KR102415999B1 - A method to identify a food menu using speech recognition and a system using the same - Google Patents

Abstract

Provided in the present invention is a method for identifying a delivery order, in a state in which a recognition module is included in an Artificial Intelligence (AI) callbot system and the recognition module for recognizing a delivery order is trained, comprising the following steps of: allowing a call center infrastructure system to receive an order call from a user terminal; allowing the AI callbot system to receive the order call from the call center infrastructure system and check a delivery order included in the order call by using the recognition module; allowing the AI callbot system to register an order sheet including the delivery order in a cloud order system; allowing a point of sale (POS) terminal in a store to receive the order sheet from the cloud order system and display information included in the order sheet for a preset time on a display so that a responsible person in the store can confirm the order; if there is no confirmation of the order for the preset time, allowing the order call to be automatically canceled, and if the order is confirmed for the preset time, allowing the POS terminal to print the order sheet through a printer in the store and deliver information included in the order sheet to a delivery agency device. A purpose of the present invention is to have a high success rate of voice recognition through the AI callbot system for the delivery order call.

Description

A method for identifying a food menu through voice recognition and a system using the same

The present invention relates to a method of identifying a food menu through voice recognition, wherein a recognition module is included in the AI callbot system, and in a state in which the recognition module for recognizing a delivery order is learned, the call center infrastructure system is receiving an order call of The AI callbot system, receiving the order call from the call center infrastructure system, and using the recognition module to check the delivery order included in the order call; registering, by the AI callbot system, an order book including the delivery order in the cloud order system; The in-store POS terminal receives the order form from the cloud order system, and displays the information included in the order form on the screen for a preset time so that the in-store person in charge confirms the order; and if there is no order confirmation for the preset time, the order call is automatically canceled, and when the order is confirmed for the preset time, the POS terminal prints the order form through an in-store printer, It relates to a method comprising the step of passing the included information to a delivery agency device.

The domestic delivery market is growing rapidly with an estimated market size of KRW 28 trillion in 2021 and an average annual growth rate of 12.3%. Although delivery orders are increasing through delivery apps, the proportion of orders made over the phone is still high.

Currently, in the case of a phone order by a commercialized AI speaker or AI voice assistant, it performs voice recognition over a wide area and shows a low recognition success rate. Due to this, there was an aspect that an active response was impossible in case of an emergency. In addition, there is a limit to voice recognition, so an AI delivery service that specializes in recognition areas such as menus and shows a high recognition success rate was needed.

Accordingly, the present inventor intends to propose a method for identifying a food menu through voice recognition and a system using the same.

An object of the present invention is to solve all of the above problems.

Another object of the present invention is to have a high voice recognition success rate through an AI call bot for a consumer's voice included in a delivery call.

Another object of the present invention is to improve service quality through similarity analysis, prediction, and accuracy improvement of recognized words.

In order to achieve the object of the present invention as described above and to realize the characteristic effects of the present invention to be described later, the characteristic configuration of the present invention is as follows.

According to an aspect of the present invention, in a method for recognizing a delivery order, a recognition module is included in the AI callbot system, and in a state in which the recognition module for recognizing a delivery order is learned, the call center infrastructure system is a consumer's terminal receiving an order call from; The AI callbot system, receiving the order call from the call center infrastructure system, and using the recognition module to check the delivery order included in the order call; registering, by the AI callbot system, an order book including the delivery order in the cloud order system; The in-store POS terminal receives the order form from the cloud order system, and displays the information included in the order form on the screen for a preset time so that the in-store person in charge confirms the order; and if there is no order confirmation for the preset time, the order call is automatically canceled, and when the order is confirmed for the preset time, the POS terminal prints the order form through an in-store printer, A method is provided, comprising the step of delivering the included information to a delivery agent device.

