Order Response System

Abstract

To provide an order response system, method and program for processing an order after obtaining a desired pick-up time and determining in advance whether or not the time can be provided when receiving an order through a voice call with a user. [Solution] An order response system that accepts order details by inputting voice conversations with a user into a large-scale language model, and includes a time acquisition unit that acquires the desired time for receiving the order details from the user, a judgment unit that determines whether the order details can be provided at the desired time, an order acquisition unit that acquires the order details from the user after determining that the order details can be provided, and an order output unit that outputs the acquired order details to the store.

Description

[Technical Field]

[0001] The present invention relates to an order response system that processes orders through voice communication, and in particular to a technology that efficiently acquires order information from users and enables linkage with stores through natural dialogue processing using a large-scale language model. [Background technology]

[0002] Patent Document 1 discloses a technology that converts a request voice from a user into text, generates a voice for judgment to confirm the content, judges whether the user responds, and then transfers the call to an operator terminal and generates provisional data as necessary. This technology aims to streamline procedures by accepting requests through an automated voice response and transferring the call to an operator. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2025-97180 Summary of the Invention [Problem to be solved by the invention]

[0004] The technology described in Patent Document 1 has features such as confirmation of the request content by the user's voice and transfer to an operator, but it does not have functions specialized for order processing, and in particular does not take into consideration the delivery of items that require immediacy, such as food and drink.

[0005] Furthermore, when it comes to voice calls with users, conventional scenario-based processing has difficulty in providing flexible dialogue, making it difficult to realize natural responses.

[0006] Therefore, an object of the present invention is to provide a technique that solves the above problems. [Means for solving the problem]

[0007] According to the present invention, An order response system that accepts an order by inputting a voice conversation with a user into a large-scale language model, a time acquisition unit that acquires a desired time for receiving the order content from the user; a determination unit that determines whether the ordered content can be provided at the desired pickup time; an order acquisition unit that acquires the order details from the user after it is determined that the food can be provided; an order output unit that outputs the acquired order details to the store; An order response system comprising: [Effects of the Invention]

[0008] According to the present invention, when accepting an order through a voice call with a user, the desired pickup time is obtained and whether or not that time can be provided is determined in advance before processing the order, thereby significantly improving the accuracy of order acceptance and the user experience.

[0009] Furthermore, natural dialogue processing using a large-scale language model enables flexible and smooth ordering dialogue, unlike conventional fixed voice responses. Furthermore, the system automatically outputs order details to the store after acquiring them, reducing human intervention and reducing the burden on store operations, while also helping to prevent ordering errors. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a sequence diagram showing the overall configuration of an order response system according to an embodiment of the present invention. [Figure 2] 1 is a functional block diagram of an order response system according to an embodiment of the present invention. [Figure 3] 1 is a process flow diagram of an order response system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0011] The present invention will be described below by listing the contents of the embodiments. The present invention has the following configuration. [Item 1] An order response system that accepts an order by inputting a voice conversation with a user into a large-scale language model, a time acquisition unit that acquires a desired time for receiving the order content from the user; a determination unit that determines whether the ordered content can be provided at the desired pickup time; an order acquisition unit that acquires the order details from the user after it is determined that the food can be provided; an order output unit that outputs the acquired order details to the store; An order response system comprising: [Item 2] In the order response system according to item 1, An order response system, wherein the voice call with the user is conducted by natural dialogue processing using a large-scale language model. [Item 3] In the order response system according to claim 1, The order response system is characterized in that the order acquisition unit confirms the acquired order details with the user by voice and confirms the order details based on the user's response. [Item 4] In the order response system according to item 1, The order response system is characterized in that the time acquisition unit presents a time when the order can be picked up depending on the business hours of the store, and after presenting the time, asks the user about the desired time of pick-up. [Item 5] In the order response system according to item 1, An order response system comprising a transfer processing unit that automatically transfers the voice call to the store when the voice call with the user fails a predetermined number of times.

