Payment system, method and device based on large model, medium and equipment

Abstract

The embodiment of the invention discloses a payment system based on a large model, and the method comprises the steps: determining a target merchant corresponding to a purchase intention if the purchase intention of a user is recognized when a third-party agent interacts with the user through a preset first LLM, and transmitting the purchase intention to a merchant agent of the target merchant, the merchant agent determines a target commodity matched with the purchase intention and sends a collection request to the payment server for the target commodity, the payment server creates a payment order and returns an order identifier to the third-party agent, and the third-party agent obtains a payment authorization message of the user and sends the payment authorization message to the payment server. And the payment authorization message is sent to a payment server, and the payment server carries out payment settlement on the payment order based on the payment authorization message. Through the payment system, the user can use the intelligent agent to place an order for shopping and also pay for the shopping through the intelligent agent, so that the operation required by online shopping and payment of the user is greatly simplified.

Description

Technical Field

[0001] This specification relates to the field of computer technology, and in particular to a payment system, method, apparatus, medium and device based on a large model. Background Technology

[0002] With the development of artificial intelligence (AI) technology, large language models (LLM) have been widely applied in various fields.

[0003] Through LLM, users can obtain various information and knowledge as if communicating with real people. In particular, generative LLM can also help users generate images and computer code, greatly simplifying the tedious operations required by users in their daily life and work.

[0004] Therefore, in online shopping and payment scenarios, how to simplify the user's operations during the shopping and payment process through LLM is an urgent problem to be solved. Summary of the Invention

[0005] This specification provides a payment system, method, apparatus, storage medium, and electronic device based on a large model to partially solve the problems existing in the prior art.

[0006] The embodiments in this specification adopt the following technical solutions: This specification provides a payment system based on a large model, comprising: a third-party intelligent agent, a merchant intelligent agent, and a payment server; the third-party intelligent agent communicates with at least one merchant's intelligent agent based on an inter-agent A2A protocol; wherein: The third-party intelligent agent is used to interact with the user through a pre-set first large language model. When the first large language model identifies the user's purchase intention based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intention and sends the purchase intention to the merchant intelligent agent of the target merchant based on the A2A protocol. The merchant agent of the target merchant is used to determine the target product that matches the received purchase intention among the products sold by the target merchant through a pre-set second language model, and send a payment request to the payment server according to the target merchant's payment account and the amount to be paid corresponding to the target product; In response to the payment request, the payment server creates a payment order containing the receiving account and the amount to be paid, and returns the order identifier corresponding to the payment order to the merchant agent. The merchant agent of the target merchant is also used to return the order identifier returned by the payment server to the third-party agent; The third-party intelligent agent is also used to obtain the payment authorization message entered by the user for the order identifier, and send the payment authorization message to the payment server; The payment server, in response to the payment authorization message, queries the user's payment account and uses the payment account to settle the payment order.

[0007] This specification provides a payment method based on a large model, the method comprising: The third-party intelligent agent interacts with the user through a pre-built primary language model; When the first language model identifies the user's purchase intention based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intention; Based on the A2A protocol between intelligent agents, the purchase intention is sent to the merchant intelligent agent of the target merchant. The merchant intelligent agent of the target merchant uses the second language model to determine the target product that matches the purchase intention from the products sold by the target merchant, and sends a payment request to the payment server according to the target merchant's payment account and the amount to be paid corresponding to the target product. Receive the order identifier returned by the payment server. The order identifier is the order identifier corresponding to the payment order returned by the payment server through the merchant smart agent after the payment server responds to the payment request, creates a payment order containing the payment account and the amount to be paid; Obtain the payment authorization message entered by the user in response to the order identifier; The payment authorization message is sent to the payment server, which responds to the payment authorization message by querying the user's payment account and using the payment account to settle the payment order.

[0008] This specification provides a payment method based on a large model, the method comprising: The target merchant's intelligent agent receives the user's purchase intention sent by a third-party intelligent agent based on the inter-agent A2A protocol; the purchase intention is the purchase intention at least corresponding to the target merchant identified by the third-party intelligent agent when interacting with the user through a pre-set first language model; Using a pre-set second language model, target products that match the received purchase intent are identified among the products sold by the target merchant. Based on the target merchant's payment account and the outstanding amount corresponding to the target product, a payment request is sent to the payment server; The system receives an order identifier returned by the payment server and sends the order identifier to the third-party intelligent agent. The third-party intelligent agent then obtains the payment authorization message entered by the user for the order identifier and sends the payment authorization message to the payment server. In response to the payment authorization message, the payment server queries the user's payment account and uses the payment account to settle the payment order corresponding to the order identifier. The order identifier is an order identifier created by the payment server in response to the payment request, containing the receiving account and the amount to be paid for the payment order.

[0009] This specification provides a payment method based on a large model, the method comprising: The payment server receives a payment request from the merchant agent of the target merchant. This payment request is generated by the merchant agent using a pre-set second language model to identify a target product among the products sold by the target merchant that matches the user's purchase intent, and then sending the request based on the target merchant's payment account and the outstanding payment amount corresponding to the target product. The user's purchase intent is determined by a third-party agent interacting with the user using a pre-set first language model. When the first language model identifies the user's purchase intent based on the interaction, it identifies at least one target merchant corresponding to the purchase intent and sends it to the target merchant's merchant agent based on the A2A protocol. Based on the payment request containing the target merchant's payment account and the amount to be paid corresponding to the target product, create a payment order containing the payment account and the amount to be paid; The order identifier corresponding to the payment order is returned to the third-party intelligent agent through the merchant intelligent agent; Receive the payment authorization message input by the user for the order identifier, sent by the third-party intelligent agent; In response to the payment authorization message, query the user's payment account; The payment account is used to settle the payment order.

[0010] This specification provides a payment device based on a large model, the device comprising: The interaction module is used to interact with the user through a pre-built primary language model; The merchant identification module is used to identify at least one target merchant corresponding to the purchase intention when the first language model identifies the user's purchase intention based on the interaction with the user. The sending module is used to send the purchase intention to the merchant agent of the target merchant based on the A2A protocol between intelligent agents. The merchant agent of the target merchant uses the second language model to determine the target product that matches the purchase intention from the products sold by the target merchant, and sends a payment request to the payment server according to the payment account of the target merchant and the amount to be paid corresponding to the target product. The receiving module is used to receive the order identifier returned by the payment server. The order identifier is the order identifier corresponding to the payment order returned by the payment server through the merchant agent after the payment server responds to the payment request, creates a payment order containing the payment account and the amount to be paid; The interaction module is also used to obtain the payment authorization message entered by the user for the order identifier; The sending module is further configured to send the payment authorization message to the payment server, so that the payment server responds to the payment authorization message, queries the user's payment account, and uses the payment account to settle the payment order.