According to another aspect of the present invention, in a system for recognizing a delivery order, a recognition module is included in the AI callbot system, and in a state in which the recognition module for recognizing a delivery order is learned, an order call from a consumer's terminal Call center infrastructure system to receive; The AI callbot system for receiving the order call from the call center infrastructure system, confirming the delivery order included in the order call using the recognition module, and registering the order book including the delivery order in the cloud order system; and receiving the order form from the cloud order system, displaying the information included in the order form on the screen for a preset time so that the person in charge in the store confirms the order, and if there is no order confirmation for the preset time, the Including an in-store POS terminal that automatically cancels the order call, and prints the order form through an in-store printing device when the order is confirmed for the preset time, and delivers the information included in the order to the delivery agency device A system is provided.

According to the present invention, the following effects are obtained.

The present invention has an effect of having a high voice recognition success rate through an AI call bot for the voice of the consumer included in the delivery call.

In addition, the present invention has an effect of improving service quality through similarity analysis, prediction, and accuracy improvement of recognized words.

1 is a diagram showing a schematic configuration of a system for an order automation service according to an embodiment of the present invention.
Figure 2 is a flow chart showing the process until the delivery of the order call to the delivery agency app according to an embodiment of the present invention.
3 is a diagram illustrating a degree of similarity of recognized food menus according to an embodiment of the present invention.
4 is a diagram illustrating a process of displaying a visual order screen on a consumer's terminal according to an embodiment of the present invention.
5 is a diagram illustrating a screen of a backup counselor terminal supported by a backup counseling server according to an embodiment of the present invention.
6 is a diagram illustrating information that can be checked in an integrated management system according to an embodiment of the present invention.
7 is a diagram showing a detailed configuration of a system for an order automation service according to an embodiment of the present invention.
8 is a view showing the progress of the AI ordering service according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [0012] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [0010] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [0010] Reference is made to the accompanying drawings, which show by way of illustration specific embodiments in which the present invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the present invention. It should be understood that the various embodiments of the present invention are different but need not be mutually exclusive. For example, certain shapes, structures, and characteristics described herein with respect to one embodiment may be implemented in other embodiments without departing from the spirit and scope of the invention. In addition, it should be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the present invention. Accordingly, the detailed description set forth below is not intended to be taken in a limiting sense, and the scope of the present invention, if properly described, is limited only by the appended claims, along with all scope equivalents as those claimed. Like reference numerals in the drawings refer to the same or similar functions throughout the various aspects.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to enable those of ordinary skill in the art to easily practice the present invention.

1 is a diagram showing a schematic configuration of a system for an order automation service according to an embodiment of the present invention.

2 is a flowchart illustrating a process up to delivery of an order call to a delivery agency according to an embodiment of the present invention.

As can be seen in Figure 1, in the overall system of the present invention, AI callbot system 100, call center infrastructure system 200, backup consulting server 300, cloud ordering system 400, POS terminal 500, delivery agency device 600 and the like, and an order automation service may be provided using each of them. A detailed structure of each system will be described with reference to FIG. 7 to be described later, and processes performed in each system will be described below. However, in the POS terminal 500, both a service solution in the form of an app or web with the same function installed in a separate terminal and a general smart phone are possible.

First, a consumer (customer) may request an order using a terminal (phone). The call center infrastructure system 200 may receive an order call from the consumer's terminal (S210). Here, the order call corresponds to a call (call) for requesting delivery of food, etc. sold in the store.

Next, the AI callbot system 100 may receive the order call from the call center infrastructure system 200 and confirm the delivery order included in the order call using the recognition module 110 ( S220 ).

The recognition module 110 included in or linked to the AI callbot system 100 includes an AI engine and the like, and can support AI-based voice recognition technology through this. For reference, the AI engine may include a speak to text (STT) engine, a text to speak (TTS) engine, or a hybrid engine that considers all of the STT function, TTS function, and NLP (Natural Language Process) function.

Specifically, the AI callbot system 100 may recognize the food menu and delivery address included in the order call through the recognition module 110 . However, in the present invention, the recognition of the food menu will be focused on. That is, it recognizes the consumer's voice order "one golden olive chicken, one olive boneless chicken" made on the order call. In order to increase the accuracy of food menu recognition, iterative learning should be performed in advance for the recognition module 110 , and the performance of the recognition module 110 can be improved by registering similar words. First, let's take a look at the learning process below.