[0012] <Details of implementation form> Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0013] <Summary> The order response system of the present invention is configured to accept orders through voice communication with the user, and can accurately and flexibly grasp the order details through natural dialogue processing using a large-scale language model.

[0014] As shown in Figure 1, this system includes a time acquisition unit that acquires the user's desired pick-up time (at the store), a determination unit that determines whether the store can provide the order based on the desired time, an order acquisition unit that acquires the order after it has been determined that the order can be provided, an order output unit that outputs the order details to the store, and a transfer processing unit that automatically transfers the order to the store if the call fails a predetermined number of times. This configuration consistently automates the process from order reception to store collaboration, simultaneously reducing the human burden and improving service quality.

[0015] <Functional configuration> As shown in FIG. 2, the order response system according to the present invention is made up of a plurality of functional units for automating the process of accepting orders through voice calls with users.

[0016] First, voice input from the user is received via a voice call interface, and this voice is converted into text by the speech recognition unit. The converted text is input into the large-scale language model processing unit, which understands the context and accurately extracts the user's intent. When accepting an order, the time acquisition unit acquires the user's desired pickup time, and the determination unit checks whether that desired time matches the store's business hours and availability. If it is determined that the order can be delivered, the order acquisition unit acquires the order details, such as the specific product name and quantity, and links them to the store system via the order output unit. Furthermore, if the call fails after a certain number of attempts, the transfer processing unit is activated and automatically transfers the call to the store, allowing the call to continue without compromising user convenience.

[0017] The voice call interface is an interface for receiving voice signals sent by users through their terminals (telephones, smartphones, VoIP devices, etc.) and transferring them within the system. This interface supports common communication protocols such as the public switched telephone network (PSTN), SIP, and VoIP, and performs basic call control such as establishing call connections, transmitting voice signals, and detecting DTMF signals.

[0018] This interface also has a function to monitor conditions for call failure, such as when no voice input is made within a specified time or when the user's speech cannot be recognized continuously, and this also triggers the sending of a connection signal to the transfer processing unit (described later). This allows for a smooth switchover to a human operator even when the system is unable to respond.

[0019] The speech recognition unit has the function of converting the user's voice data input from the voice call interface into a character string (text) based on an acoustic model and a language model. In speech recognition, it is preferable to use a deep learning acoustic model (DNN-HMM, RNN-T, etc.) to absorb acoustic variations such as noise environments, the user's speaking rate, and intonation.

[0020] The converted text is treated as primary data for determining what the user is trying to order and is sent to the subsequent large-scale language model processing unit. Note that it is preferable that the speech recognition processing is performed in real time, and a configuration in which intermediate results are sent in stages during speech may also be adopted.

[0021] The large-scale language model processing unit is a core component that applies natural language processing (NLP) to the text data received from the speech recognition unit to understand order intent, extract slots, control dialogue, etc. The language model used here uses pre-trained transformer models such as GPT (Generative Pre-trained Transformer) and BERT to interpret the context from the natural language expressions spoken by the user and generate structured order data (e.g., menu name, quantity, desired pickup time, etc.).

[0022] Furthermore, even when a user uses ambiguous expressions or presents multiple options, LLM can appropriately interpret their meaning and generate follow-up questions as necessary to continue the natural dialogue, enabling flexible ordering dialogue that was difficult with conventional rule-based IVRs.

[0023] The output generated by the LLM is distributed to the time acquisition unit, order acquisition unit, etc., and converted into a data format appropriate for each process.

[0024] The time acquisition unit has the function of acquiring the date and time when the user wishes to receive the ordered product (desired pickup time). This configuration extracts time information (e.g., "I want to receive it at 6 p.m. today" or "30 minutes later would be better") contained in the output of the large-scale language model processing unit, converts it into a standardized time format (e.g., timestamp), and handles it. By appropriately converting relative time expressions in natural language (e.g., "tomorrow noon" or "15 minutes later"), the accuracy and consistency of the time information is ensured.

