Air ticket sales service processing method and system, electronic equipment and storage medium

Abstract

The invention relates to the technical field of computers, and provides an air ticket sales service processing method and system, electronic equipment and a storage medium, and the method comprises the steps: determining the demand information of a user in an air ticket sales service; based on the demand information and the cue word template, in combination with an NDC knowledge base tool, calling an LLM (Large Language Model), and generating a request message of an NDC interface; calling an NDC interface tool by using the request message for generating the NDC interface to obtain an NDC interface response message; based on the NDC interface response message and the cue word template, in combination with an NDC knowledge base tool, calling LLM, and analyzing content information matched with the demand information; and sending the analyzed content information to the user terminal. According to the scheme, the AI intelligent agent technology and the NDC service are combined, an NDC interface tool is shielded for the downstream, and the adaptation difficulty of WEB and APP ends is reduced.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, system, electronic device and storage medium for processing air ticket sales business. Background Art

[0002] NDC (New Distribution Capability) was proposed by IATA (International Air Transport Association) in 2012 and has gradually matured and stabilized. Airlines, service providers or aggregators have provided stable and standard interfaces. However, NDC interface message nodes have many attributes, rich interfaces, large differences between versions, different business process interface calling methods, and interfaces cannot be directly used by end users.

[0003] At present, NDC air ticket sales to agents or individual users are usually implemented through the WEB or APP, which helps system users operate and complete the entire NDC order operation through general or customized page interactions, complex forms, and rich instruction documents.

[0004] However, the NDC interface version iteration is fast, there are many interfaces for upstream airlines, and the number of NDC interface protocol updates or expanded access codes has increased, which makes downstream access costs high and makes WEB and APP adaptation difficult from a technical implementation perspective. Under the premise of ensuring the stability and reliability of the system, as the NDC market expands, it is relatively difficult for WEB and APP to follow market demand and NDC interface iteration. Summary of the invention

[0005] To solve the above problems, the present invention provides a method and system for processing air ticket sales business. The AI ​​intelligent body technology is combined with the NDC business, which shields the NDC interface tool from the downstream and reduces the difficulty of adaptation on the WEB and APP ends.

[0006] In a first aspect, a method for processing air ticket sales business is provided, the method comprising:

[0007] Determine the user's demand information in the air ticket sales business;

[0008] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to generate the request message of the NDC interface;

[0009] Use the request message generated by the NDC interface to call the NDC interface tool and obtain the NDC interface response message;

[0010] Based on the NDC interface response message and prompt word template, combined with the NDC knowledge base tool, LLM is called to parse out the content information that matches the required information;

[0011] The parsed content information is sent to the user terminal.

[0012] Furthermore, based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to generate the request message of the NDC interface, including:

[0013] Determine the relevant information that matches the required information;

[0014] Based on the requirement information and prompt word template, combined with the NDC knowledge base tool, call LLM to summarize the NDC interfaces that need to be used;

[0015] Use the required NDC interface to call the NDC knowledge base tool to obtain the NDC interface protocol content and samples;

[0016] Using the prompt word template, with the obtained NDC interface protocol content and samples as the background, and the determined associated information as the condition, combined with the NDC knowledge base tool, calling LLM, and generating the request message of the NDC interface, wherein the prompt word template used is the prompt word template that has a corresponding relationship with the summarized NDC interface that needs to be used.

[0017] Furthermore, based on the NDC interface response message and prompt word template, combined with the NDC knowledge base tool, LLM is called to parse out content information that matches the required information, including:

[0018] Using the prompt word template, with the NDC interface protocol content and samples as the background, and the NDC interface response content as the condition, combined with the NDC knowledge base tool, calling LLM, parses out the content information that matches the required information.

[0019] Furthermore, the ticket sales business is a flight query link; the demand information is itinerary information, and the itinerary information includes: departure place, arrival place and planned time information;

[0020] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including:

[0021] Based on the itinerary information and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interfaces to be used are AirShopping and OfferPrice;

[0022] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0023] Use AirShopping and OfferPrice to call the NDC knowledge base tool to obtain the AirShopping and OfferPrice interface protocol content, flight query samples, and best price search samples.

[0024] Furthermore, before determining the itinerary information of the user in the flight query phase, the method further includes:

[0025] Receive itinerary plan description information from a user terminal and obtain user environment information; the itinerary plan description information includes: events, event-related information and booking requirements, and the user environment information includes: login user name and login environment;

[0026] Use the prompt word template, combined with the itinerary description information and user environment information, to call LLM to summarize the user's itinerary information;

[0027] Send the summarized itinerary information to the user for confirmation;

[0028] Determine the user's itinerary information during the flight query phase, including:

[0029] The itinerary information confirmed by the user is determined as the itinerary information of the user in the flight query phase.

[0030] Furthermore, the associated information matching the demand information is determined, including:

[0031] Using the prompt word template, analyzing the user's implicit preference information based on the event and event-related information, adjusting the planned time, and obtaining the adjusted planned time;

[0032] According to the departure and arrival places in the itinerary information, historical order information is obtained, combined with the NDC knowledge base tool, and LLM is called to summarize the user's itinerary preferences;

[0033] Determine itinerary planning suggestions based on the adjusted planned time information, event information and prompt word template;

[0034] Using the prompt word template, combined with the NDC knowledge base tool, and calling LLM, the itinerary preferences and itinerary planning suggestions are summarized to obtain itinerary-related information that matches the itinerary information.

[0035] Further, parsing the content information matching the demand information into a flight result list;

[0036] Before sending the parsed content information to the user terminal, the method further includes:

[0037] Use the prompt word template to combine the user's itinerary preferences and the parsed flight combination list, use LLM to label each flight combination for recommendation and sorting, and recommend and sort according to the label order;

[0038] Sending the parsed content information to the user terminal specifically includes:

[0039] The flight result list recommended and sorted in the order of tags is sent to the user terminal.

[0040] Furthermore, the ticket sales business is a flight reservation link; the demand information is flight information and passenger information;

[0041] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, determine the associated information that matches the demand information, including:

[0042] Use the prompt word template, combine the flight list in short-term memory with the flight combination selected by the user and the passenger information, and parse the OfferId and passenger list of the flight combination;

[0043] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including:

[0044] Based on flight information and passenger information, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interface to be used is OrderCreate;

[0045] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0046] Use OrderCreate to call the NDC knowledge base tool to obtain the OrderCreate interface protocol content and create an order sample.

[0047] Furthermore, using the prompt word template, with the NDC interface protocol content and samples as the background, and the NDC interface response content as the condition, combined with the NDC knowledge base tool, calling LLM, parses out the content information that matches the required information, including:

[0048] Use the prompt word template, take the OrderCreate interface protocol content and the order creation sample as the background, and the OrderCreate interface response content as the condition, combine with the NDC knowledge base tool, call LLM, and parse the order number and specific order content.

[0049] Furthermore, all unpaid orders are tagged and the insurance and payment deadlines for each order are recorded;

[0050] Sort the current unpaid orders by the insurance time limit and payment time limit, and use LLM to identify the time that is about to expire based on the current time point and the sorted current unpaid orders;

[0051] Remind users to pay based on identified orders that are about to expire;

[0052] When an unpaid order exceeds the insurance time limit, call the NDC knowledge base tool to obtain the next action of the airline to which the unpaid order belongs after the payment time limit has expired;

[0053] Follow the next steps to process the unpaid order.

[0054] Furthermore, the air ticket sales business is a payment and ticket issuance process; the demand information is a ticket issuance demand;

[0055] Determine the relevant information that matches the required information, including:

[0056] Determine the payment order, payment method and payment voucher that match the user's ticketing needs;

[0057] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including:

[0058] Based on the ticketing requirement information and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interface to be used is OrderChange;

[0059] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0060] Use OrderChange to call the NDC knowledge base tool to obtain the OrderChange interface protocol content and payment and invoicing samples.