The recognition module 110 includes a plurality of parameters (a kind of weight, weight, etc.) including a first parameter and a second parameter, and a process of deriving a correct result value by changing them will correspond to a learning process.

In the present invention, in order to increase the speech recognition rate, the recognition module can be updated by learning the name of the sales menu and learning the words with a quantity of 4 or less.

First, the recognition module 110 includes a first parameter and a second parameter, which may each have a relationship with a food menu or quantity. Of course, one parameter rather than different parameters may be included to have relevance to the food menu and quantity.

The AI callbot system 100 may derive a first difference value by passing the first character for the test through the recognition module 110 and comparing the calculated result value with the first character for the correct answer. Here, the first character for the test may include a word representing a food menu, such as 'crispy chicken' and 'crispy chicken'. In addition, the first character for the correct answer may include a word indicating an accurate food menu name, such as 'crispy chicken'.

Next, the AI colbot system 100 may update the first parameter based on the first difference value. Specifically, if the first character for the test is 'Crunchy Chicken', there is a difference (first difference value) from 'Crunchy Chicken', which is the first character for the correct answer. can

In a state in which the first parameter included in the recognition module 110 is updated, the AI callbot system 100 may perform a similar word registration operation for a specific food menu. Of course, in some cases, the similar word registration operation may be performed regardless of whether the first parameter is updated. In the following, a similar word registration operation will be described with reference to FIG. 3 .

3 is a view exemplarily showing the degree of similarity of recognized food menus according to an embodiment of the present invention.

The AI callbot system 100 passes a plurality of test 1-1 characters (ex crispy chicken) representing a specific food menu through the recognition module, thereby allowing a plurality of predictive 1-1 characters (ex basacan chicken, crispy chicken) , Crispy Chicken, etc.)

That is, the voice recognition is performed while passing the 1-1 character (ex crispy chicken) for a test for a specific food menu through the recognition module 110, and the voice recognition results for the plurality of test 1-1 characters are displayed. It may be a plurality of predictive 1-1 characters. The plurality of test 1-1 characters have different voices, pitches, etc., so that the recognition module 110 can perform voice recognition differently, and thus a plurality of different predictive 1-1 characters (ex Basakan) chicken, crispy chicken, crispy chicken, etc.) may be derived.

The AI callbot system 100 may calculate each coordinate value according to the frequency of the plurality of 1-1 characters for prediction. As can be seen in FIG. 3( a ), a graph (a mountain-shaped graph) that is usually the highest in the middle and becomes smaller as it goes away can be formed. For reference, the y-axis may represent the frequency, and the x-axis may represent the different 1-1 characters for prediction (eg, written according to the mountain-shaped graph, Korean dictionary order, etc.).

Next, the AI callbot system 100 sets the 1-1 character for a specific prediction having the highest frequency as the 1-1 character for a specific correct answer, based on this, the first 1-for prediction for a predetermined distance within a preset distance One character may be determined as the 1-1 character for the specific correct answer.

When looking at FIG. 3( a ), Crispy Chicken is the 1-1 character for specific prediction with the highest frequency and is located in the center of the graph, and 'Crunchy Chicken' is located when moving to the left from it, and it moves more from the center to the right. You can see where 'Basakan Chicken' is located.

The 'crispy chicken' and 'basacan chicken' are both located within a preset distance from the center (the position of the 'crispy chicken'), so 'crispy chicken' and 'basacan chicken' are the 1-1 characters for a predetermined prediction and are used for the specific correct answer It may be determined as the 1-1 character ('crispy chicken'). That is, predetermined 1-1 characters for prediction located within a preset distance from the center may be set as one group and all may be determined as 1-1 characters for a specific correct answer.

The graph according to the frequency may be represented as a group in a circle shape as shown in FIG. 3(b) instead of FIG. 3(a).