[0025] The time acquisition unit may also play a role in presenting "candidate times for pickup" that take into account constraints such as the store's business hours and the preparation time required for product provision. In this case, after the candidate times are presented, a dynamic dialogue can be configured in which the user can select or verbally answer the request for a pickup time again.

[0026] The determination unit determines whether the desired pickup time acquired by the time acquisition unit is a time when the product can actually be provided. Information used for the determination includes the store's business hours, the current time, order status, product inventory, and staff capacity. For example, even if the user requests pickup at 7 p.m., if the store closes at 6 p.m., the determination is made as "unavailable."

[0027] The determination unit may also have a function to generate and re-suggest alternative times for the user's desired time (e.g., "7 PM is unavailable, but 6:30 PM is available.") This allows for a high level of user satisfaction to be achieved while balancing the constraints of the user and the store.

[0028] The order acquisition unit has the function of acquiring order details (product name, quantity, options, notes, etc.) from the user after the judgment unit has determined that the product is "available." Because order details are acquired in a free-form input format during a voice call, the order acquisition unit normalizes order elements (slots) based on the results of LLM processing, and organizes them into structured data while reducing the risk of misrecognition.

[0029] Furthermore, the order acquisition unit has the function of first reading out the acquired order details to the user for confirmation, and then confirming the order based on the response. For example, it generates a confirmation prompt such as "Is it correct that your order is one fried chicken bento and one bottle of cola?", and if the user responds "yes," it executes the confirmation process. If there is an incorrect response, it also performs branching control to return to order acquisition again.

[0030] The order output unit is responsible for outputting the order data confirmed by the order acquisition unit to the store's management system (POS system, kitchen display system, slip printer, etc.). This output format can be designed to suit the store's business operations, such as API integration, CSV output, or FTP transmission.

[0031] In addition, the order output unit can also add metadata such as the user's contact information and pick-up time to the order details, allowing the store to smoothly prepare and manage the order.

[0032] The transfer processing unit has the function of automatically transferring the call to the store's manned telephone line when predetermined conditions are met, such as when a dialogue with the user is not established a certain number of times during a voice call, or when misrecognition or confusion continues. This transfer allows the user to smoothly switch to human assistance without giving up on their order, ensuring service continuity without compromising the UX (user experience).

[0033] Forwarding conditions can also be determined by combining multiple conditions using logical expressions, such as ``if the order cannot be confirmed within three round trips,'' ``if the same question is repeated more than twice,'' or ``if the user clearly expresses dissatisfaction.''

[0034] <Processing flow> The order response system according to the present invention automatically accepts orders from users via voice calls, and performs a series of processes to confirm the order and coordinate with the store. The overall process flow is explained below based on the flowchart shown in Figure 3.

[0035] The process starts with a "voice call initiation" step in response to a user making a call or receiving a call signal. In this step, it is detected that the user has started a call to the store to place an order, and a call session is established via a voice call interface.

[0036] Next, in the "Acquisition of Voice Signals (Call IF)" step, the user's voice is acquired in real time. The acquired voice signal is sent to the voice recognition unit, and then converted into character string data in the subsequent "Voice to Text Conversion (Voice Recognition Unit)" step. This voice recognition process is performed with high accuracy and speed using deep neural networks and end-to-end models.

[0037] The converted text data undergoes natural language processing using a large-scale language model (LLM) in the Semantic Analysis and Intent Extraction (LLM) step. This process extracts order information in slot format, such as the user's desired pickup time, product name, and quantity, and also processes it to generate a context-based response.

[0038] Next, in the "Acquisition of desired pick-up time (time acquisition section)," the time information included in the LLM processing result is extracted and converted into a standardized format. The converted desired pick-up time is passed to the "Determination of availability (determination section)," which checks it against the store's business hours, product availability, and congestion status to determine whether delivery is possible at the specified time.