[0061] Furthermore, the ticket sales business is an additional service reservation link; the demand information is an additional service reservation demand including an additional service type;

[0062] Determine the relevant information that matches the required information, including:

[0063] Using the prompt word template, combined with the user's additional service reservation needs and environmental information, combined with the NDC knowledge base tool, calling LLM, disassemble the user's additional service reservation needs, and obtain related information that matches the additional service reservation needs;

[0064] Based on the additional service reservation requirements and prompt word template, combined with the NDC knowledge base tool, LLM is called to summarize the NDC interfaces that need to be used, including:

[0065] Based on the additional service reservation requirements and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interfaces that need to be used are ServiceList and OrderChange;

[0066] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0067] Use ServiceList and OrderChange to call the NDC knowledge base tool to obtain the ServiceList and OrderChange interface protocol content, additional service query samples, and additional service samples.

[0068] Furthermore, the additional service reservation requirement also includes additional service preference information. Before sending the parsed content information to the user terminal, the method further includes:

[0069] Query the additional service types purchased by the current user through the order management module;

[0070] Query the temperature and humidity of the departure place through the external interface;

[0071] Call the NDC interface to obtain a list of available additional services;

[0072] Based on the list of available additional services, the temperature and humidity of the departure point, and the additional service preference information, combined with the NDC knowledge base tool, the LLM is called to summarize, mark, and sort the additional service list according to the additional service preference;

[0073] Send the parsed content information to the user terminal, including:

[0074] The additional service list marked and sorted according to the additional service preference is sent to the user terminal.

[0075] Furthermore, the ticket sales business is an additional service payment link; the demand information is an additional service payment demand;

[0076] Determine the relevant information that matches the required information, including:

[0077] Determine the payment additional service items and current orders that match the additional service payment requirements;

[0078] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including:

[0079] Based on the additional service payment requirements and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interface to be used is OrderChange;

[0080] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0081] Use OrderChange to call the NDC knowledge base tool to obtain the OrderChange interface protocol content and purchase additional service samples.

[0082] Furthermore, the air ticket sales business is a process of proposing to modify the itinerary; the demand information is a demand for proposing to modify the itinerary;

[0083] Determine the relevant information that matches the required information, including:

[0084] Using the prompt word template, combined with the user's request to modify the itinerary and environmental information, combined with the NDC knowledge base tool, calling LLM, disassemble the user's original itinerary and target itinerary;

[0085] Use the prompt word template, based on the user's original itinerary, combined with the NDC knowledge base tool, call LLM, generate order query parameters, and call the order management function to query the user's changed orders;

[0086] The user's changed order, original itinerary and target itinerary are used as related information matching the itinerary modification request;

[0087] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including:

[0088] Based on the proposed itinerary modification requirements and the prompt word template, combined with the NDC knowledge base tool, LLM is called and it is concluded that the NDC interface to be used is OrderReshop;

[0089] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0090] Call the NDC knowledge base tool to obtain the OrderReshop interface protocol content and ticket change query sample.

[0091] Furthermore, the ticket sales business is a process of confirming and modifying the itinerary; the demand information is a process of confirming and modifying the itinerary;

[0092] Determine the relevant information that matches the required information, including:

[0093] Determine the user's rebooking order that matches the confirmed itinerary modification needs and confirm the rebooking flight combination;

[0094] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including:

[0095] Based on the confirmation of the itinerary modification requirements and the prompt word template, the NDC interface required to call the LLM summary is OrderChange;

[0096] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0097] Use OrderChange to call the NDC knowledge base tool to obtain the OrderChange interface protocol content and payment change samples.

[0098] Furthermore, the air ticket sales business is a process of proposing to cancel a trip;

[0099] Before determining that the user has requested to cancel the trip, the method further includes:

[0100] Using the prompt word template, combined with the user-entered cancellation order description and environmental information, combined with the NDC knowledge base tool, calling LLM, disassembled the key order information and the user's needs to cancel the order;

[0101] Use the prompt word template to call LLM to generate query parameters based on the key information of the order;

[0102] Use the query parameters to call the order management function and query the current valid order list that matches the query parameters;

[0103] Use the prompt word template, combined with the NDC knowledge base tool, to call LLM and summarize the refund and change rules information of the queried order list;

[0104] Return the order that needs to be canceled and the refund and change rules information to the user terminal for user confirmation.

[0105] Furthermore, the demand information is a request to cancel a trip;

[0106] Determine the relevant information that matches the required information, including:

[0107] Determine the refund information that matches the cancellation request;

[0108] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to summarize the NDC interfaces that need to be used, including:

[0109] Based on the request to cancel the trip and the prompt word template, the large language model LLM is called to summarize the OrderReshop when the NDC interface needs to be used;

[0110] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0111] Use OrderReshop to call the NDC knowledge base tool to obtain the OrderReshop interface protocol content and refund query samples.

[0112] Furthermore, the air ticket sales business is a process of confirming and canceling a trip; the demand information is a demand for confirming and canceling a trip;

[0113] Determine the relevant information that matches the required information, including:

[0114] Identify the orders that match the cancellation requirements;

[0115] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to summarize the NDC interfaces that need to be used, including:

[0116] Using the cancellation request and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interface to be used is OrderChange;

[0117] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0118] Call the NDC knowledge base tool to obtain the OrderChange interface protocol content and refund sample.

[0119] Furthermore, the ticket sales business is a flight change reminder link;

[0120] Before determining the user's demand information in the air ticket sales business, the method further includes:

[0121] Receive flight change notifications from airlines;

[0122] Call the NDC knowledge base tool to obtain the OrderChangeNotif interface protocol content and flight change notification sample;

[0123] Using the prompt word template, with the OrderChangeNotif interface protocol content and flight change notification sample as the background, combined with the NDC knowledge base tool, call LLM to summarize the airline's OrderChangeNotif interface request message content, including the order number, change type and change reason;

[0124] A flight change message is sent to the user terminal, wherein the flight change message includes the order number, change type and change reason.

[0125] Furthermore, the demand information is to accept flight rescheduling;

[0126] Determine the relevant information that matches the required information, including:

[0127] Identify orders and flight change notifications that match accepted flight changes;

[0128] Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to summarize the NDC interfaces that need to be used, including:

[0129] The NDC interface that needs to be used to call the LLM summary to accept flight changes is OrderRetrieve;

[0130] Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including:

[0131] Use OrderRetrieve to call the NDC knowledge base tool to obtain the OrderRetrieve interface protocol content and accept flight change samples.

[0132] Furthermore, the method further comprises:

[0133] When there is an error or warning message, obtain the current NDC interface protocol content and sample containing the error or warning message;

[0134] Use prompt word templates, obtain the current NDC interface protocol content and samples as the background, and the NDC interface response content as the condition. Combined with the NDC knowledge base tool, call LLM to parse the error code and error prompt content.

[0135] Call the NDC knowledge base tool to obtain the cause and solution of the error code and error message content;

[0136] Call the log tracking module function in the flight query module to obtain the NDC interface request message and return content of the operation process before the error;

[0137] Use prompt word templates, the causes and solutions of error codes, and the request messages and return contents in the operation process as the background, combined with the NDC knowledge base tool, call LLM, and summarize the real error causes and solutions of the current operation process;

[0138] If the solution needs to call the NDC interface for processing, the interface call will be automatically completed and the processing result will be returned. Otherwise, the actual error cause and solution will be returned to the user terminal.