Using the most frequent specific prediction 1-1 character (ex crispy chicken, golden olive chicken, etc.) as the reference coordinate, the frequency of the specific prediction 1-1 character differs within a preset distance (r) based on this coordinate (Ex. Crispy Chicken, Basakan Boneless Chicken, Golden Olive Chicken Boneless, etc.) can be expressed as a circle-shaped group (centered on the 1-1 character for specific prediction).

As a result, all of the 1-1 characters for a predetermined prediction included in one circle-shaped group may be determined as 1-1 characters for a specific correct answer.

On the other hand, the AI callbot system 100 passes the second character (ex one, one, one, two, hidden word, etc.) for testing indicating the quantity through the recognition module 110 and returns the calculated result value to the second correct answer A second difference value may be derived by comparing the characters (eg, one, two, twenty, etc.), and the second parameter may be updated based on the second difference value.

In addition, the AI callbot system 100 passes the second character for testing (ex one, two, three, etc.) representing a value of 4 or less through the recognition module 110 and returns the calculated result value to the second correct answer A second difference value may be derived by comparing with the text, and the second parameter may be updated based on the second difference value.

That is, the AI colbot system 100 may learn the recognition module 110 using a number indicating a value of 4 or less. This may enable more accurate voice recognition by intensively learning the recognition module 110 for four or less numbers rather than performing learning for each of a large number of numbers.

Meanwhile, the AI callbot system 100 may perform Natural Language Processing (NLP) processing based on Named-entity recognition (NER) using the recognition module 110 .

Any one food menu (eg whole chicken) can be divided into detailed categories (ex bones, boneless meat, etc.), and the detailed category can be set as an option.

In this case, the AI callbot system 100 may use the recognition module 110 to recognize a food menu, an option, and a quantity based on the order expression included in the order call, respectively, and confirm the delivery order.

For example, the AI callbot system 100 uses the recognition module 110 to use the recognition module 110, based on the order expression included in the order call, food menus (eg seasoning, fried), options ( You can check the delivery order by recognizing each bone, boneless meat) and quantity (ex 1).

This is to recognize and confirm the exact delivery order in the state where the customer's expression is various (ex, please give it a fried set, two boneless meat, two fried seasonings, etc.).

The call center infrastructure system 200 can support multiple phone orders in one store, so that when a plurality of consumers place an order by phone in one store at the same time, it is possible to place a simultaneous call order instead of a call. This is because the AI callbot system 100 can automatically respond to an order call received in the call center infrastructure system 200 .

In addition, the AI callbot system 100 may recognize the delivery address through the recognition module 110 and compare the delivery address with addresses included in an external database to determine the delivery address. It is possible to recognize the consumer's voice order "96, Guuigangbyeon-ro, Gwangjin-gu, Seoul" made on the order call. In some cases, secondary access to the address DB of the Ministry of Public Administration and Security determines whether the recognized address ("96 Guuigangbyeon-ro, Gwangjin-gu, Seoul") exists, and if the matching address is included in the DB, the corresponding delivery address can be confirmed. In order to increase the accuracy of the delivery address recognition, iterative learning should be performed on the recognition module 110, and in this case, the standard address included in the standard address DB of the Ministry of Public Administration and Security may be used.

4 is a diagram illustrating a process of displaying a visual order screen on a consumer's terminal according to an embodiment of the present invention.

When the AI callbot system 100 determines that the delivery order has been recognized, it transmits the delivery order to the consumer's terminal and, through verification, registers the order book including the delivery order in the cloud order system 400 based on this. can

Conversely, when it is determined that the AI callbot system 100 does not recognize the delivery order, it displays the order screen (visible ARS) on the consumer's terminal, and obtains the delivery order based on the input value from the consumer's terminal. have.

Specifically, in a state in which the order screen is displayed on the consumer's terminal, the AI callbot system 100 obtains a delivery order based on an input value (eg fried, bone, one, etc.) input from the consumer's terminal. it can be When the accuracy of voice recognition using the recognition module 110 is low, the AI callbot system 100 may support obtaining an accurate delivery order using the order screen (visible ARS).

