target tool, so that the target tool queries business data according to the structured task parameters; and execute the business management task through the LLM corresponding to the second agent based on the business data queried by the target tool to obtain the business management result.

[0053] Optionally, the business management task includes: managing the status of the payment acceptance equipment deployed by the second service provider for the first service provider; The business data includes the equipment deployment data that the second service provider has uploaded to the server of the first service provider, and / or the operating status data reported by the payment acceptance equipment deployed by the second service provider to the server of the first service provider during operation. The task execution module 303 is specifically used to perform statistical analysis on the business data queried by the target tool through the LLM corresponding to the second intelligent agent, using at least a statistical method corresponding to the task type of the business management task, and obtain statistical results.

[0054] Optionally, the integration presentation module 304 is specifically used to integrate the business management results obtained by each of the second intelligent agents through the LLM corresponding to the first intelligent agent to obtain the final management result; and to display the final management result to the second service provider.

[0055] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, can be used to perform the above-described agent-based business management method.

[0056] This specification also provides a computer program product that, when executed by a processor, can be used to perform the above-described agent-based business management method.

[0057] based on Figure 2 The agent-based business management method shown in this specification also provides embodiments that... Figure 4 The diagram shows the structure of the electronic device. Figure 4 At the hardware level, the electronic device includes a processor, internal bus, network interface, memory, and non-volatile storage, and may also include other hardware required for business operations. The processor reads the corresponding computer program from the non-volatile storage into memory and then runs it to implement the aforementioned agent-based business management method.

[0058] The above description is merely an embodiment of this specification and is not intended to limit this specification. Various modifications and variations can be made to this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of the claims of this specification.

Claims

1. A business management method based on intelligent agents, the method comprising: The server of the first service provider receives input information in natural language form sent by the second service provider; The input information is input into the first intelligent agent. Through the large language model (LLM) corresponding to the first intelligent agent, at least one target intent of the second service provider is determined from each preset business management intent based on the input information. For each target intent, the first intelligent agent generates the corresponding business management task. The business management task is input into the second intelligent agent corresponding to the target intent, and the business management task is executed by the second intelligent agent. Obtain the business management results obtained by each second intelligent agent after executing its respective received business management task, and display the business management results to the second service provider.

2. The method as described in claim 1, wherein the server of the first service provider receives input information in natural language form sent by the second service provider, specifically including: The server of the first service provider receives a login request sent by the second service provider. The login request carries the login account and login verification information used by the second service provider to log in to the server. In response to the login request, after the login account and the login verification information are verified, the second service provider is allowed to log in to the server using the login account; Receive input information in natural language form sent by the second service provider based on the login account.

3. The method as described in claim 1, wherein generating the business management task corresponding to the target intent through the first intelligent agent specifically includes: The LLM corresponding to the first agent executes the following: Based on the target intent, the task type of the business management task corresponding to the target intent is determined; based on the task type and the input information in natural language form, the structured task parameters required for the task type are extracted from the input information; and based on the task type and the structured task parameters, the business management task corresponding to the target intent is generated.

4. The method as described in claim 3, wherein inputting the business management task into the second intelligent agent corresponding to the target intent specifically includes: The LLM corresponding to the first intelligent agent executes the following: query the intelligent agent identity information of each second intelligent agent, determine the capability description information of each second intelligent agent based on the intelligent agent identity information of each second intelligent agent, and select the second intelligent agent among the second intelligent agents whose task type matches the business management task corresponding to the target intent based on the capability description information of each second intelligent agent. The business management task is input into the selected second agent.

5. The method of claim 2, wherein before inputting the business management task into the second intelligent agent corresponding to the target intent, the method further comprises: Query the task execution permissions of the logged-in account; Based on the task execution permissions, it is determined that the login account has the authority to execute the business management task.

6. The method as described in claim 3, wherein the business management task is executed by the second intelligent agent, specifically including: Execution via the LLM corresponding to the second agent: Based on the task type of the input business management task and the capability description information of each available tool pre-configured for the second agent, select the target tool required to execute the business management task from among the available tools; The second intelligent agent invokes the target tool and passes the structured task parameters contained in the business management task to the target tool, enabling the target tool to query business data based on the structured task parameters; The business management task is executed by the LLM corresponding to the second intelligent agent based on the business data queried by the target tool, and the business management result is obtained.

7. The method of claim 6, wherein the business management task includes: The task of managing the status of payment acceptance equipment deployed by the second service provider for the first service provider; The business data includes the equipment deployment data that the second service provider has uploaded to the server of the first service provider, and / or the operating status data reported by the payment acceptance equipment deployed by the second service provider to the server of the first service provider during operation. The business management task is executed by the LLM corresponding to the second intelligent agent based on the business data queried by the target tool, and the execution result is obtained, specifically including: Using the LLM corresponding to the second intelligent agent, at least a statistical method corresponding to the task type of the business management task is used to statistically analyze the business data queried by the target tool and obtain statistical results.

8. The method as described in claim 1, wherein displaying the business management results to the second service provider specifically includes: By integrating the business management results obtained by each of the second agents through the LLM corresponding to the first agent, the final management result is obtained. The final management results are then displayed to the second service provider.

9. A business management device based on intelligent agents, the device comprising: The receiving module is used to receive input information in natural language form sent by the second service provider; The intent recognition module is used to input the input information into the first intelligent agent, and through the large language model (LLM) corresponding to the first intelligent agent, determine at least one target intent of the second service provider from among the preset business management intents based on the input information. The task generation module is used to generate a business management task corresponding to each target intent through the first intelligent agent. The task execution module is used to input the business management task into the second intelligent agent corresponding to the target intent, and execute the business management task through the second intelligent agent; The integrated presentation module obtains the business management results obtained by each second intelligent agent after executing its respective received business management tasks, and displays the business management results to the second service provider.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in any one of claims 1-8.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method according to any one of claims 1-8.

Images