[0139] In a second aspect, a ticket sales business processing system is provided, comprising: an agent auxiliary module, an order management module, and a flight query module having an NDC interface tool;

[0140] The intelligent agent auxiliary module is used to determine the user's demand information in the air ticket sales business; based on the demand information and the prompt word template, combined with the NDC knowledge base tool, call the large language model LLM to generate an NDC interface request message; use the generated NDC interface request message to call the NDC interface tool to obtain the NDC interface response message; based on the NDC interface response message and the prompt word template, combined with the NDC knowledge base tool, call the LLM to parse out the content information that matches the demand information; and send the parsed content information to the user terminal.

[0141] In a third aspect, an electronic device is provided, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0142] Memory, used to store computer programs;

[0143] The processor is used to implement the steps of the above method when executing the program stored in the memory.

[0144] In a fourth aspect, a computer storage medium is provided, wherein the computer storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0145] Compared with the prior art, the present invention has the following advantages:

[0146] 1. Intelligent agent technology is used to convert the interaction between the WEB system or APP and the NDC interface tool into the interaction between the WEB or APP and the intelligent agent auxiliary module, and the interaction between the intelligent agent auxiliary module and the NDC interface tool. Through the intelligent agent auxiliary module, the downstream access is shielded, the complexity of the WEB system and APP is reduced, and there is no need to adapt to the personalized business processes of multiple airlines, which reduces the difficulty of adaptation on the WEB and APP sides.

[0147] 2. Use natural language, combined with end-user usage habits and NDC interface response content, to recommend content and automatically process processes, especially for complex business scenarios such as special passenger types and preferential price inquiries;

[0148] 3. Add order status and flight change notifications to your own environment, respond to changes in the environment and provide various reminders, thereby improving passengers' travel efficiency and reducing operation and maintenance work;

[0149] 4. For abnormal processes, we can learn the business problems of the aggregator and the airline in the NDC business, combine the log tracking system and the NDC interface standard document as well as the previous manual work order email processing problems to learn and handle abnormal processes, and reduce manual operation and maintenance costs.

[0150] Other features and advantages of the present disclosure will be described in the following description, and partly become apparent from the description, or be understood by implementing the present disclosure. The purpose and other advantages of the present disclosure can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0151] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0152] Figure 1 The overall business process diagram of current NDC air ticket sales is shown;

[0153] Figure 2 A schematic diagram of an overall business process of NDC air ticket sales according to an embodiment of the present disclosure is shown;

[0154] Figure 3 A schematic diagram of an agent-assisted NDC flight query according to an embodiment of the present disclosure is shown;

[0155] Figure 4 A schematic diagram of ticketing based on agent-assisted NDC reservation according to an embodiment of the present disclosure is shown;

[0156] Figure 5 A schematic diagram of an agent-assisted NDC additional service according to an embodiment of the present disclosure is shown;

[0157] Figure 6 A schematic diagram of an additional service based on agent-assisted NDC rebooking according to an embodiment of the present disclosure is shown;

[0158] Figure 7 A schematic diagram of canceling a trip based on agent-assisted NDC according to an embodiment of the present disclosure is shown;

[0159] Figure 8 A schematic diagram of an agent-assisted NDC flight change notification according to an embodiment of the present disclosure is shown;

[0160] Fig. 9 A schematic diagram of a functional module diagram of a ticket sales business system according to an embodiment of the present disclosure is shown;

[0161] Fig.10 A flowchart of a method for processing air ticket sales business according to an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0162] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0163] The core of the technical solution disclosed in the present invention is to combine AI intelligent agent technology with NDC business, and realize the full process management of natural language NDC intelligent reservation through LLM fine-tuning and prompt word engineering, which includes NDC intelligent agent model design, deep integration of intelligent agent and NDC business process, environmental response of intelligent agent based on NDC order status, long-term and short-term memory processing of intelligent agent about NDC business and application of prompt word engineering in NDC business.

[0164] In order to clearly illustrate the solution of the embodiment of the present disclosure, the overall business process of the technical solution of the present disclosure is first described below, followed by a description of the system modules, and finally, in combination with an air ticket sales case, the air ticket sales business processing method of the present disclosure is described.

[0165] 1. Overall business process

[0166] NDC booking can be divided into AirShopping (flight inquiry, price calendar), OfferPrice (price verification, re-price verification, Up Sell), ServiceList / SeatAvail (additional service inquiry), and Oder Create (order creation) before ordering; and can be divided into OrderChange (order payment, change confirmation, order separation), OrderRetrieve (order view), Order Cancel (order cancellation), OrderReshop (change inquiry, change price verification, refund and change fee inquiry), and ServiceList / SeatAvail (additional service inquiry) after ordering.

[0167] Based on the intelligent agent to assist NDC booking full business process Figure 2 As shown in the figure, the original interaction process is simplified. The overall natural language interaction reduces complex business logic and personalized business needs, receives end-user needs based on LLM (Large Language Model) capabilities, and calls NDC interface tool capabilities to complete reservations and subsequent processes, and meet end-user preferences and personalized needs.

[0168] The six key business processes are as follows:

[0169] 1.1 Flight Query

[0170] like Figure 3 , the end user needs to clarify the itinerary plan, including at least the departure place, arrival place and planned time. If there is a discount code such as a key customer code, promotion code, or tripartite agreement code, it can be described together or supplemented after the prompt. The agent uses the above information to generate an NDC interface request message using the LLM and prompt word template and calls the NDC interface to obtain flight information. It also recommends and sorts flights based on fare preferences, price calendars, brand fares, and preferential price query information to help end users reduce redundant information. Of course, it can also expand to view all flight data.

[0171] 1.2 Payment and Ticketing

[0172] like Figure 4 After the terminal user confirms the itinerary, the system automatically obtains the frequent flyer information that has been filled in and submits it to the terminal user to confirm the passenger and itinerary information together. After confirmation, the order reservation is completed. The intelligent experience will parse its key information based on the order information returned by the NDC interface, including the order item content, refund and change rules, insurance time limit, payment time limit, and warning information. If there is any inconsistency with the itinerary confirmation, a prompt will be given. At the same time, the terminal user will be prompted to complete the ticketing operation within the insurance time limit and payment time limit. After the ticket is issued, the passenger information can also be modified according to the rules of different airlines. The intelligent experience will give prompts according to the airline rules, and the terminal customer does not need to handle it offline.

[0173] 1.3 Purchase additional services

[0174] like Figure 5 , the end user describes the type of additional services they want to purchase and their preferences, and the intelligent experience calls the SeatAvail or ServiceList interface of NDC, and filters and sorts the additional services according to their preferences. The purchase of additional services is completed after the end user confirms. When purchasing multiple additional services or the end user has purchased additional services in the past, the intelligent experience summarizes and recommends additional services. At the same time, you can view the current free additional services or the additional services you have purchased through a simple description and take the next step, such as canceling or viewing details.

[0175] 1.4 Modify itinerary

[0176] like Figure 6 As the most complex business in the air ticket business, the intelligent agent first trains all the refund and change rules of all connected airlines and previous change data and results. When the end user selects and modifies the itinerary, the intelligent agent will check the refund and change rule information returned by the NDC interface, and prompt if the next step cannot be performed. For some people who change their tickets, the intelligent agent will automatically separate their orders and record the relationship between the two orders for subsequent end-user inquiries. Then, based on the previous refund and change data, the change plan queried through the OrderReshop interface of NDC is recommended and sorted, and finally the end user confirms the change itinerary. If there is an abnormality in the change process or results, no human intervention is required. The intelligent agent can answer the end user one by one and recommend other solutions based on the refund and change rules, airline business and previous summary of reasons for failed changes.

[0177] 1.5 Cancellation of Trip

[0178] like Figure 7 When the terminal user chooses to cancel the trip, the agent will prompt the next step of canceling the trip according to the refund and change rules message in the NDC interface, the OrderReshop interface content and the business logic of each airline. Usually, the terminal user prefers to cancel the ticket without any handling fee, but due to time, rebooking, airline policies and other factors, they can only cancel the ticket by refunding or retaining the ticket. The agent can answer the terminal user and feedback the specific refund fee or fine according to the refund and change rules and the business logic of the airline.