[0011] This specification provides a payment device based on a large model, the device comprising: The receiving module is used to receive the user's purchase intention sent by a third-party intelligent agent based on the inter-agent A2A protocol; the purchase intention is the purchase intention of at least the target merchant identified by the third-party intelligent agent when interacting with the user through a preset first language model. The product selection module is used to determine, through a pre-set second language model, target products that match the received purchase intent from among the products sold by the target merchant. The payment initiation module is used to send a payment request to the payment server based on the target merchant's payment account and the amount to be paid corresponding to the target product; The forwarding module is used to receive the order identifier returned by the payment server and send the order identifier to the third-party intelligent agent, so that the third-party intelligent agent can obtain the payment authorization message entered by the user for the order identifier, and send the payment authorization message to the payment server, so that the payment server can respond to the payment authorization message, query the user's payment account, and use the payment account to perform payment settlement for the payment order corresponding to the order identifier; the order identifier is the order identifier of the payment order, which includes the receiving account and the amount to be paid, created by the payment server in response to the payment request.

[0012] This specification provides a payment device based on a large model, the device comprising: The receiving module is used to receive a payment request sent by the merchant agent of the target merchant. The payment request is sent by the merchant agent of the target merchant after determining the target product that matches the user's purchase intention among the products sold by the target merchant through a preset second language model, and based on the target merchant's payment account and the amount to be paid corresponding to the target product. The user's purchase intention is determined by a third-party agent interacting with the user through a preset first language model. When the first language model recognizes the user's purchase intention based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intention and sends it to the merchant agent of the target merchant based on the A2A protocol. The order creation module is used to create a payment order containing the payment account and the amount to be paid for the target product, based on the payment request containing the target merchant's payment account and the amount to be paid for the target product. The sending module is used to return the order identifier corresponding to the payment order to the third-party intelligent agent through the merchant intelligent agent; The receiving module is further configured to receive a payment authorization message sent by the third-party intelligent agent, which is input by the user for the order identifier; The query module is used to query the user's payment account in response to the payment authorization message; The payment module is used to process payment settlements for the payment orders using the payment account.

[0013] This specification provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned payment method based on a large model.

[0014] This specification provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the aforementioned payment method based on a large model.

[0015] This specification provides a computer program product, which includes a computer program that, when executed by a processor, implements the aforementioned payment method based on a large model.

[0016] The above-described at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects: This specification discloses a payment system based on a large model. When a third-party intelligent agent interacts with a user through a pre-set first LLM (Limited Least Model), if it identifies the user's purchase intention, it determines the target merchant corresponding to that intention and sends the purchase intention to the merchant's intelligent agent. The merchant intelligent agent then determines the target product matching the purchase intention and sends a payment request to the payment server for that product. The payment server creates a payment order and returns an order identifier to the third-party intelligent agent. After obtaining the user's payment authorization message, the third-party intelligent agent sends the payment authorization message to the payment server, which then processes the payment order based on the authorization message. Through this payment system, users can place orders and make payments using the same intelligent agent, greatly simplifying the online shopping and payment process. Attached Figure Description

[0017] 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 schematic diagram of a payment system structure based on a large model is provided for embodiments of this specification; Figure 2 A flowchart of the payment method provided in the embodiments of this specification; Figure 3 A schematic diagram of a first type of payment device based on a large model, provided for embodiments of this specification; Figure 4 A schematic diagram of a second type of payment device based on a large model, provided for embodiments of this specification; Figure 5 A schematic diagram of a third type of payment device based on a large model provided in the embodiments of this specification. Figure 6 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this specification. Detailed Implementation

[0018] 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.

[0019] The technical solutions provided in the various embodiments of this specification are described in detail below with reference to the accompanying drawings.

[0020] Figure 1 This document provides a schematic diagram of a payment system architecture based on a large model, as illustrated in an embodiment of this specification. The payment system specifically includes: a third-party intelligent agent, a merchant intelligent agent, and a payment server. Wherein: The third-party intelligent agent has a pre-installed first LLM (Local Level Manager), which is used to interact with the user through at least one form, such as voice, text, or vision. The third-party intelligent agent can be installed on a user terminal. The user terminal described in the embodiments of this specification may include various screen-equipped terminals with display devices, such as personal computers, mobile phones, and tablets, or various screenless terminals without display devices, such as smart glasses, smart bracelets, smart headphones, and smart speakers. When the third-party intelligent agent is installed on a screenless terminal, the first LLM can interact with the user through voice.

[0021] In one possible embodiment, the front-end of the third-party intelligent agent can be installed on the user terminal, and the back-end of the third-party intelligent agent can be deployed on a back-end server. The first LLM is deployed on the back-end of the intelligent agent, and the user can interact with the first LLM deployed on the back-end of the third-party intelligent agent through the front-end of the third-party intelligent agent on the user terminal. In another possible embodiment, the third-party intelligent agent may not distinguish between the front-end and back-end, and the third-party intelligent agent containing the first LLM can be directly installed on the user terminal. For the sake of convenience, the front-end and back-end of the third-party intelligent agent will not be distinguished below.

[0022] The third-party intelligent agent can also integrate a payment software development kit (SDK) corresponding to the payment server. This payment SDK includes an application programming interface (API) for the payment server, allowing the third-party intelligent agent to interact with the payment server. The payment SDK can also directly provide at least some of the functions of the payment client installed on the user's terminal, enabling users to use these functions directly within the third-party intelligent agent's interface without switching to the payment client. Furthermore, the third-party intelligent agent can communicate with at least one merchant intelligent agent via an agent-to-agent (A2A) protocol. Through the A2A protocol, the third-party intelligent agent can interact with each merchant intelligent agent it interfaces with.

[0023] A merchant's intelligent agent connects to at least one of its merchant systems. A merchant may have its own stores in different geographical locations, each with its own independent merchant system to provide users with the online shopping functions offered by that store. In traditional online shopping scenarios, the merchant system displays information about the various products it sells. Users can select their desired products and place orders within the merchant system. The merchant system can also activate a payment client, and the payment for the purchased products is processed through the activated payment client and its corresponding payment server.

[0024] In the embodiments described in this specification, a merchant's intelligent agent interfaces with all of that merchant's systems. This intelligent agent has a pre-installed second LLM (Local Level Manager) that can interact with third-party intelligent agents via voice, text, vision, structured data, or unstructured data to determine the user's desired product and place an order for that product in at least one of the merchant systems it interfaces with. Thus, a user can directly interface with multiple different merchants' intelligent agents through a third-party intelligent agent installed on their terminal; that is, they can place orders with multiple merchant intelligent agents through a single third-party intelligent agent, without needing to individually interface with each of these different merchants' intelligent agents.

[0025] The payment server, located in the background, is used to collect the triples required for payment, namely the payment account, the receiving account, and the amount to be paid. It generates a payment order based on the triples required for payment and completes the payment for the order according to the triples.