[0039] If the result of the determination is "Available," the process proceeds to the "Acquire order details (order acquisition section)" step, where order details (menu, quantity, options, etc.) are acquired through voice dialogue with the user. After that, in the "Confirm order details and wait for response" step, the acquired order details are presented to the user again by voice, and a process is performed to wait for a confirmation response such as "Yes / No."

[0040] If the confirmation response here is "Yes," the process moves to the "Confirm Order" step, where the acquired order is officially confirmed. After confirmation, the order information is automatically sent to the store's management system (POS, kitchen display, slip printing, etc.) in the "Order Output Unit → Send to Store" step, and the process ends.

[0041] On the other hand, if the confirmation response is "No" or a similar negative response, the process proceeds to the "return to re-input process" step, and a loop process is performed in which the order acquisition process is restarted from the beginning.

[0042] Although not shown, during the order process, a condition determination is made as to whether the number of failed calls is within a predetermined number. This condition is triggered when the process fails a predetermined number of times due to voice recognition failure, an incomplete user response, unclear speech, etc. If it is determined that the predetermined number of attempts has been exceeded, the process proceeds to the "Automatically transfer call to store" step, where the current call is handed over to a manned store operator. This transfer process is realized by means of SIP control, number switching, etc., via the call control interface.

[0043] As described above, the order response system of the present invention has an advanced automated dialogue processing structure that includes everything from receiving an order via voice call to determining whether or not it can be provided, confirming the order, coordinating with the store, and even recovery measures in case the call fails. This series of processes makes it possible to simultaneously improve the user experience and increase the efficiency of store operations.

[0044] <Dialogue example> A specific example of a dialogue via this system will be described below.

[0045] <Dialogue example 1> The example of the dialogue example 1 is a case where the user wants to make a reservation by phone to pick up one fried chicken bento at the store at 6:00 PM. System: "Hello, this is the automated order entry system for store X. Would you like to order today?" User: "Yes, I'd like one fried chicken bento, please." System: "Okay. What time would you like to pick it up?" User: "Around 6pm today please." System: "You're picking up today at 6 PM. Let me check. Yes, we can make a reservation for 6 PM." System: "So, would you like to order one fried chicken bento and pick it up at 6:00 PM?" User: "Yes." System: "Your order has been received. Please pick it up at XX store at 6:00 PM. Thank you."

[0046] <Dialogue example 2> In the example of dialogue example 2, the user wishes to receive the item at 8 p.m., but the store's business hours are until 7 p.m. System: "Hello, this is the automated order taker. Please place your order." User: "I'd like to pick up a bowl of curry rice at 8pm." System: "Sorry, but our store is open until 7 PM, so we cannot accept pick-up orders until 8 PM. Would 7 PM work for you?" User: "Okay, I'll make it 6:30pm." System: "It's 6:30 PM. Would you like to order one curry rice and pick it up at 6:30 PM?" User: "Yes." System: "Thank you very much. We'll meet you at XX store at 6:30 PM."

[0047] <Dialogue example 3> The example in this dialogue example 3 is a case where the user's speech is unclear or inaudible, and the call is transferred to a store after a predetermined number of failed attempts. System: "Hello, this is the automated order taker. Please place your order." User: "Um... it's the same one as yesterday... what was it again...?" System: "Sorry. Could you please repeat the product name clearly?" User: "That one from yesterday. The set..." System: "I'm sorry, I didn't hear you correctly. Could you please repeat the name of the item you ordered?" (→The failed call count reaches the limit.) System: "Sorry, but our automated receptionist can't help you. We'll connect you to a live person." (→Automatically redirect to store)

[0048] <Effects of the present invention> According to the order response system of the present invention, the following specific effects can be obtained according to each component.

[0049] By acquiring the desired pick-up time through a voice call with the user and determining in advance whether the product can be provided at that time before processing the order, it is possible to prevent orders that cannot be fulfilled, improving work efficiency and satisfaction for both users and stores.In addition, by structuring the order details and automatically outputting them to the store, human error and workload can be reduced.