[0179] like Figure 8After receiving the airline's OCN message, the agent will extract the OCN interface message content, obtain the order change status, and extract the current order content to notify the end user. The next step is based on the user's choice: accept the change, the agent will obtain the updated order content and return the difference before and after the order change to the end user; involuntary change, the end user does not accept the itinerary change, the agent will query the terminal user for the flight information that can be changed and recommend it according to the involuntary change business rules. If the change process or result is abnormal and the itinerary is modified, the agent can answer and handle the problem for the end user; and cancel the order.

[0180] 2. System Module

[0181] NDC booking based on agent assistance can be divided into modules: flight query module, including NDC interface, which is provided by airlines, service providers and Aggregators.

[0182] The NDC interface has a log tracking module, which has log cleaning and log collection functions. The airline has a flight change notification function. The Aggregator has flight query, additional service query, seat query, in-flight service and brand fare query functions.

[0183] The order management module holds the business order status management functions, including creating orders, order payments, modifying orders, canceling orders, and sales reports.

[0184] Add intelligent agent auxiliary modules, including LLM large model, prompt word management, booking assistant, itinerary butler, and intelligent customer service. Fig. 9 shown.

[0185] The booking assistant agent is responsible for natural language interactive booking and order operations, including automatic processing of booking procedures, flight recommendations, fare preferences, brand fare analysis, additional service recommendations, and itinerary planning.

[0186] The trip manager agent is responsible for reminding and processing order status changes, including order status management, travel reminders, and flight change notifications;

[0187] The intelligent customer service agent is responsible for business answers and exception handling, including the NDC knowledge base and error codes. The NDC knowledge base is the NDC knowledge base tool. The NDC interface is the NDC interface tool.

[0188] like Fig.10 As shown, the present disclosure provides a flowchart of a method for processing a ticket sales business, and the ticket sales business processing method includes the following steps:

[0189] Step 101: Determine the user's demand information in the air ticket sales business;

[0190] Step 102: Based on the demand information and the prompt word template, in combination with the NDC knowledge base tool, the large language model LLM is called to generate a request message for the NDC interface;

[0191] Specifically, this step 102 includes:

[0192] Determine the relevant information that matches the required information;

[0193] Based on the requirement information and prompt word template, combined with the NDC knowledge base tool, call LLM to summarize the NDC interfaces that need to be used;

[0194] Use the required NDC interface to call the NDC knowledge base tool to obtain the NDC interface protocol content and samples;

[0195] Using the prompt word template, with the obtained NDC interface protocol content and samples as the background, and the determined associated information as the condition, combined with the NDC knowledge base tool, calling LLM, and generating the request message of the NDC interface, wherein the prompt word template used is the prompt word template that has a corresponding relationship with the summarized NDC interface that needs to be used.

[0196] Step 103: Use the request message for generating the NDC interface to call the NDC interface tool and obtain the NDC interface response message;

[0197] Step 104: Based on the NDC interface response message and the prompt word template, in combination with the NDC knowledge base tool, the LLM is called to parse out the content information that matches the requirement information;

[0198] Specifically, this step 104 includes: using a prompt word template, taking the NDC interface protocol content and samples as the background, the NDC interface response content as the condition, combining the NDC knowledge base tool, calling LLM, and parsing out content information that matches the demand information.

[0199] Step 105: Send the parsed content information to the user terminal.

[0200] In a specific implementation, demand information and user behavior habits can be combined to complete the ticket sales business through the collaboration of one or more intelligent agents. The functional division and collaborative use of each intelligent agent are not limited to the intelligent agents in the embodiment of the present disclosure.

[0201] The specific implementation of the above solution will be illustrated through the following specific sales case.

[0202] 3. Air ticket sales business processing method based on specific cases:

[0203] 3.1 Flight inquiries can be divided into new user inquiries and old user inquiries

[0204] The user only needs to describe his / her travel plan. For example, the user enters "I have a meeting in Shanghai on the evening of May 4th, please help me book a flight ticket for that day".

[0205] To reserve an assistant agent, first disassemble the agent according to the pre-training steps and execute:

[0206] Obtaining environmental information, including login user name, login environment, the login user's existing order information, and used frequent flyer information;

[0207] Using the prompt word template, combined with the user input information and environmental information, LLM is called to extract the user's departure place as Beijing, the location of the login IP address, the arrival place as Shanghai, and the departure time as May 4;

[0208] Use the prompt word template to analyze the user's implicit preference information. According to the conditions, there is a meeting in the evening, and the flight departure time + flight duration needs to be earlier than 6 pm. Considering flight delays and traffic conditions, it is necessary to arrive before 3 pm.

[0209] The trip manager agent performs the following operations based on the conditions analyzed by the booking assistant agent and the pre-trained steps:

[0210] Call the order management module function to obtain the user's order information for flights from Beijing to Shanghai;

[0211] Call LLM to summarize the user's itinerary preferences, which include: brand fare preference is free refund and change and in-flight WIFI, aircraft type preference is large passenger aircraft, and seat preference is close to the entrance and exit;

[0212] The itinerary has specific travel time requirements, so there is no need to call the price calendar;

[0213] Based on the user's itinerary preference information summarized in the booking assistant, the meeting on the same day and the arrival before 3 o'clock, LLM is called to form the next itinerary planning suggestion: whether it is necessary to book the return information on the evening of May 4 or the next day.

[0214] The Booking Assistant Agent receives further user itinerary preference information and itinerary planning content from the Itinerary Manager Agent and executes:

[0215] Use the prompt word template to summarize the itinerary: Itinerary 1, Beijing to Shanghai on May 4; Itinerary 2, Beijing to Shanghai and back on May 4; Itinerary 3, Beijing to Shanghai on May 4, Shanghai to Beijing and back on May 5;

[0216] The NDC interfaces required to call LLM summary are AirShopping and OfferPrice;

[0217] Call the NDC knowledge base tool to obtain the AirShopping and OfferPrice interface protocol content and flight query samples and best price search samples;

[0218] Use the prompt word template, take the NDC interface protocol and sample as the background, and trip 1-3 as the conditions to call LLM to generate the NDC interface request message;

[0219] Use the generated NDC interface request message to call the NDC interface tool to obtain the response messages of AirShopping and OfferPrice respectively;

[0220] Using the prompt word template, with the NDC interface protocol and samples as the background and the NDC interface response content as the condition, call LLM to parse the flight combination list of the three itineraries and the brand fare content that can be selected under each flight combination list;

[0221] Using the prompt word template, the user's preference content (i.e., user itinerary preference and implicit preference) and the parsed flight result list are combined, and LLM is used to label each flight combination for recommendation and sorting.

[0222] Return to the interactive page and tell the user that the flights from Beijing to Shanghai on May 4 are as follows (arranged in order of labels) and whether they need to purchase round-trip flights. When the user confirms that they need round-trip flights from Beijing to Shanghai on May 4 and from Shanghai to Beijing on May 5, the recommended flight list for itinerary 3 is displayed. After the user confirms the flight information, the user is prompted to select whether the passenger is the current user.

[0223] After the user confirms the passenger information and the reservation, the reservation assistant agent will perform the following operations:

[0224] Use the prompt word template, combine the flight list in short-term memory with the flight combination selected by the user and the passenger information, and parse the OfferId and passenger list of the flight combination;

[0225] Among them, OfferId is a concept in NDC. NDC ticket sales use Offer as the sales unit, and OfferId is the unique identifier of the sales unit.