[0026] To enable users to directly order their desired products from the merchant system through the third-party intelligent agent during their first LLM interaction with the third-party intelligent agent, and to also directly pay for the products through the third-party intelligent agent, without having to switch back and forth between the third-party intelligent agent, the merchant intelligent agent or the merchant system, and the payment client corresponding to the payment server, thus simplifying the user's operation in the shopping and payment process, based on the above... Figure 1 The system shown in this specification provides embodiments as follows: Figure 2 The payment methods shown.

[0027] Figure 2 The payment method flowchart provided for the embodiments of this specification specifically includes the following steps: S200: A third-party intelligent agent interacts with the user through a pre-configured first LLM.

[0028] In one embodiment provided in this specification, a user may input one or more forms of interactive information, such as text, voice, and vision, into a third-party intelligent agent. The third-party intelligent agent then analyzes the interactive information input by the user through the reasoning ability of the first LLM and / or a preset knowledge base, and outputs corresponding response information to the user to complete the interaction between the user and the first LLM.

[0029] S201: When the first LLM identifies the user's purchase intent based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intent.

[0030] In the embodiments described in this specification, the third-party intelligent agent can first identify the semantics of the user's interaction with the first LLM through the first LLM, and then identify the type of goods the user needs based on the semantics. This is because in general chat conversations, users mostly only say what type of goods they need, without deliberately emphasizing that they need goods from a specific merchant. For example, users will mostly say to the first LLM "I want to drink a cup of coffee", and may not deliberately emphasize "I want to drink coffee from a certain brand". Therefore, the first LLM can first identify the type of goods the user needs.

[0031] After identifying the product type, since the third-party agent may be connected to more than one merchant agent, the first LLM can determine the merchant agent that matches the identified product type based on the natural language descriptions of each merchant agent, and use this as the target merchant agent. Specifically, the description information of a merchant agent must include at least a description of the product type sold by the merchant in natural language. The first LLM can then determine the target merchant agent that matches the product type required by the user based on this description information.

[0032] If there are multiple merchant agents that match the type of goods the user needs, the first LLM can either use all of these merchant agents as the target merchant agent, or it can continue to interact with the user and ask the user about their preferences for these matching merchant agents, allowing the user to select the target merchant agent from these matching merchant systems. For example, the first LLM can reply to the user, "I can order coffee from brand A, brand B, and brand C for you. Which brand do you prefer?"

[0033] S202: Based on the A2A protocol, the merchant's intelligent agent sends the user's purchase intention to the target merchant.

[0034] Generally, there are two ways for third-party intelligent agents to connect with merchant systems. The first is that the third-party intelligent agent directly embeds the API corresponding to the merchant system. For example, if the third-party intelligent agent integrates the e-commerce SDK corresponding to the merchant system, and this e-commerce SDK contains the API corresponding to the merchant system, then the third-party intelligent agent can interact with the merchant system through the API corresponding to the merchant system based on Remote Procedure Call (RPC) communication. The second method involves the merchant also deploying its own merchant intelligent agent, which interfaces with all of the merchant's merchant systems. In this case, the third-party intelligent agent can directly connect to the merchant intelligent agent through the A2A protocol. This specification only discusses the second communication connection method in the embodiments, and does not limit the first communication connection method.

[0035] Once a third-party intelligent agent identifies the target merchant's intelligent agent based on the user's purchase intent, it can directly send that intent to the target merchant's intelligent agent. Specifically, the user's purchase intent can be sent to the target merchant's intelligent agent in natural language text format, or it can be converted into structured data (such as JSON data) before being sent.

[0036] In this scenario, a third-party intelligent agent can first log in to the target merchant's intelligent agent using the user's user ID, and then send the aforementioned purchase intention to the target merchant's intelligent agent using the user ID that is logged into the target merchant's intelligent agent.

[0037] Specifically, a third-party intelligent agent can first send a login request carrying the user's identifier to the target merchant's intelligent agent. At this point, the user's identifier becomes the user's login account used to log in to the target merchant's intelligent agent. The target merchant's intelligent agent can then verify the user identifier based on the login request, including but not limited to password verification, voiceprint verification, and facial recognition. Upon successful login verification, the intelligent agent allows the third-party intelligent agent to log in to the target merchant's intelligent agent using the user identifier, generates a login status maintenance token, and then returns the login status maintenance token to the third-party intelligent agent.

[0038] The login status maintenance token is a token used to maintain the login status of a third-party agent using the user's user ID to log in to the merchant's intelligent agent. This is because the entity logging into the merchant's intelligent agent at this time is not the merchant's intelligent agent client installed on the user's terminal, but a third-party intelligent agent independent of the merchant's intelligent agent. Therefore, the merchant's intelligent agent needs to send this login status maintenance token to the third-party intelligent agent to maintain the third-party intelligent agent's login status using the user's user ID. Specifically, when the merchant's intelligent agent generates the login status maintenance token, it can set a validity period for the token, such as 30 minutes. Each time the third-party intelligent agent interacts with the merchant's intelligent agent, it must carry the login status maintenance token. The merchant's intelligent agent then checks whether the login status maintenance token sent by the third-party intelligent agent is a login status maintenance token already generated by the merchant's intelligent agent, and whether the login status maintenance token is within its validity period. If both checks are yes, the merchant's intelligent agent accepts the interaction with the third-party intelligent agent for that time; otherwise, it rejects the interaction with the third-party intelligent agent.

[0039] The second function of the login status maintenance token is to enable the merchant agent to know which user the third-party agent is representing in the current interaction. Specifically, when the merchant agent generates the login status maintenance token, it can establish a correspondence between the token and the user identifier. Each time the third-party agent interacts with the merchant agent, it must carry this login status maintenance token. The merchant agent can then determine the user represented by the third-party agent in this interaction based on the user identifier corresponding to the login status maintenance token sent by the third-party agent.

[0040] In step S202, after the third-party intelligent agent receives the login status maintenance token returned by the merchant intelligent agent, it can send the user's purchase intention and the login status maintenance token to the target merchant's intelligent agent.

[0041] S203: The merchant agent of the target merchant uses a pre-set second language model to determine the target product that matches the received purchase intention among the products sold by the target merchant.

[0042] After receiving the purchase intent and login status maintenance token sent by the third-party intelligent agent, the merchant intelligent agent can first determine whether the login status maintenance token is valid (i.e., whether the login status maintenance token is generated by the merchant intelligent agent itself, and whether the login status maintenance token is within its validity period). If it is valid, the merchant intelligent agent can use the second LLM pre-set in the merchant intelligent agent to determine the target product that matches the received purchase intent among the products sold by the target merchant.

[0043] Specifically, the product information of each item sold by the target merchant can be pre-stored in the knowledge base of the second LLM. After receiving the purchase intention sent by the third-party intelligent agent, the second LLM can query the product information of each item stored in the knowledge base, and based on the reasoning ability of the second LLM, select the target product that matches the purchase intention according to the purchase intention and the product information of each item.