[0050] By implementing natural dialogue processing using a large-scale language model, interactions with users become much more flexible and natural compared to conventional scenario-based IVR, reducing stress during the ordering process. It can also handle complex order details and ambiguous expressions.

[0051] After an order is received, the details are confirmed by voice and the order is confirmed based on the user's response. This minimizes discrepancies in information between the user and the system and significantly reduces the risk of placing an incorrect order.

[0052] By configuring the system to present users with available pickup times based on the store's business hours and supply capacity, realistic scheduling becomes possible, reducing the rate of order rejections and contributing to store capacity management.

[0053] By providing a function that automatically transfers the call to the store if a certain number of failed conversations occur during a voice call, it is possible to continuously follow up with the user without interrupting their ordering experience, preventing opportunity loss.

[0054] <Other embodiments> <1> Delivery order The order response system of the present invention is not limited to accepting takeout orders, but can also be modified to accommodate home delivery orders or delivery orders, in which products are delivered to a delivery address specified by the user. In this modification, in addition to the order details obtained via voice communication, the user's desired delivery time and delivery address information are also acquired.

[0055] To handle such delivery orders, the time acquisition unit acquires information on the desired delivery time from the user and also has the function of receiving the address or location information of the delivery destination. Since the desired delivery time needs to take into account the time it takes to deliver compared to regular takeout, the time acquisition unit calculates the standard delivery time by referring to information such as the distance along the delivery route, past delivery times, and traffic conditions.

[0056] The determination unit compares the calculated delivery time with the user's desired delivery time, and determines whether the product can be delivered to the user's home while maintaining the specified quality if it is prepared and shipped from the store. For example, if the user requests delivery in 30 minutes, and the estimated delivery time taking traffic conditions into account is 45 minutes, the determination is "not possible." On the other hand, if the store's preparation time + delivery time is before the user's desired time, the determination is "possible."

[0057] Furthermore, in addition to product information, the order acquisition unit can also acquire additional items specific to home delivery through voice dialogue, such as delivery method (motorcycle, bicycle, walking, etc.), payment method (cash, electronic money, etc.), and whether or not the customer wishes to be notified of delivery status via chat.

[0058] When the order output unit sends the acquired delivery order information to the store system, it outputs it in a format that includes information such as the delivery address, desired delivery time, and delivery method. This allows store staff to accurately prepare for delivery. Furthermore, if the store collaborates with an external delivery company, the order output unit may be configured to send the necessary information directly to the company's platform via an API.

[0059] In this way, by applying the configuration of the present invention to delivery orders, it is possible to consistently realize automatic reception, automatic acceptance / rejection, and automatic output processing of delivery orders in the same way as takeout orders, through natural dialogue using voice calls. This makes it possible to provide a delivery service with high customer satisfaction while reducing the human resources required for telephone reception.

[0060] That is, the present invention also includes the configurations disclosed below. An order response system that accepts delivery orders by inputting voice conversations with users into a large-scale language model, a time acquisition unit that acquires the order details and desired delivery time from the user and also acquires delivery destination information; a determination unit that determines whether the ordered content can be delivered within the desired delivery time based on the distance to the delivery destination and traffic conditions; an order acquisition unit that acquires the order details after it is determined that delivery is possible; an order output unit that outputs the acquired order details and delivery address information to the store or delivery company; A delivery order response system comprising:

[0061] <2> Drive-through pickup The order response system of the present invention can also be applied to order reception between a store and a user who arrives by car in a drive-through lane installed in the store. In the drive-through mode, when the user enters the lane, a voice call is initiated through a microphone / speaker terminal installed in the store, and the order response system of the present invention automatically responds by voice.

[0062] This system receives voice input from users in real time and acquires order details in a natural conversational format by performing dialogue processing using a large-scale language model. When a user communicates the desired product and quantity, the order acquisition unit temporarily stores the information, and then the judgment unit checks the store's sales status (stock, cooking status, etc.) to determine whether the product can be provided.