That is, the AI callbot system 100 recognizes the consumer's voice (ex. one fried boneless meat) and provides a corresponding pre-stored voice (ex is one fried boneless meat correct? Please say yes or no) to communicate with the consumer Conversation can be conducted by voice, but in some cases, it recognizes the consumer's voice (ex. one fried boneless meat) and displays the corresponding ARS screen (“Is one fried boneless meat correct? Enter Yes or No on the screen of the consumer terminal) Please do it" display), the consumer may conduct a conversation by voice or input (ex terminal touch, etc.), and the AI callbot system 100 displays a changing order screen.

In addition, the AI callbot system 100 may not clearly recognize the consumer's voice. For example, although the consumer said "two crispy chickens" and "Seoul Yeoksam Station", the AI callbot system 100 may not recognize it as "two Basakan chickens" and "Seoul Yeongsam Station".

The AI callbot system 100 recognizes that the delivery menu or delivery address is unclear based on the pre-stored delivery menu (pre-stored for food sold at each store) and the pre-stored address (in the standard address DB of the Ministry of Public Administration and Security). A support request message may be transmitted to the backup consultation server 300 .

5 is a diagram illustrating a screen of a backup counselor terminal supported by a backup counseling server according to an embodiment of the present invention.

The backup consultation server 300 says that the recognition is unclear among the information about the customer and the store who made the order call ('1' in the drawing), the connection step, the delivery menu recognition step using the AI engine, and the delivery address recognition step using the AI engine. The determined specific step ('2' in the drawing), delivery information including delivery or packaging ('3' in the drawing), the recognized delivery menu order product ('4' in the drawing), the recognized delivery menu and recognition The payment amount based on the delivered delivery address ('5' on the drawing), the recognized delivery address ('6' on the drawing), the consumer's order history using the order call ('7' on the drawing), At least one of a consultation conversation between the consumer and the AI engine may be displayed on the screen of the terminal of the backup agent.

First, the information on the consumer and the store who made the order call may include a telephone number of the consumer (customer), a store branch name, and the like. In addition, on the screen of the backup agent's terminal, the AI counseling departure step includes a connection step, a menu recognition step, and an address recognition step, among which communication connection is difficult or voice recognition is unclear in the process of processing an order call. The determined specific step may be displayed. That is, it indicates whether an error occurred at any stage.

Here, the connection step may mean a case in which an error occurs in the connection (communication) process between the consumer's terminal and the AI callbot system 100, and in some cases, may not be included in the specific step.

In addition, as delivery information, whether it is a package order or a delivery order may be included and displayed, and an AI voice-recognized order product, AI-voice-recognized delivery address, etc. may be displayed on the consumer's terminal. The backup consultation server 300 may support the backup agent to change/add/delete orders for products and delivery addresses.

In addition, in the case of the payment amount displayed on the consumer's terminal, the delivery cost based on the location of the address to be delivered as well as the ordered product may be considered.

The consumer's order history using the AI consultation conversation may also be displayed on the consumer's terminal, and the date and time of each AI consultation conversation, whether the consultation was successful, the contents of the consultation conversation, etc. may be included.

After all, the backup consultation server 300 allows the backup agent to supplement the delivery menu or delivery address based on the order history, whether the AI consultation is deviated from which stage, and the conversation between the consumer and the recognition module (AI engine). You can complete your order. For reference, the backup consultation server 300 may play a supplementary role to the voice recognition method by the AI engine, but in some cases (eg, when a consumer requests, etc.) the voice by the recognition module (AI engine) It is also possible to have the backup agent complete the order form (delivery menu, delivery address, etc.) without realizing it.

The AI callbot system 100 may register an order book including a delivery order (order menu, order address) in the cloud order system 400 ( S230 ). The order menu and order address may correspond to results derived from AI voice recognition or supplementation with a backup consultation server.

In addition, after the ordered food is delivered to the consumer, when a complaint is made that the delivery content is wrong, the function of reproducing the recorded voice file or the recorded text of the telephone order is a backup counseling server service or solution 300 You can check the difference between the customer's order and delivery.