[0044] Furthermore, since the different stores of the target merchant are located in different geographical locations, this also means that each merchant system connected to the merchant agent is located in a different geographical location. Therefore, the third-party agent can also obtain its own current geographical location and send this current geographical location along with the aforementioned login status maintenance token to the merchant agent. After verifying the validity of the login status maintenance token, the merchant agent, based on the third-party agent's current geographical location and the geographical locations of each merchant system connected to the target merchant, determines each candidate merchant system and returns the information of each candidate merchant system to the third-party agent. Specifically, when determining candidate merchant systems, the merchant agent can identify merchant systems whose distance from the third-party agent's current geographical location is within a preset distance range.

[0045] The third-party intelligent agent can then display information about each candidate merchant system and, based on the user's actions, determine the target merchant system selected by the user from among the candidate systems. It then returns the target merchant system information, along with the aforementioned login status maintenance token, to the target merchant's intelligent agent. Specifically, when the user's terminal is a screenless terminal, the third-party intelligent agent can display information about each candidate merchant system via voice broadcast and determine the target merchant system selected by the user based on their voice.

[0046] After the merchant agent determines that the login status maintenance token is valid, it can use a second LLM to identify the target product that matches the received purchase intention among the products sold in the target merchant system.

[0047] The step of the aforementioned third-party intelligent agent determining its current location and sending it to the merchant intelligent agent can occur either before or after sending the purchase intent and login status maintenance token to the merchant intelligent agent; this specification does not impose any restrictions on this.

[0048] Of course, third-party intelligent agents can also directly receive the target merchant system information input by the user and send the target merchant system information and login status maintenance token to the merchant intelligent agent, so that the merchant intelligent agent can directly determine the target merchant system without sending its own current geographical location to the merchant intelligent agent. This will not be elaborated on here.

[0049] S204: Send a payment request to the payment server based on the target merchant's payment account and the amount due for payment corresponding to the target product.

[0050] After identifying a target product that matches the user's purchase intent, the merchant's intelligent agent can first return the product information to a third-party intelligent agent, which then displays that product information. When the third-party intelligent agent receives an order instruction from the user regarding the target product's information, it sends the order instruction and the aforementioned login status maintenance token to the merchant's intelligent agent. Especially when the merchant's intelligent agent identifies more than one target product, the third-party intelligent agent needs to determine the target product corresponding to the user's order instruction among these products before sending that order instruction, the corresponding product information, and the login status maintenance token to the merchant's intelligent agent. When the user's terminal is a screenless terminal, the third-party intelligent agent can display the target product information via voice announcement and determine whether it has received an order instruction from the user regarding a specific target product based on the user's voice.

[0051] After confirming the legitimacy of the login status maintenance token, the merchant agent, on the one hand, identifies the user identifier corresponding to the login status maintenance token and sends an order instruction for the target product to the target merchant system based on the user identifier. This causes the target merchant system to generate a fulfillment order for the target product based on the user identifier. On the other hand, it can also generate a payment request carrying the pre-stored payment account and the amount to be paid for the target product based on the pre-stored payment account of the target merchant and the amount to be paid for the target product, and send it to the payment server. In other words, in this embodiment of the specification, the user's payment process for the target product is initiated by the merchant agent.

[0052] The aforementioned fulfillment order is an order used by the target merchant's system to confirm that the user has purchased the target product and to order the product to be shipped to the user. It is not a payment order used by the payment server to pay for the target product. When generating this fulfillment order, the target merchant's system can query the pre-saved logistics address and / or contact information corresponding to the user identifier based on the user identifier, and generate a fulfillment order that includes at least the product information of the target product and the logistics address and / or contact information corresponding to the user identifier.

[0053] S205: In response to the payment request, the payment server creates a payment order that includes the receiving account and the amount to be paid.

[0054] After receiving a payment request from the merchant agent, that is, after the merchant agent initiates the user's payment process for the target product, the payment server can generate a payment order containing the payment account and the amount to be paid, based on the payment account and the amount to be paid carried in the payment request, and at the same time generate an order identifier for the payment order.

[0055] At this point, the receiving account and the amount to be paid are complete in the triple required for payment, but the payment account has not yet been collected.

[0056] S206: Return the order identifier corresponding to the payment order to the merchant's intelligent agent.

[0057] S207: The merchant agent of the target merchant returns the order identifier returned by the payment server to the third-party agent.

[0058] Since the payment process for the target product is initiated by the merchant agent in this specification, and the payment account is still missing from the triple required for payment, the payment server can return the order identifier corresponding to the payment order to the merchant agent after creating a payment order containing only the receiving account and the amount to be paid in step S205. The merchant agent will then return the order identifier to the third-party agent.

[0059] S208: The third-party intelligent agent obtains the payment authorization message entered by the user for the order identifier, and sends the payment authorization message to the payment server.

[0060] In this specification, a payment SDK is integrated into the third-party intelligent agent. This payment SDK can directly provide users with at least some of the functions of the payment client corresponding to the payment server within the third-party intelligent agent. These functions include user identity verification.

[0061] After receiving the order identifier of the payment order created by the payment server forwarded by the merchant's intelligent agent, the third-party intelligent agent responds to the trigger of the order identifier by prompting the user to enter a payment authorization message for the order identifier. After obtaining the user's payment authorization message for the order identifier, the third-party intelligent agent verifies the user's identity through the payment SDK. After the identity verification is successful, the payment authorization message is sent to the payment server.

[0062] When the user terminal where the third-party intelligent agent is located is a screenless terminal, prompting the user to enter the aforementioned payment authorization message can also be done by broadcasting a preset voice message. User verification can also be performed using the user's voiceprint. Specifically, the third-party intelligent agent can first log in to the payment server using the user's user identifier. This user identifier used to log in to the payment server can be the same as or different from the user identifier used to log in to the merchant intelligent agent; this manual only illustrates the case of a different identifier.

[0063] After a third-party intelligent agent logs into the payment server using the user's user ID, it can collect the user's voiceprint through the payment SDK and upload the collected voiceprint as a standard voiceprint to the payment server. The payment server saves the standard voiceprint and establishes an association between the standard voiceprint and the user's user ID. Alternatively, the third-party intelligent agent can directly save the standard voiceprint locally. The third-party intelligent agent can also upload the device ID of the user's terminal to the payment server after logging into the payment server using the user's user ID. The payment server saves the device ID and establishes an association between the device ID and the user's user ID. In this way, the payment server saves the association between the user's user ID and the standard voiceprint, as well as the correspondence between the user ID and the device ID. The device ID mentioned in this specification may include, but is not limited to, the MAC address, device serial number, or any other device identifier that can uniquely identify the user's terminal. This specification does not impose any restrictions on this.

[0064] If the third-party intelligent agent stores the user's standard voiceprint locally, when verifying the user's identity through the payment SDK, the third-party intelligent agent can collect the user's current voiceprint through the payment SDK, and verify the user's identity based on the collected current voiceprint and the pre-saved standard voiceprint. After the identity verification is successful, the device identifier of the user's terminal and the payment authorization message are sent to the payment server.