[0063] Depending on the result of the judgment, the user will receive an automated voice message saying, "We can prepare the item you ordered," or "Sorry, XX is currently sold out." If it is determined that the item can be provided, the acquired order information is sent directly to the kitchen (cooking department) in real time by the order output unit, and cooking begins.

[0064] In this way, by applying the order response system of the present invention to a drive-through system, it becomes possible to receive orders smoothly and accurately even from users in vehicles, reducing the human burden compared to conventional manual responses while ensuring the speed of delivery and consistency of service.

[0065] <3> Ticket collection The order response system of the present invention can also be applied to a system in which users place orders using terminals (unmanned ordering machines) or ticket vending machines installed inside or outside the store. In this configuration, in addition to ordering by voice call, it is possible to provide a function for presenting and confirming order details and available pick-up times through cooperation with a kiosk terminal.

[0066] In particular, the time acquisition unit responds to user specifications such as "desired pick-up time period" or "desired pick-up time" while the user is placing an order on the terminal, and displays the available pick-up time in real time based on the store's supply capacity, cooking schedule, terminal usage status, etc. For example, it may display something like "You can pick up your order in 15 minutes from now during the current time period," and the user can agree to or change this.

[0067] The decision unit comprehensively considers the store's business hours, supply capacity, stock status, etc., and decides whether to provide the requested pick-up time input through a kiosk or ticket vending machine. Regarding business hours, if the hours during which a specific terminal is in use (e.g., the operating hours of the ticket vending machine) differ from the store's business hours, the decision unit can be configured to handle these constraints as well as other conditions for determination.

[0068] Furthermore, when the voice communication system of the present invention is integrated with a kiosk / ticket vending machine, a distributed configuration may be adopted in which part or all of the order processing is performed by voice communication, and then payment processing and ticket issuing are performed at the kiosk terminal. This increases user convenience while reducing the workload on the store.

[0069] <4> Tabletop ordering system The order response system of the present invention can also be applied to a so-called tabletop ordering system, where users place orders using tablet terminals or smartphones installed at each table in the store. In this configuration, orders and their serving times are controlled not through voice calls but through an interactive interface (voice or text-based) with the in-store terminal.

[0070] In this case, the time acquisition unit acquires the desired pick-up or delivery time specified by the user (e.g., "order immediately" or "order after dinner"), and calculates whether delivery is possible at that time by referring to the store's delivery status and operation status. At the same time, it also determines whether the time period exceeds the store's last order time.

[0071] The determination unit determines whether or not the order can be provided based on the last order time set by the store, depending on the time the order was entered. For example, if a store has a last order time of 9:30 PM and a user orders a main dish and dessert together at 9:29 PM, it is possible to make a partial determination, such as "only the main dish can be provided, but not the dessert."

[0072] This embodiment enables detailed order acceptance that takes into account the length of time customers spend in the restaurant, contributing to balancing the kitchen load and managing operations before and after the last order. In addition, by visualizing the order status and serving time for each table, it is possible to improve the service quality and optimize turnover efficiency throughout the restaurant.

[0073] <Modification> The order response system of the present invention can be implemented in a more versatile manner by the following modifications.

[0074] (Variation 1: Multilingual system) In the order response system of the present invention, the large-scale language model processing unit uses a multilingual natural language model (e.g., multilingual GPT, mBERT, etc.), making it possible to respond to orders in languages ​​other than Japanese (English, Chinese, Korean, etc.), thereby further improving convenience for foreign visitors to Japan and multinational users.

[0075] (Variation 2: Customizing store voice using voice synthesis) By using a speech synthesis engine that can set different speaker voices, phrases, dialects, etc. for each store, AI responses can be provided while maintaining the store's brand image and customer service style, providing a more natural and friendly user experience.

[0076] (Variation 3: Smartphone app linkage) This system can be configured to work with a smartphone application, in which case it can add a function that uses the user's location information, order history, point information, etc. to suggest stores where items can be picked up and recommended products, making it possible to provide a more personalized ordering experience.