[0226] The NDC interface required to call LLM summary is OrderCreate;

[0227] Call the NDC knowledge base tool to obtain the OrderCreate interface protocol content and order creation sample;

[0228] Use the prompt word template, NDC interface protocol and sample as the background, OfferId and passenger list as conditions to call LLM to generate NDC interface request message;

[0229] Use the generated NDC interface request message to call the NDC interface tool and obtain the OrderCreate response message;

[0230] Use the prompt word template, NDC interface protocol and sample as the background, NDC interface response content as the condition, and call LLM to parse the order number and specific order content;

[0231] Use prompt word templates to summarize order content into a standard order management structure;

[0232] Call the order management module to store order content;

[0233] Returns the order number and order content.

[0234] 3.2 Payment and Ticketing

[0235] Users only need to select the order and payment method to complete the order payment. In the process, the trip manager will manage and remind you of delayed payment through environmental changes. For example, if the user has completed the round-trip reservation from Beijing to Shanghai on May 4 and from Shanghai to Beijing on May 5, the insurance period of the order is 12:00 on April 15, and the payment period of the order is 22:00 on April 20.

[0236] The Trip Manager agent will label all unpaid orders and record the insurance time limit and payment time limit for each order. Every half an hour, the Trip Manager will sort the current unpaid orders by insurance time limit and payment time limit, and use the LLM capability to identify the time that is about to expire based on the current time point and the sorted orders. If the current time is 14:00 on April 14, the Trip Manager agent will inform the user in the form of a message notification that "the order price will change at 12:00 on April 15, please pay in time." If the current time is 9:00 on April 20, the insurance time limit has expired, and the Trip Manager will execute:

[0237] Call the NDC knowledge base tool to obtain the next step of the airline after the payment deadline has expired, which is to re-check the price through the OrderReshop interface to obtain the new price, and obtain the message protocol and re-check price sample;

[0238] Use the prompt word template to call LLM through the order number to generate the request message of the NDC interface;

[0239] Use the generated NDC interface request message to call the NDC interface tool to obtain the response message of OrderReshop;

[0240] Use the prompt word template, take the NDC interface protocol and sample as the background, and the NDC interface response content as the condition to call LLM to parse the current price to be paid for the order;

[0241] Inform the user via message notification that "the order price will be automatically cancelled at 22:00 on April 20, please pay in time. The current amount to be paid is 10,000."

[0242] After the user confirms the payment method and pays, the Booking Assistant Agent executes:

[0243] The NDC interface that needs to be used to call LLM summary is OrderChange;

[0244] Call the NDC knowledge base tool to obtain the OrderChange interface protocol content and payment and invoicing samples;

[0245] Use the prompt word template, NDC interface protocol and sample as the background, payment order, payment method and payment credential as conditions to call LLM to generate NDC interface request message;

[0246] Use the generated NDC interface request message to call the NDC interface tool to obtain the OrderChange response message;

[0247] Use the prompt word template, NDC interface protocol and sample as the background, NDC interface response content as the condition, and call LLM to parse the payment result;

[0248] Returns payment result information, usually the ticket number.

[0249] Here, after payment, the user can also describe and modify the incorrect passenger information.

[0250] 3.3 Purchase additional services

[0251] When the user needs to purchase additional services, he / she can confirm the order content by checking the order and describe the additional service needs, such as: "Cancel the paid baggage allowance from Beijing to Shanghai, and purchase a meal."

[0252] 3.3.1 Booking Assistant Execution:

[0253] Using the prompt word template, combined with the user input information and environmental information, the LLM is called to break down the user's needs into two: one is to cancel the current baggage supplement service on the Beijing-Shanghai segment, and the other is to purchase a meal on the Beijing-Shanghai segment;

[0254] The NDC interfaces required to call LLM summary are ServiceList and OrderChange;

[0255] Call the NDC knowledge base tool to obtain the ServiceList and OrderChange interface protocol content and additional service query samples and additional service cancellation samples;

[0256] Use the prompt word template, take the NDC interface protocol and sample as the background, cancel the baggage service from Beijing to Shanghai, and purchase meals from Beijing to Shanghai as the condition, and call LLM to generate ServiceList and OrderChange interface request messages;

[0257] Use the generated NDC interface request message to call the NDC interface tool to obtain the response messages of ServiceList and OrderChange respectively;

[0258] Using the prompt word template, with the NDC interface protocol and sample as the background and the NDC interface response content as the condition, LLM is called to parse the result of canceling the baggage additional service and the query result of the meal additional service; among them, the query result of the meal additional service is also the list of additional services to be purchased.

[0259] Call the order management module to update the service information in the order.

[0260] In the scheme of the embodiment of the present disclosure, there is a corresponding relationship between the prompt word template and the NDC interface. After identifying the NDC interface required for the user sales stage, the NDC interface will be called in sequence to complete a series of business processes. In this process, each different NDC interface corresponds to a corresponding prompt word template.

[0261] 3.3.2 Execution of the trip manager agent:

[0262] Query the types of meals purchased by the current user through the order management module;

[0263] Query the temperature and humidity of the departure place through the external interface;

[0264] Based on the list of purchasable additional services, environment, and preferences obtained by the booking assistant in 3.3.1, call the LLM summary and mark and sort the additional service list according to possible preferences, where the additional service list includes meals, luggage, and seats;

[0265] Return to the front-end user. The user finally confirms the type of meal purchased, and the booking assistant agent finally executes:

[0266] The NDC interface that needs to be used to call LLM summary is OrderChange;

[0267] Call the NDC knowledge base tool to obtain the OrderChange interface protocol content and purchase additional service samples;

[0268] Using the prompt word template, with the NDC interface protocol and sample as the background, and the user confirming the purchased additional service options and the current order as conditions, LLM is called to generate an OrderChange interface request message;

[0269] Use the generated NDC interface request message to call the NDC interface tool to obtain the OrderChange response message;

[0270] Using the prompt word template, with the NDC interface protocol and sample as the background and the NDC interface response content as the condition, call LLM to parse the purchase result of meal additional services;

[0271] Call the order management function to add additional service information in the order;

[0272] Returns the front-end user's purchase results and the status of the current order.

[0273] 3.4 Modify itinerary

[0274] 3.4.1 The user needs to modify the return flight and enter "Change the flight from Shanghai to Beijing on May 5 to May 6".

[0275] Booking Assistant Agent Execution:

[0276] Using the prompt word template, combined with the user input information and environmental information, LLM is called to decompose the user's need to modify the itinerary. The modification content is that the original departure time is May 5, the departure place is Shanghai, and the arrival place is Beijing; the departure time after modification is May 6, the departure place is Shanghai, and the arrival place is Beijing;

[0277] Use the prompt word template to call LLM to generate query parameters, call the order management function to query the current orders, and filter out the orders with a departure time of May 6, a departure point of Shanghai, and a destination of Beijing, as well as their itinerary lists;

[0278] The NDC interface required to call LLM summary is OrderReshop;

[0279] Call the NDC knowledge base tool to obtain the OrderReshop interface protocol content and ticket change query sample;

[0280] Use the prompt word template, NDC interface protocol and sample as the background, the user's changed order, original itinerary, and target itinerary as conditions, and call LLM to generate the OrderReshop interface request message;

[0281] Use the generated NDC interface request message to call the NDC interface tool to obtain the response message of OrderReshop;

[0282] Use the prompt word template to call LLM to summarize the flight list, and filter, mark and sort the flights and their preferred contents with the flight query before booking.

[0283] 3.4.2 Return the list of flights recommended by the front-end user, and the user confirms the flight combination to be changed, which is then executed by the booking assistant agent:

[0284] The NDC interface that needs to be used to call LLM summary is OrderChange;

[0285] Call the NDC knowledge base tool to obtain the OrderChange interface protocol content and payment change sample;

[0286] Using the prompt word template, with the NDC interface protocol and sample as the background, and the user's changed order and confirmed flight combination as conditions, LLM is called to generate an OrderChange interface request message;

[0287] Use the generated NDC interface request message to call the NDC interface tool to obtain the OrderChange response message;

[0288] Use the prompt word template to call LLM to analyze the change result and the order content after the change;

[0289] Call the order management function to record the ticket change process and results.