[0065] If the payment server stores the user's standard voiceprint, then when a third-party intelligent agent verifies the user's identity through the payment SDK, it can collect the user's current voiceprint through the payment SDK and send the current voiceprint, the device identifier of the user's terminal, and the payment authorization message to the payment server.

[0066] S209: In response to the payment authorization message, the payment server queries the user's payment account and uses the payment account to settle the payment order.

[0067] If the third-party intelligent agent locally stores the user's standard voiceprint, the payment server, upon receiving the device identifier and payment authorization message sent by the third-party intelligent agent, can respond to the triggering of the payment authorization message by querying the user identifier corresponding to the device identifier, and then querying the user's payment account based on the user identifier. After finding the payment account, it can be added to the payment order created in step S205. At this point, the required triple for payment is complete, and the payment terminal can use the payment account to settle the payment order, that is, transfer the amount corresponding to the amount to be paid in the payment account to the receiving account.

[0068] If the payment server stores the user's standard voiceprint, upon receiving the current voiceprint, the user's terminal device identifier, and the payment authorization message from the third-party intelligent agent, the payment server, in response to the triggering of the payment authorization message, queries the user identifier corresponding to the device identifier, then queries the stored standard voiceprint corresponding to the user identifier based on the user identifier, and finally determines the user's identity verification result based on the current voiceprint sent by the third-party intelligent agent and the queried standard voiceprint. If the result is successful, the payment server queries the user's payment account based on the user identifier. After finding the payment account, it can be added to the payment order created in step S205. At this point, the required three-tuple for payment is complete, and the payment terminal can use the payment account to settle the payment order, that is, transfer the amount corresponding to the amount to be paid in the payment account to the receiving account.

[0069] It should be noted that the user identifier used by the user to log in to the payment service as described in this specification is not the same concept as the user's payment account. For payment security, the user's payment account information is not the same as described above. Figure 2 Throughout the entire interaction process shown, the third-party intelligent agent did not appear. The payment authorization message that the user entered on the third-party intelligent agent for the aforementioned order identifier is actually an authorization message used to authorize the payment server to allow the user's payment account to make payment for the payment order.

[0070] Furthermore, regardless of whether the aforementioned user terminals are screen-equipped or screenless terminals, in Figure 2 Throughout the entire interaction process shown, the user terminal can remain in a state where only the third-party intelligent agent is activated. It is not necessary to activate the merchant client corresponding to any merchant system, nor the payment client corresponding to the payment server. The interaction with the merchant system is completed through the interaction with the corresponding merchant intelligent agent, while the interaction with the payment server can be completed through the payment SDK.

[0071] Through the aforementioned payment system and methods, users can place orders directly using a third-party intelligent agent and also pay for the purchase through the same agent, greatly simplifying the online shopping and payment process for users.

[0072] The above describes a payment system and method based on a large model, as provided in the embodiments of this specification. Based on the same idea, this specification also provides corresponding devices, storage media, and electronic devices.

[0073] Figure 3 This is a schematic diagram of a first type of payment device based on a large model, provided in the embodiments of this specification. This first payment device can be applied to a third-party intelligent agent, and the device includes: Interaction module 300 is used to interact with the user through a pre-built first language model; Merchant determination module 301 is used to determine at least one target merchant corresponding to the purchase intention when the first language model identifies the user's purchase intention based on the interaction with the user; The sending module 302 is used to send the purchase intention to the merchant agent of the target merchant based on the inter-agent A2A protocol, so that the merchant agent of the target merchant can determine the target product that matches the purchase intention from the products sold by the target merchant through the second language model, and send a payment request to the payment server according to the payment account of the target merchant and the amount to be paid corresponding to the target product. The receiving module 303 is used to receive the order identifier returned by the payment server. The order identifier is the order identifier corresponding to the payment order returned by the payment server through the merchant agent after the payment server responds to the payment request, creates a payment order containing the payment account and the amount to be paid; The interaction module 300 is also used to obtain the payment authorization message entered by the user for the order identifier; The sending module 302 is further configured to send the payment authorization message to the payment server, so that the payment server responds to the payment authorization message, queries the user's payment account, and uses the payment account to settle the payment order.

[0074] Optionally, the sending module 302 is further configured to send a login request carrying the user's user identifier to the merchant agent of the target merchant before sending the purchase intention to the merchant agent of the target merchant; The receiving module 303 is further configured to receive a login status maintenance token returned by the merchant agent of the target merchant, wherein the login status maintenance token is generated and returned by the merchant agent of the target merchant after the login verification of the user identifier is passed according to the login request, while allowing the third-party agent to log in to the merchant agent of the target merchant through the user identifier; and to store the login status maintenance token. The sending module 302 is specifically used to send the purchase intention and the login status maintenance token to the merchant agent of the target merchant.

[0075] Optionally, the merchant agent of the target merchant is connected to at least one merchant system of the target merchant, and the different merchant systems of the target merchant are located in different geographical locations; The sending module 302 is specifically configured to, after the merchant determination module 301 determines at least one target merchant corresponding to the purchase intention, obtain the current geographical location of the third-party intelligent agent; send the current geographical location to the merchant intelligent agent of the target merchant; receive and display the information of each candidate merchant system returned by the merchant intelligent agent of the target merchant, wherein each candidate merchant system is determined by the merchant intelligent agent of the target merchant based on the current geographical location of the third-party intelligent agent and the geographical location of each merchant system of the target merchant; determine the information of the target merchant system selected by the user from the information of each candidate merchant system; and return the information of the target merchant system to the merchant intelligent agent of the target merchant, so that the merchant intelligent agent of the target merchant can determine the target product matching the received purchase intention from the products sold by the target merchant system through a preset second language model.

[0076] Optionally, the sending module 302 is further configured to: receive and display the product information of the target product returned by the merchant agent of the target merchant before receiving the order identifier returned by the payment server; when receiving the order instruction input by the user for the product information of the target product, send the order instruction to the merchant agent of the target merchant, so that the merchant agent of the target merchant sends the order instruction for the target product to the target merchant system, and sends a payment request to the payment server according to the payment account of the target merchant and the amount to be paid corresponding to the target product.

[0077] Optionally, the third-party intelligent agent integrates a payment SDK; The sending module 302 is specifically used to verify the user's identity through the payment SDK; after the identity verification is successful, it sends the payment authorization message to the payment server.

[0078] Optionally, the third-party intelligent agent is installed on the user's user terminal, and the user terminal includes a screenless terminal; The payment server pre-stores the correspondence between the device identifier of the user terminal and the user identifier of the user; The sending module 302 is specifically used to collect the user's current voiceprint; and to verify the user's identity based on the current voiceprint and a standard voiceprint pre-saved for the user. The sending module 302 is specifically used to send the device identifier of the user terminal and the payment authorization message to the payment server, so that the payment server responds to the payment authorization message, queries the user identifier corresponding to the device identifier, and queries the user's payment account based on the user identifier.