[0077] (Variation 4: Optimizing response accuracy through automatic learning) It can also be implemented as a self-learning configuration that records and analyzes user interaction logs, order confirmation rates, and confirmation responses, and uses this data to improve speech recognition accuracy and the response accuracy of the dialogue model. With this type of implementation, the system can improve its response accuracy with repeated operation.

[0078] <Hardware configuration example> Each of the above-mentioned functional blocks can be configured by, for example, hardware provided in a server device (terminal device), a DSP (Digital Signal Processor), or software. For example, when configured by software, each of the above-mentioned functional blocks is actually configured with a CPU, RAM, ROM, etc. of a computer, and is realized by the operation of a program stored in a recording medium such as RAM, ROM, a hard disk, or a semiconductor memory.

[0079] The above-described embodiment is merely an example for facilitating understanding of the present invention, and is not intended to limit the present invention. The present invention can be modified and improved without departing from the spirit thereof, and it goes without saying that the present invention includes equivalents thereof.

Claims

1. An order response system that accepts an order by inputting a voice conversation with a user into a large-scale language model, a time acquisition unit that acquires a desired time for receiving the order content from the user; a determination unit that determines whether the ordered contents can be provided at the desired pickup time according to the store's business status and provision capacity; an order acquisition unit that acquires the order details from the user after it is determined that the food can be provided; an order output unit that outputs the acquired order details to the store; An order response system comprising: The business status includes the store's business hours, current time, order status, or congestion status, An order response system characterized in that the supply capacity includes product inventory or staff response capacity.

2. 2. The order response system according to claim 1, An order response system, wherein the voice call with the user is conducted by natural dialogue processing using a large-scale language model.

3. 2. The order response system according to claim 1, The order response system is characterized in that the order acquisition unit confirms the acquired order details with the user by voice and confirms the order details based on the user's response.

4. 2. The order response system according to claim 1, The order response system is characterized in that the time acquisition unit presents a time when the order can be picked up depending on the business hours of the store, and after presenting the time, asks the user about the desired time of pick-up.

5. 2. The order response system according to claim 1, An order response system comprising a transfer processing unit that automatically transfers the voice call to the store when the voice call with the user fails a predetermined number of times.

6. An order response program that accepts order details by inputting voice conversations with a user into a large-scale language model, Computer, functioning as a time acquisition means for acquiring a desired time for receiving the order contents from the user; functioning as a determination means for determining whether the ordered contents can be provided at the desired pickup time according to the store's business status and provision capacity; After it is determined that the product can be provided, the device functions as an order acquisition unit that acquires the order details from the user; functioning as an order output means for outputting the acquired order details to the store; 1. An order response program, comprising: The business status includes the store's business hours, current time, order status, or congestion status, The order response program is characterized in that the supply capacity includes product inventory or staff response capacity.

7. An order response method for accepting an order by inputting a voice conversation with a user into a large-scale language model, comprising: obtaining a desired time to receive the order from the user; a step of determining whether the ordered contents can be provided at the desired pickup time according to the store's business status and provision capacity; a step of acquiring the order details from the user after it is determined that the product can be provided; outputting the acquired order details to the store; 1. An order response method comprising: The business status includes the store's business hours, current time, order status, or congestion status, An order response method characterized in that the supply capacity includes product inventory or staff response capacity.

8. An order response server that receives order details by inputting voice conversations with users into a large-scale language model, a time acquisition unit that acquires a desired time for receiving the order content from the user; a determination unit that determines whether the ordered contents can be provided at the desired pickup time according to the store's business status and provision capacity; an order acquisition unit that acquires the order details from the user after it is determined that the food can be provided; an order output unit that outputs the acquired order details to the store; An order response server comprising: The business status includes the store's business hours, current time, order status, or congestion status, An order response server characterized in that the supply capacity includes product inventory or staff response capacity.

Images