[0290] 3.4.3 Return the ticket change status and order content after the change to the front-end user.

[0291] 3.5 Cancellation of Trip

[0292] 3.5.1 Due to the cancellation of the meeting, the user needs to cancel the order and enter "Help me cancel the flight from Beijing to Shanghai on May 4th".

[0293] Booking Assistant Agent Execution:

[0294] Using the prompt word template, combined with the user input information and environmental information, LLM is called to deconstruct the user's demand to cancel the order. The order content is: departure time May 4, departure place Beijing, arrival place Shanghai;

[0295] Use the prompt word template to call LLM to generate query parameters, and call the order management function to query the currently valid orders, and the order list with the departure time of May 4, the departure place of Beijing and the arrival place of Shanghai;

[0296] Use the prompt word template to call LLM to summarize the refund and change rules of the queried order list.

[0297] 3.5.2 Return the front-end user to confirm the order cancellation. The trip manager agent receives the request to cancel the order and the order in the environment, and queries the detailed refund information for the order. The trip manager agent executes:

[0298] The NDC interface required to call LLM summary is OrderReshop;

[0299] Call the NDC knowledge base tool to obtain the OrderReshop interface protocol content and refund query sample;

[0300] Use the prompt word template, NDC interface protocol and sample as the background, order as the condition, and call LLM to generate OrderReshop interface request message;

[0301] Use the generated NDC interface request message to call the NDC interface tool to obtain the response message of OrderReshop;

[0302] Use the prompt word template, the NDC interface protocol and sample as the background, and the NDC interface response content as the condition to call LLM to parse the order refund fee and penalty.

[0303] 3.5.3 Return to the front-end user and remind the user of the specific fees and amounts for canceling the order. After the user confirms, the booking assistant agent executes:

[0304] The NDC interface that needs to be used to call LLM summary is OrderChange;

[0305] Call the NDC knowledge base tool to obtain the OrderChange interface protocol content and refund samples;

[0306] Use the prompt word template, NDC interface protocol and sample as the background, order as the condition, and call LLM to generate OrderChange interface request message;

[0307] Use the generated NDC interface request message to call the NDC interface tool to obtain the OrderChange response message;

[0308] Use the prompt word template, NDC interface protocol and sample as the background, NDC interface response content as the condition, and call LLM to parse the order cancellation result or status.

[0309] 3.5.4 Return the final order processing results and status to the front-end user.

[0310] 3.6 Flight Change Reminder

[0311] 3.6.1 After the airline sends a flight change notification, the itinerary manager agent responds to the environmental changes and executes:

[0312] Call the NDC knowledge base tool to obtain the OrderChangeNotif interface protocol content and flight change notification sample;

[0313] Use the prompt word template, with the NDC interface protocol and sample as the background, to call LLM to summarize the airline's OrderChangeNotif interface request content, including the order number, change type and change reason, and may include the change content;

[0314] Notify front-end users of flight changes and the details of the changes in the form of messages.

[0315] 3.6.2 The front-end user confirms the change of flight segments and the subsequent operations of the flight change. When the user chooses to accept the flight change (the process of not accepting the involuntary change is similar to the process of modifying the itinerary, but there is no handling fee; the process of not accepting the cancellation of the order is similar to the process of cancelling the order, but there is no refund fee and penalty), the booking assistant agent executes:

[0316] The NDC interface that needs to be used to call the LLM summary to accept flight changes is OrderRetrieve;

[0317] Call the NDC knowledge base tool to obtain the OrderRetrieve interface protocol content and the flight change acceptance sample;

[0318] Use the prompt word template, NDC interface protocol and sample as the background, order and flight change notification as conditions, and call LLM to generate OrderRetrieve interface request message;

[0319] Use the generated NDC interface request message to call the NDC interface tool and obtain the OrderRetrieve response message;

[0320] Use the prompt word template, NDC interface protocol and sample as the background, NDC interface response content as the condition, and call LLM to parse the changed order content;

[0321] Call the order management module, query the original order information, and update the order content;

[0322] Use the prompt word template to call LLM to summarize the differences between the original order information and the updated order information, and give the user a change prompt.

[0323] 3.7 Exception Handling

[0324] If there are errors or warnings, the Smart Customer Service Agent performs:

[0325] Get the current interface content through environmental information;

[0326] Use the prompt word template, NDC interface protocol and sample as the background, NDC interface response content as the condition, and call LLM to parse the error code and error prompt content;

[0327] Call the NDC knowledge base tool to obtain the cause and solution of the error code and error message content;

[0328] Call the log tracking module function in the flight query module to obtain the NDC interface request message and return content of the operation process before the error;

[0329] Use the prompt word template, the cause and solution of the error code, and the request message and return content in the operation process as the background, and call LLM to summarize the real error cause and solution of the current process;

[0330] If the solution needs to call the NDC interface for processing, the interface call will be automatically completed and the processing result will be returned. Otherwise, the actual error cause and solution will be returned to the user terminal.

[0331] The disclosed embodiments solve the following seven problems in current NDC ticket sales in terms of technical implementation, business process, and sales operations:

[0332] 1. NDC interface versions are iterating fast, there are many interfaces for upstream airlines, many NDC interface protocol updates or extension access codes, high downstream access costs, and high difficulty in WEB and APP adaptation. Under the premise of ensuring the stability and reliability of the system, as the NDC market expands, it is relatively difficult to keep up with market demand and the iterative development of NDC interfaces.

[0333] 2. To be compatible with different business scenarios and user groups for booking NDC air tickets, different WEB or APP interaction processes and pages need to be customized, which requires additional development and operation and maintenance costs.

[0334] 3. The NDC interface returns a rich selection of flights, but lacks sorting and recommendations based on the different preferences of end users, making the content complex and difficult to choose.

[0335] 4. The NDC interface ticket sales business process is different for different market airlines, especially in the additional service options and refund and change business processes. It is not user-friendly for end users and has a high learning cost.

[0336] 5. The payment deadline and insurance deadline for delayed payment in the NDC interface require additional message systems to notify the end user or lack of message notifications. Flight change information also requires additional message systems to notify the end user, otherwise it will affect the check-in of passengers.

[0337] 6. The NDC interface usually returns error descriptions or codes in text form for abnormal information, which will bring huge operation and maintenance pressure. For end customers, first, they are not familiar with the business process, and second, they do not understand these abnormal descriptions or codes. At this stage, they must contact the platform provider offline and then contact the airline. In addition, the processing process is long, which will affect the travel plans of passengers in actual business scenarios.

[0338] 7. The NDC interface ticket sales business process is complex and changeable, requiring frequent business training for agents and regular updates of user manuals and other relevant business documents for individual users.

[0339] Regarding technical issue 1: In the disclosed embodiment, a large language model and the functions of three intelligent agents are used to call the NDC interface tool, which shields downstream access. The WEB and APP ends do not need to connect to the NDC interfaces of different airlines, reducing the difficulty of adaptation on the WEB and APP ends.

[0340] Regarding technical problem 2: In the embodiment of the present disclosure, a large language model and the functions of three agents are used to return corresponding results to the user based on the user's demand description information and display them on the page, so that different business scenarios and user groups can use a unified WEB or APP interaction process and page;

[0341] In response to technical problem 3, in the embodiment of the present disclosure, a large language model, environmental information and the user's historical order information are used to obtain the user's preference information, and the returned flight options are recommended and sorted based on the preference information. The content is organized to make it easy for users to choose.

[0342] Regarding technical issue 4: In the disclosed embodiment, a large language model and a booking assistant agent are used to provide users with additional service recommendations and refund and change rule prompts for their orders, directly reminding customers of relevant information, which is user-friendly for end users.