[0079] Figure 4 This is a schematic diagram of a second type of payment device based on a large model, provided in the embodiments of this specification. This second type of payment device can be applied to merchant intelligent agents, and the device includes: The receiving module 400 is used to receive the user's purchase intention sent by a third-party intelligent agent based on the inter-intelligent agent A2A protocol; the purchase intention is the purchase intention of at least the target merchant identified by the third-party intelligent agent when interacting with the user through a preset first language model. The product selection module 401 is used to determine, through a pre-set second language model, a target product that matches the received purchase intention from the products sold by the target merchant. The payment initiation module 402 is used to send a payment request to the payment server based on the payment account of the target merchant and the amount to be paid corresponding to the target product. The forwarding module 403 is used to receive the order identifier returned by the payment server and send the order identifier to the third-party intelligent agent, so that the third-party intelligent agent can obtain the payment authorization message input by the user for the order identifier, and send the payment authorization message to the payment server, so that the payment server can respond to the payment authorization message, query the user's payment account, and use the payment account to perform payment settlement for the payment order corresponding to the order identifier; the order identifier is the order identifier of the payment order containing the receiving account and the amount to be paid, created by the payment server in response to the payment request.

[0080] Optionally, the receiving module 400 is further configured to, before receiving the user's purchase intention sent by the third-party intelligent agent based on the inter-intelligent A2A protocol, receive a login request sent by the third-party intelligent agent with the user's user identifier; perform login verification on the user identifier according to the login request; when the login verification is successful, allow the third-party intelligent agent to log in to the merchant intelligent agent of the target merchant through the user identifier, generate a login status maintenance token, return the login status maintenance token to the third-party intelligent agent, and enable the third-party intelligent agent to store the login status maintenance token; The receiving module 400 is specifically used to receive the user's purchase intention and the login status maintenance token sent by the third-party intelligent agent based on the inter-intelligent A2A protocol.

[0081] Optionally, the merchant agent of the target merchant is connected to at least one merchant system of the target merchant, and the different merchant systems of the target merchant are located in different geographical locations; The product selection module 401 is specifically used to: receive the current geographical location of the third-party intelligent agent; determine candidate merchant systems among the merchant systems based on the current geographical location of the third-party intelligent agent and the geographical locations of the target merchant's various merchant systems; return the information of each candidate merchant system to the third-party intelligent agent; receive the information of the target merchant system returned by the third-party intelligent agent, wherein the information of the target merchant system is the information of the target merchant system selected by the user from the information of each candidate merchant system after the third-party intelligent agent displays the information of each candidate merchant system; and determine the target product that matches the received purchase intention among the products sold by the target merchant system using a preset second language model.

[0082] Optionally, the payment initiation module 402 is specifically configured to: return the determined product information of the target product to the third-party intelligent agent; receive the order instruction input by the user regarding the product information of the target product sent by the third-party intelligent agent, wherein the order instruction sent by the third-party intelligent agent is sent after the third-party intelligent agent displays the product information of the target product and receives the order instruction input by the user regarding the product information of the target product; send the order instruction for the target product to the target merchant system, and send a payment request to the payment server according to the target merchant's payment account and the amount to be paid corresponding to the target product.

[0083] Figure 5 This is a schematic diagram of a third type of payment device based on a large model, provided in the embodiments of this specification. This third type of payment device can be applied to a payment server. The device includes: The receiving module 500 is used to receive a payment request sent by the merchant agent of the target merchant. The payment request is sent by the merchant agent of the target merchant after determining the target product that matches the user's purchase intention among the products sold by the target merchant through a preset second language model, and based on the target merchant's payment account and the amount to be paid corresponding to the target product. The user's purchase intention is determined by a third-party agent interacting with the user through a preset first language model. When the first language model recognizes the user's purchase intention based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intention and sends it to the merchant agent of the target merchant based on the A2A protocol. The order creation module 501 is used to create a payment order containing the payment account and the amount to be paid for the target product based on the payment request containing the payment account of the target merchant and the amount to be paid for the target product. Sending module 502 is used to return the order identifier corresponding to the payment order to the third-party intelligent agent through the merchant intelligent agent; The receiving module 500 is further configured to receive a payment authorization message sent by the third-party intelligent agent, which is input by the user for the order identifier; Query module 503 is used to query the user's payment account in response to the payment authorization message; The payment module 504 is used to perform payment settlement on the payment order using the payment account.

[0084] Optionally, the third-party intelligent agent is installed on the user's user terminal, and the user terminal includes a screenless terminal; The payment server pre-stores the correspondence between the device identifier of the user terminal and the user identifier of the user; The receiving module 500 is specifically used to receive the device identifier of the user terminal and the payment authorization message entered by the user for the order identifier sent by the third-party intelligent agent; The query module 503 is specifically used to query the user identifier corresponding to the device identifier; and to query the user's payment account based on the queried user identifier.

[0085] 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 large-model-based payment method described above.

[0086] based on Figure 2 The payment method based on a large model shown in this specification also provides embodiments. Figure 6 The diagram shows the structure of the electronic device. Figure 6 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 payment method based on a large model.

[0087] 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 payment system based on a large model, the system comprising: A third-party intelligent agent, a merchant intelligent agent, and a payment server; the third-party intelligent agent communicates with at least one merchant's intelligent agent based on an inter-agent A2A protocol; wherein: The third-party intelligent agent is used to interact with the user through a pre-set first large language model. When the first large language model identifies the user's purchase intention based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intention and sends the purchase intention to the merchant intelligent agent of the target merchant based on the A2A protocol. The merchant agent of the target merchant is used to determine the target product that matches the received purchase intention among the products sold by the target merchant through a pre-set second language model, and send a payment request to the payment server according to the target merchant's payment account and the amount to be paid corresponding to the target product; In response to the payment request, the payment server creates a payment order containing the receiving account and the amount to be paid, and returns the order identifier corresponding to the payment order to the merchant agent. The merchant agent of the target merchant is also used to return the order identifier returned by the payment server to the third-party agent; The third-party intelligent agent is also used to obtain the payment authorization message entered by the user for the order identifier, and send the payment authorization message to the payment server; The payment server, in response to the payment authorization message, queries the user's payment account and uses the payment account to settle the payment order.

2. The system as described in claim 1, wherein the third-party intelligent agent is further configured to send a login request carrying the user's identifier to the merchant intelligent agent of the target merchant; The merchant agent of the target merchant is also used to perform login verification on the user identifier according to the login request, and when the login verification is successful, allow the third-party agent to log in to the merchant agent of the target merchant through the user identifier, generate a login status maintenance token, and return the login status maintenance token to the third-party agent. Specifically, the third-party intelligent agent is used to store the login status maintenance token and, based on the A2A protocol, send the purchase intention and the login status maintenance token to the merchant intelligent agent of the target merchant.