In addition, the backup counseling server 300 clearly displays the part that the recognition module 110 has not recognized indefinitely with a separate display in order to shorten the service time, so that the backup counselor does not ask for duplicate voice recognition contents. can do. Specifically, separate indications (ex display of different colors, check marks, etc.) for the unclearly recognized part (ex. 1 bowl of Jjamppong) among the recognized plurality of delivery menus (ex 1 bowl of Jjajangmyeon, 1 bowl of Jjambbong, 1 bowl of sweet and sour pork) can be displayed with

Next, the POS terminal 500 installed in the store may receive an order form from the cloud order system 400 and display information included in the order form on the screen for a preset time.

The POS terminal 500 is a device controlled by a person in charge in the store, and can also be accessed from the cloud order system 400 . In some cases, unlike in FIG. 1 , a franchise server may be placed between the cloud ordering system 400 and the POS terminal 500 to support communication between the two.

In addition, the POS terminal 500 may display the information included in the order form on the screen for a preset time so that the person in charge in the store can confirm the order. That is, the POS terminal 500 may pop-up the order food and order place included in the order form on the screen for a preset time (ex 5 minutes) so that the person in charge in the store clicks the order confirmation button to confirm the order ( S240).

When the information included in the order is in the pop-up state (for a preset time), if there is no order confirmation, the order call is canceled and the pop-up information may also disappear from the screen. In some cases, the person in charge of the store may cancel the order call while inputting the reason for canceling the order call (eg, delayed order, exhausted materials, etc.).

When it is confirmed that the order call has been canceled, the AI callbot system 100 or the cloud order system 400 may transmit an order call cancellation message and a reason for cancellation to the consumer's terminal. In addition, it is possible to recommend a menu and a store corresponding to the ordered food that the consumer wants to order.

For example, if a consumer orders '1 garlic chicken' from store A, but the order call is canceled, a store that sells 'garlic chicken' or 'whole chicken' among other stores located a certain distance from the customer's delivery address can be recommended. A store that sells the same food (ex garlic chicken) as the food ordered by the consumer may be recommended, but in some cases, a store that sells food (ex whole chicken) that is a higher level of the food ordered by the consumer may be recommended. .

However, the recommendation service may not be provided depending on specific cancellation reasons. For example, if it is determined that the customer who ordered is a black consumer, there is no need to recommend another store.

In addition, the recommendation order may be based on a recommendation score (C) calculated based on a preset method (Ax + By = C). In the above A/x + By = C, A and B correspond to constants, x is the distance from the customer's delivery address to the store, and y is the store's evaluation grade (evaluated by customers), so that the recommendation score C is that can be calculated.

As a result, the smaller the distance from the delivery address to the store, the higher the evaluation grade of the store, the higher the recommendation score C may be.

On the other hand, when the information included in the order form is in the pop-up state (for a preset time), if the order is confirmed, the POS terminal 500 prints the order form through an in-store printing device, and the information included in the order form can be delivered to the delivery agency device 600 (S250).

In addition, when the order is confirmed for a preset time, the POS terminal 500 may display a plurality of estimated delivery times and allow a store manager to select a specific delivery expected time. Among a plurality of estimated delivery times (eg 15 minutes, 20 minutes, 30 minutes, etc.), the store manager selects a specific delivery estimated time.

The POS terminal 500 may print an order book including the selected specific delivery expected time (ex 30 minutes) and transmit an order reception completion message including the specific delivery expected time to the consumer's terminal.

It can be assumed that a plurality of delivery agencies interlocked with the POS terminal 500 exist. The cloud ordering system 400 may deliver a list of delivery agencies in which deliverable manpower exists to the POS terminal 500 in a state in which information on delivery manpower is received in real time from each delivery agent.

The POS terminal 500 allows the person in charge of the store to select a specific delivery agency from the list of delivery agencies, select a specific delivery request time from among a plurality of delivery request times, and specify the order including the specific delivery request time. It can be delivered to the delivery agency device (600).

Here, the delivery request time may mean the time it takes to deliver the cooked food to the order address, and the above-mentioned expected delivery time may correspond to the sum of the time it takes to cook the food and the time it takes to deliver it. have.