[0343] Regarding technical issue 5: The payment deadline, insurance deadline and flight change notification for delayed payment in the NDC interface can be sent directly to the end user in the form of a dialog box through the APP.

[0344] Regarding technical issue 6: The intelligent agent learns the business problems of the Aggregator and the airline in the NDC business, combines the log tracking system and the NDC interface standard document with the previous manual work order email processing problems to learn and handle abnormal processes, and reduce manual operation and maintenance costs.

[0345] At present, artificial intelligence technology is developing rapidly, and end users have an increasingly strong demand for convenience and intelligence. The current sales method through WEB or APP, where end users need to understand the business process to complete the NDC order operation, can no longer meet the needs of system users. The technical solution disclosed in this disclosure uses intelligent agent technology, which can complete the ticket sales business through dialogue communication, can reduce complex business logic, meet user needs, and improve user experience.

[0346] Based on the above method, the embodiment of the present disclosure also provides a ticket sales business processing system corresponding to the above method, such as Fig. 9 It includes: an agent auxiliary module, an order management module and a flight query module with an NDC interface tool; wherein the agent auxiliary module includes LLM, a prompt word management module, a booking assistant agent and an itinerary butler agent:

[0347] The intelligent agent auxiliary module is used to determine the user's demand information in the air ticket sales business; based on the demand information and the prompt word template, combined with the NDC knowledge base tool, call the large language model LLM to generate an NDC interface request message; use the generated NDC interface request message to call the NDC interface tool to obtain the NDC interface response message; based on the NDC interface response message and the prompt word template, combined with the NDC knowledge base tool, call the LLM to parse out the content information that matches the demand information; and send the parsed content information to the user terminal.

[0348] Based on the same inventive concept as the above disclosed content, the present disclosure also provides an electronic device accordingly. The electronic device of the present disclosure embodiment includes at least one electrically connected processor and only one memory, the memory is electrically connected to the processor, wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the method as described above.

[0349] It should be noted that the electrical connection between the above-mentioned units does not necessarily mean the connection between lines. An indirect connection method can be applied to the embodiments of the present disclosure as long as the purpose of the present disclosure is achieved.

[0350] Based on the same inventive concept, the present disclosure also provides a computer storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of the above method are implemented.

[0351] Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. A method for processing air ticket sales business, characterized in that: The method comprises, Determine the user's demand information in the air ticket sales business; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to generate the request message of the NDC interface; Use the request message generated by the NDC interface to call the NDC interface tool and obtain the NDC interface response message; Based on the NDC interface response message and prompt word template, combined with the NDC knowledge base tool, LLM is called to parse out the content information that matches the required information; The parsed content information is sent to the user terminal.

2. The method according to claim 1, characterized in that Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to generate the request message of the NDC interface, including: Determine the relevant information that matches the required information; Based on the requirement information and prompt word template, combined with the NDC knowledge base tool, call LLM to summarize the NDC interfaces that need to be used; Use the required NDC interface to call the NDC knowledge base tool to obtain the NDC interface protocol content and samples; Using the prompt word template, with the obtained NDC interface protocol content and samples as the background, and the determined associated information as the condition, combined with the NDC knowledge base tool, calling LLM, and generating the request message of the NDC interface, wherein the prompt word template used is the prompt word template that has a corresponding relationship with the summarized NDC interface that needs to be used.

3. The method according to claim 2, characterized in that Based on the NDC interface response message and prompt word template, combined with the NDC knowledge base tool, LLM is called to parse the content information that matches the required information, including: Using the prompt word template, with the NDC interface protocol content and samples as the background, and the NDC interface response content as the condition, combined with the NDC knowledge base tool, calling LLM, parses out the content information that matches the required information.

4. The method according to claim 2 or 3, characterized in that: The air ticket sales business is the flight enquiry link; The demand information is itinerary information, which includes: departure place, arrival place and planned time information; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including: Based on the itinerary information and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interfaces to be used are AirShopping and OfferPrice; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Use AirShopping and OfferPrice to call the NDC knowledge base tool to obtain the AirShopping and OfferPrice interface protocol content, flight query samples, and best price search samples.

5. The method according to claim 4, characterized in that Before determining the itinerary information of the user in the flight query phase, the method further includes: Receive itinerary plan description information from the user terminal and obtain user environment information; The itinerary plan description information includes: events, event related information and booking requirements, and the user environment information includes: login user name and login environment; Use the prompt word template, combined with the itinerary description information and user environment information, to call LLM to summarize the user's itinerary information; Send the summarized itinerary information to the user for confirmation; Determine the user's itinerary information during the flight query phase, including: The itinerary information confirmed by the user is determined as the itinerary information of the user in the flight query phase.

6. The method according to claim 5, characterized in that Determine the relevant information that matches the required information, including: Using the prompt word template, analyzing the user's implicit preference information based on the event and event-related information, adjusting the planned time, and obtaining the adjusted planned time; According to the departure and arrival places in the itinerary information, historical order information is obtained, combined with the NDC knowledge base tool, and LLM is called to summarize the user's itinerary preferences; Determine itinerary planning suggestions based on the adjusted planned time information, event information and prompt word template; Using the prompt word template, combined with the NDC knowledge base tool, and calling LLM, the itinerary preferences and itinerary planning suggestions are summarized to obtain itinerary-related information that matches the itinerary information.

7. The method according to claim 6, characterized in that Parse the content information matching the demand information into a flight result list; Before sending the parsed content information to the user terminal, the method further includes: Use the prompt word template to combine the user's itinerary preferences and the parsed flight combination list, use LLM to label each flight combination for recommendation and sorting, and recommend and sort according to the label order; Sending the parsed content information to the user terminal specifically includes: The flight result list recommended and sorted in the order of tags is sent to the user terminal.

8. The method according to claim 3, characterized in that The ticket sales business refers to the flight reservation process; the demand information refers to flight information and passenger information; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, determine the associated information that matches the demand information, including: Use the prompt word template, combine the flight list in short-term memory with the flight combination selected by the user and the passenger information, and parse the OfferId and passenger list of the flight combination; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including: Based on flight information and passenger information, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interface to be used is OrderCreate; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Use OrderCreate to call the NDC knowledge base tool to obtain the OrderCreate interface protocol content and create an order sample.

9. The method according to claim 8, characterized in that Using the prompt word template, with the NDC interface protocol content and samples as the background, and the NDC interface response content as the condition, combined with the NDC knowledge base tool, calling LLM, parses the content information that matches the required information, including: Use the prompt word template, take the OrderCreate interface protocol content and the order creation sample as the background, and the OrderCreate interface response content as the condition, combine with the NDC knowledge base tool, call LLM, and parse the order number and specific order content.

10. The method according to claim 9, characterized in that Label all unpaid orders and record the insurance and payment deadlines for each order; Sort the current unpaid orders by the insurance time limit and payment time limit, and use LLM to identify the time that is about to expire based on the current time point and the sorted current unpaid orders; Remind users to pay based on identified orders that are about to expire; When an unpaid order exceeds the insurance time limit, call the NDC knowledge base tool to obtain the next action of the airline to which the unpaid order belongs after the payment time limit has expired; Follow the next steps to process the unpaid order.

11. The method according to claim 2 or 3, characterized in that: The ticket sales business is the payment and ticket issuance process; the demand information is the ticket issuance demand; Determine the relevant information that matches the required information, including: Determine the payment order, payment method and payment voucher that match the user's ticketing needs; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including: Based on the ticketing requirement information and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interface to be used is OrderChange; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Use OrderChange to call the NDC knowledge base tool to obtain the OrderChange interface protocol content and payment and invoicing samples.