3. The system as described in claim 1, wherein the merchant agent of the target merchant is connected to at least one merchant system of the target merchant, and the different merchant systems of the target merchant are located in different geographical locations; The third-party intelligent agent is also used to obtain its own current geographical location and send the current geographical location to the merchant intelligent agent of the target merchant; The target merchant's intelligent agent is used to determine each candidate merchant system in each merchant system based on the current geographical location of the third-party intelligent agent and the geographical location of each merchant system of the target merchant, and return the information of each candidate merchant system to the third-party intelligent agent; The third-party intelligent agent is also used to display information of each candidate merchant system, determine the information of the target merchant system selected by the user from the information of each candidate merchant system, and return the information of the target merchant system to the merchant intelligent agent of the target merchant. The merchant agent of the target merchant is specifically used to determine, through a pre-set second language model, the target product sold in the target merchant system that matches the received purchase intention.

4. In the system as described in claim 3, the merchant agent of the target merchant is specifically used to return the determined product information of the target product to the third-party agent; The third-party intelligent agent is also used to display the product information of the target product, and when it receives the order instruction input by the user for the product information of the target product, it sends the order instruction to the merchant intelligent agent of the target merchant; The merchant agent of the target merchant is specifically used to send an order instruction for the target product to the target merchant system, and to send a payment request to the payment server based on the target merchant's payment account and the amount to be paid corresponding to the target product.

5. The system as described in claim 1, wherein the third-party intelligent agent integrates a payment SDK; The third-party intelligent agent is specifically used to verify the user's identity through the payment SDK, and after the identity verification is successful, send the payment authorization message to the payment server.

6. The system as described in claim 5, wherein the third-party intelligent agent is installed on the user's user terminal, and the user terminal includes a screenless terminal; The payment server pre-stores the correspondence between the device identifier of the user terminal and the user identifier of the user; The third-party intelligent agent is specifically used to collect the user's current voiceprint through the payment SDK, and to verify the user's identity based on the current voiceprint and the standard voiceprint pre-saved for the user. After the identity verification is successful, the device identifier of the user's terminal and the payment authorization message are sent to the payment server. The payment server is specifically used to, in response to the payment authorization message, query the user identifier corresponding to the device identifier, and query the user's payment account based on the user identifier.

7. A payment method based on a large model, the method comprising: The third-party intelligent agent interacts with the user through a pre-built primary language model; When the first language model identifies the user's purchase intention based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intention; Based on the A2A protocol between intelligent agents, the purchase intention is sent to the merchant intelligent agent of the target merchant. The merchant intelligent agent of the target merchant uses the second language model to determine the target product that matches the purchase intention from the products sold by the target merchant, and sends a payment request to the payment server according to the target merchant's payment account and the amount to be paid corresponding to the target product. Receive the order identifier returned by the payment server. The order identifier is the order identifier corresponding to the payment order returned by the payment server through the merchant smart agent after the payment server responds to the payment request, creates a payment order containing the payment account and the amount to be paid; Obtain the payment authorization message entered by the user in response to the order identifier; The payment authorization message is sent to the payment server, which responds to the payment authorization message by querying the user's payment account and using the payment account to settle the payment order.

8. The method of claim 7, wherein before sending the purchase intention to the merchant agent of the target merchant, the method further comprises: Send a login request carrying the user's identifier to the merchant agent of the target merchant; Receive the login status maintenance token returned by the merchant agent of the target merchant. The login status maintenance token is generated and returned by the merchant agent of the target merchant after the login request verifies the user identifier and allows the third-party agent to log in to the merchant agent of the target merchant through the user identifier. Store the login status maintenance token; Sending the purchase intent to the target merchant's intelligent agent specifically includes: The purchase intent and the login status maintenance token are sent to the merchant agent of the target merchant.

9. The method as described in claim 7, wherein the merchant agent of the target merchant is connected to at least one merchant system of the target merchant, and the different merchant systems of the target merchant are located in different geographical locations; After determining at least one target merchant corresponding to the purchase intention, the method further includes: Obtain the current geographical location of the third-party intelligent agent; Send the current geographic location to the merchant's intelligent agent at the target merchant; Receive and display information on each candidate merchant system returned by the merchant agent of the target merchant, wherein each candidate merchant system is determined by the merchant agent of the target merchant based on the current geographical location of the third-party agent and the geographical location of each merchant system of the target merchant. Determine the information of the target merchant system selected by the user from the information of each candidate merchant system; The information from the target merchant system is returned to the target merchant's intelligent agent, enabling the target merchant's intelligent agent to determine, through a pre-set second language model, the target product that matches the received purchase intention among the products sold by the target merchant system.

10. The method of claim 9, further comprising, before receiving the order identifier returned by the payment server: Receive and display the product information of the target product returned by the merchant's intelligent agent; Upon receiving an order instruction from the user regarding the product information of the target product, the order instruction is sent to the merchant agent of the target merchant, causing the merchant agent to send an order instruction for the target product to the target merchant system, and to send a payment request to the payment server based on the target merchant's payment account and the amount to be paid corresponding to the target product.

11. The method as described in claim 7, wherein the third-party intelligent agent integrates a payment SDK; Sending the payment authorization message to the payment server specifically includes: The user is verified using the payment SDK; After identity verification is successful, the payment authorization message is sent to the payment server.

12. The method of claim 11, wherein the third-party intelligent agent is installed on the user's user terminal, and the user terminal includes a screenless terminal; The payment server pre-stores the correspondence between the device identifier of the user terminal and the user identifier of the user; Verifying the identity of the user specifically includes: Collect the user's current voiceprint; The user is verified based on the current voiceprint and the standard voiceprint pre-saved for the user; Sending the payment authorization message to the payment server specifically includes: The device identifier of the user terminal and the payment authorization message are sent to the payment server, so that the payment server responds to the payment authorization message, queries the user identifier corresponding to the device identifier, and queries the user's payment account based on the user identifier.

13. A payment method based on a large model, the method comprising: The target merchant's intelligent agent receives the user's purchase intent sent by a third-party intelligent agent based on the inter-agent A2A protocol; The purchase intent is the purchase intent at least corresponding to the target merchant that the third-party intelligent agent identifies based on its interaction with the user when interacting with the user through a pre-set first language model. Using a pre-set second language model, target products that match the received purchase intent are identified among the products sold by the target merchant. Based on the target merchant's payment account and the outstanding amount corresponding to the target product, a payment request is sent to the payment server; The system receives an order identifier returned by the payment server and sends the order identifier to the third-party intelligent agent. The third-party intelligent agent then obtains the payment authorization message entered by the user for the order identifier and sends the payment authorization message to the payment server. In response to the payment authorization message, the payment server queries the user's payment account and uses the payment account to settle the payment order corresponding to the order identifier. The order identifier is an order identifier created by the payment server in response to the payment request, containing the receiving account and the amount to be paid for the payment order.