12. The method according to claim 2 or 3, characterized in that: The ticket sales business is an additional service reservation link; the demand information is an additional service reservation demand including an additional service type; Determine the relevant information that matches the required information, including: Using the prompt word template, combined with the user's additional service reservation needs and environmental information, combined with the NDC knowledge base tool, calling LLM, disassemble the user's additional service reservation needs, and obtain related information that matches the additional service reservation needs; Based on the additional service reservation requirements and prompt word template, combined with the NDC knowledge base tool, LLM is called to summarize the NDC interfaces that need to be used, including: Based on the additional service reservation requirements and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interfaces that need to be used are ServiceList and OrderChange; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Use ServiceList and OrderChange to call the NDC knowledge base tool to obtain the ServiceList and OrderChange interface protocol content, additional service query samples, and additional service samples.

13. The method according to claim 12, characterized in that The additional service reservation requirement also includes additional service preference information. Before the parsed content information is sent to the user terminal, the method further includes: Query the additional service types purchased by the current user through the order management module; Query the temperature and humidity of the departure place through the external interface; Call the NDC interface to obtain a list of available additional services; Based on the list of available additional services, the temperature and humidity of the departure point, and the additional service preference information, combined with the NDC knowledge base tool, the LLM is called to summarize, mark, and sort the additional service list according to the additional service preference; Send the parsed content information to the user terminal, including: The additional service list marked and sorted according to the additional service preference is sent to the user terminal.

14. The method according to claim 2 or 3, characterized in that: The ticket sales business is an additional service payment link; the demand information is an additional service payment demand; Determine the relevant information that matches the required information, including: Determine the payment additional service items and current orders that match the additional service payment requirements; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including: Based on the additional service payment requirements and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interface to be used is OrderChange; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Use OrderChange to call the NDC knowledge base tool to obtain the OrderChange interface protocol content and purchase additional service samples.

15. The method according to claim 2 or 3, characterized in that: The air ticket sales business is the process of proposing to modify the itinerary; The demand information is a request to modify the itinerary; Determine the relevant information that matches the required information, including: Using the prompt word template, combined with the user's request to modify the itinerary and environmental information, combined with the NDC knowledge base tool, calling LLM, disassemble the user's original itinerary and target itinerary; Use the prompt word template, based on the user's original itinerary, combined with the NDC knowledge base tool, call LLM, generate order query parameters, and call the order management function to query the user's changed orders; The user's changed order, original itinerary and target itinerary are used as related information matching the itinerary modification request; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including: Based on the proposed itinerary modification requirements and the prompt word template, combined with the NDC knowledge base tool, LLM is called and it is concluded that the NDC interface to be used is OrderReshop; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Call the NDC knowledge base tool to obtain the OrderReshop interface protocol content and ticket change query sample.

16. The method according to claim 2 or 3, characterized in that: The air ticket sales business mentioned above is the process of confirming and modifying the itinerary; The demand information is to confirm the need to modify the itinerary; Determine the relevant information that matches the required information, including: Determine the user's rebooking order that matches the confirmed itinerary modification needs and confirm the rebooking flight combination; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, call LLM and summarize the NDC interfaces that need to be used, including: Based on the confirmation of the itinerary modification requirements and the prompt word template, the NDC interface required to call the LLM summary is OrderChange; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Use OrderChange to call the NDC knowledge base tool to obtain the OrderChange interface protocol content and payment change samples.

17. The method according to claim 2 or 3, characterized in that: The air ticket sales business is the process of requesting cancellation of the trip; Before determining that the user has requested to cancel the trip, the method further includes: Using the prompt word template, combined with the user-entered cancellation order description and environmental information, combined with the NDC knowledge base tool, calling LLM, disassembled the key order information and the user's needs to cancel the order; Use the prompt word template to call LLM to generate query parameters based on the key information of the order; Use the query parameters to call the order management function and query the current valid order list that matches the query parameters; Use the prompt word template, combined with the NDC knowledge base tool, to call LLM and summarize the refund and change rules information of the queried order list; Return the order that needs to be canceled and the refund and change rules information to the user terminal for user confirmation.

18. The method according to claim 17, characterized in that The demand information is a request to cancel the trip; Determine the relevant information that matches the required information, including: Determine the refund information that matches the cancellation request; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to summarize the NDC interfaces that need to be used, including: Based on the request to cancel the trip and the prompt word template, the large language model LLM is called to summarize the OrderReshop when the NDC interface needs to be used; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Use OrderReshop to call the NDC knowledge base tool to obtain the OrderReshop interface protocol content and refund query samples.

19. The method according to claim 2 or 3, characterized in that: The air ticket sales business is the process of confirming and canceling the itinerary; the demand information is the demand for confirming and canceling the itinerary; Determine the relevant information that matches the required information, including: Identify orders that match the cancellation requirements; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to summarize the NDC interfaces that need to be used, including: Using the cancellation request and prompt word template, combined with the NDC knowledge base tool, LLM is called to conclude that the NDC interface to be used is OrderChange; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Call the NDC knowledge base tool to obtain the OrderChange interface protocol content and refund sample.

20. The method according to claim 2 or 3, characterized in that The ticket sales business is the flight change reminder link; Before determining the user's demand information in the air ticket sales business, the method further includes: Receive flight change notifications from airlines; Call the NDC knowledge base tool to obtain the OrderChangeNotif interface protocol content and flight change notification sample; Using the prompt word template, with the OrderChangeNotif interface protocol content and flight change notification sample as the background, combined with the NDC knowledge base tool, call LLM to summarize the airline's OrderChangeNotif interface request message content, including the order number, change type and change reason; A flight change message is sent to the user terminal, wherein the flight change message includes the order number, change type and change reason.

21. The method according to claim 20, characterized in that The demand information is to accept flight changes; Determine the relevant information that matches the required information, including: Identify orders and flight change notifications that match accepted flight changes; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to summarize the NDC interfaces that need to be used, including: The NDC interface that needs to be used to call the LLM summary to accept flight changes is OrderRetrieve; Use the NDC interface you want to use and call the NDC knowledge base tool to obtain the NDC interface protocol content and samples, including: Use OrderRetrieve to call the NDC knowledge base tool to obtain the OrderRetrieve interface protocol content and accept flight change samples.

22. The method according to claim 1, characterized in that The method further comprises: When there is an error or warning message, obtain the current NDC interface protocol content and sample containing the error or warning message; Use prompt word templates, obtain the current NDC interface protocol content and samples as the background, and the NDC interface response content as the condition. Combined with the NDC knowledge base tool, call LLM to parse the error code and error prompt content. Call the NDC knowledge base tool to obtain the cause and solution of the error code and error message content; Call the log tracking module function in the flight query module to obtain the NDC interface request message and return content of the operation process before the error; Use prompt word templates, the causes and solutions of error codes, and the request messages and return contents in the operation process as the background, combined with the NDC knowledge base tool, call LLM, and summarize the real error causes and solutions of the current operation process; If the solution needs to call the NDC interface for processing, the interface call will be automatically completed and the processing result will be returned. Otherwise, the actual error cause and solution will be returned to the user terminal.

23. A ticket sales business processing system, characterized in that: include: Agent assistance module, order management module and flight query module with NDC interface tool; The agent auxiliary module is used to determine the user's demand information in the ticket sales business; Based on the demand information and prompt word template, combined with the NDC knowledge base tool, the large language model LLM is called to generate the request message of the NDC interface; Use the request message generated by the NDC interface to call the NDC interface tool and obtain the NDC interface response message; Based on the NDC interface response message and prompt word template, combined with the NDC knowledge base tool, LLM is called to parse out the content information that matches the required information; The parsed content information is sent to the user terminal.

24. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; Memory, used to store computer programs; A processor is used to implement the steps of any method described in claims 1-22 when executing a program stored in a memory.

25. A computer storage medium, characterized in that The computer storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any method described in claims 1-22 are implemented.

Images