14. The method of claim 13, prior to receiving the user's purchase intention sent by a third-party intelligent agent based on an inter-agent A2A protocol, the method further includes: Receive a login request sent by the third-party intelligent agent, which includes the user's identifier. Login verification is performed on the user identifier based on the login request; Upon successful login verification, the third-party intelligent agent is allowed to log in to the merchant intelligent agent of the target merchant through the user identifier, and a login status maintenance token is generated. The login status maintenance token is then returned to the third-party intelligent agent, which stores the login status maintenance token. Receive user purchase intentions sent by third-party intelligent agents based on the inter-agent A2A protocol, specifically including: Receive the user's purchase intent and the login status maintenance token sent by the third-party intelligent agent based on the inter-intelligent A2A protocol.

15. The method of claim 13, wherein the merchant agent of the target merchant is connected to at least one merchant system of the target merchant, and the different merchant systems of the target merchant are located in different geographical locations; Among the products sold by the target merchant, the target products that match the received purchase intention are identified, specifically including: Receive the current geographical location of the third-party intelligent agent sent by the third-party intelligent agent; Based on the current geographical location of the third-party intelligent agent and the geographical location of each merchant system of the target merchant, each candidate merchant system is determined within each merchant system; The information of each candidate merchant system is returned to the third-party intelligent agent; The system receives information about the target merchant system returned by the third-party intelligent agent. This information is the target merchant system selected by the user from the information of each candidate merchant system after the third-party intelligent agent displays the information of each candidate merchant system. Using a pre-built second language model, the target product that matches the received purchase intention is identified among the products sold by the target merchant system.

16. The method as described in claim 15, wherein a collection request is sent to the payment server based on the target merchant's payment account and the outstanding amount corresponding to the target product, specifically including: The product information of the identified target product is returned to the third-party intelligent agent; The system receives an order instruction from the user regarding the product information of the target product, sent by the third-party intelligent agent. The order instruction sent by the third-party intelligent agent is sent after the third-party intelligent agent displays the product information of the target product and receives the order instruction from the user regarding the product information of the target product. Send an order instruction for the target product to the target merchant system, and send a payment request to the payment server based on the target merchant's payment account and the amount to be paid for the target product.

17. A payment method based on a large model, the method comprising: The payment server receives a payment request from the target merchant's smart agent; The payment request is sent by the merchant agent of the target merchant through a pre-set second language model, after identifying the target product that matches the user's purchase intention among the products sold by the target merchant, and based on the target merchant's payment account and the amount to be paid corresponding to the target product. The user's purchase intention is determined by a third-party agent interacting with the user through a pre-set first language model. When the first language model identifies the user's purchase intention based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intention and sends it to the merchant agent of the target merchant based on the A2A protocol. Based on the payment request containing the target merchant's payment account and the amount to be paid corresponding to the target product, create a payment order containing the payment account and the amount to be paid; The order identifier corresponding to the payment order is returned to the third-party intelligent agent through the merchant intelligent agent; Receive the payment authorization message input by the user for the order identifier, sent by the third-party intelligent agent; In response to the payment authorization message, query the user's payment account; The payment account is used to settle the payment order.

18. The method of claim 17, wherein the third-party intelligent agent is installed on the user's user terminal, and the user terminal includes a screenless terminal; The payment server pre-stores the correspondence between the device identifier of the user terminal and the user identifier of the user; Receiving the payment authorization message input by the user for the order identifier from the third-party intelligent agent, specifically including: Receive the device identifier of the user terminal and the payment authorization message entered by the user for the order identifier from the third-party intelligent agent; Querying the user's payment account specifically includes: Query the user identifier corresponding to the device identifier; Based on the user identifier obtained, query the user's payment account.

19. A payment device based on a large model, the device comprising: The interaction module is used to interact with the user through a pre-built primary language model; The merchant identification module is used to identify at least one target merchant corresponding to the purchase intention when the first language model identifies the user's purchase intention based on the interaction with the user. The sending module is used to send the purchase intention to the merchant agent of the target merchant based on the A2A protocol between intelligent agents. The merchant agent of the target merchant uses the second language model to determine the target product that matches the purchase intention from the products sold by the target merchant, and sends a payment request to the payment server according to the payment account of the target merchant and the amount to be paid corresponding to the target product. The receiving module is used to receive the order identifier returned by the payment server. The order identifier is the order identifier corresponding to the payment order returned by the payment server through the merchant agent after the payment server responds to the payment request, creates a payment order containing the payment account and the amount to be paid; The interaction module is also used to obtain the payment authorization message entered by the user for the order identifier; The sending module is further configured to send the payment authorization message to the payment server, so that the payment server responds to the payment authorization message, queries the user's payment account, and uses the payment account to settle the payment order.

20. A payment device based on a large model, the device comprising: The receiving module is used to receive the user's purchase intention sent by a third-party intelligent agent based on the inter-intelligent A2A protocol; The purchase intent is the purchase intent at least corresponding to the target merchant that the third-party intelligent agent identifies based on its interaction with the user when interacting with the user through a pre-set first language model. The product selection module is used to determine, through a pre-set second language model, target products that match the received purchase intent from among the products sold by the target merchant. The payment initiation module is used to send a payment request to the payment server based on the target merchant's payment account and the amount to be paid corresponding to the target product; The forwarding module is used to receive the order identifier returned by the payment server and send the order identifier to the third-party intelligent agent, so that the third-party intelligent agent can obtain the payment authorization message entered by the user for the order identifier, and send the payment authorization message to the payment server, so that the payment server can respond to the payment authorization message, query the user's payment account, and use the payment account to perform payment settlement for the payment order corresponding to the order identifier; the order identifier is the order identifier of the payment order, which includes the receiving account and the amount to be paid, created by the payment server in response to the payment request.

21. A payment device based on a large model, the device comprising: The receiving module is used to receive payment requests sent by the merchant's intelligent agent from the target merchant; The payment request is sent by the merchant agent of the target merchant through a pre-set second language model, after identifying the target product that matches the user's purchase intention among the products sold by the target merchant, and based on the target merchant's payment account and the amount to be paid corresponding to the target product. The user's purchase intention is determined by a third-party agent interacting with the user through a pre-set first language model. When the first language model identifies the user's purchase intention based on the interaction with the user, it determines at least one target merchant corresponding to the purchase intention and sends it to the merchant agent of the target merchant based on the A2A protocol. The order creation module is used to create a payment order containing the payment account and the amount to be paid for the target product, based on the payment request containing the target merchant's payment account and the amount to be paid for the target product. The sending module is used to return the order identifier corresponding to the payment order to the third-party intelligent agent through the merchant intelligent agent; The receiving module is further configured to receive a payment authorization message sent by the third-party intelligent agent, which is input by the user for the order identifier; The query module is used to query the user's payment account in response to the payment authorization message; The payment module is used to process payment settlements for the payment orders using the payment account.

22. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in any one of claims 7-20.

23. 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 described in any one of claims 7-20.

24. A computer program product comprising a computer program that, when executed by a processor, implements the method described in any one of claims 7-19.

Images