Accordingly, a plurality of delivery request times (ex 15 minutes, 20 minutes, 25 minutes, 30 minutes, etc.) displayed on the screen of the POS terminal 500 will be less than or equal to each specific delivery expected time (ex 30 minutes). This is because the delivery time cannot be greater than the sum of the food preparation and delivery times.

In addition, the plurality of delivery request times will be greater than or equal to the time calculated in consideration of the delivery distance and the average delivery speed. Specifically, assuming that the distance from the store to the ordering address is 15 km and the average delivery speed (based on a scooter) is 60 km/h, the average time calculated as 15 km / 60 km/h will be 15 minutes. That is, the plurality of delivery request times may correspond to a value greater than or equal to 15 minutes.

The POS terminal 500 may display a plurality of delivery request times within the range (15 minutes to 30 minutes) calculated as above on the screen, and allow the person in charge of the store to select a specific delivery request time (ex 20 minutes). . In addition, the POS terminal 500 delivers an order form including a specific delivery request time (ex 20 minutes) to a specific delivery agency device 600, and the order menu is delivered to the order place within the specific delivery request time (20 minutes) You can request to be

6 is a diagram illustrating information that can be checked in an integrated management system according to an embodiment of the present invention.

A plurality of stores may use the order automation service of the present invention, and the plurality of stores may receive an order call using the cloud order system 400 . At this time, the cloud order system 400 will be accessible through an integrated management system or a web page used by the person in charge of each of the plurality of stores.

In the integrated management system, which is a control tower for managing the order automation service of the present invention, the number of a plurality of first order calls delivered to the cloud order system 400 using AI voice recognition for a preset time, a plurality of first order calls It is possible to check information about the total of the total payment amount for , the average call time for a plurality of first order calls, and the like.

For example, you can check information about 200 order calls delivered using AI voice recognition over the past week, total payment amount of 3 million won, and average call time of 50 seconds.

In addition, the integrated management system, the number of the plurality of second order calls delivered using the backup consultation server 300 during the preset time (ex the last week), the total sum of the total payment amount for the plurality of second order calls , it is also possible to check information about the average call time for a plurality of second order calls.

Additionally, the integrated management system is the number of a plurality of third order calls delivered using both the recognition module 110 and the backup consultation server 300, the total of the total payment amount for the plurality of third order calls, a plurality of 3 You can also check information on average talk time for order calls.

In addition, the total number of orders 1-1 to 3-1 order calls, the total payment amount for orders 1-1 to 3-1 order calls during this week (1st week period), 1-1 to Average call time for each of the 3-1 order calls in the past week (second week period), the number of order calls 1-2 to 3-2, and the number of order calls 1-2 to 3-2 A changed numerical value (ex %, number of cases) may appear compared to the total of the total payment amount and the average call time for the first 1-2 to 3-2 order calls.

According to one embodiment, the integrated management system uses the recognition module 110 or the backup consultation server 300 to the number of order calls accumulated up to the preset time (ex. one week), and the total sum of the accumulated payment amount. can also be checked.

In a plurality of stores, the cloud ordering system 400 may be accessed through a web page used by each person in charge. For reference, the log-in information (eg ID/password, etc.) of the web page and the aforementioned POS terminal 500 may be the same.

On the web page, it is possible to change the price and payment type for products sold by the order call in each store, and support to check the order time, recording history, etc. ordered by the order call. That is, the person in charge of each store can easily modify the price and payment type (packaging or delivery) of the sold product through the web page, and can check the previous details such as order time and recording history.

7 is a diagram showing a detailed configuration of a system for an order automation service according to an embodiment of the present invention.

First, when an order call is received from a terminal of a consumer (customer), the order call is Voice G/W through Server Based Computing (SBC), Session Initiation Protocol (SIP) server, etc. included in the call center infrastructure system 200 . (Gateway) can be reached. The Voice G/W may support AI voice recognition while transmitting and receiving information with the API Server, visible ARS, and STT/TTS/Hybrid engine included in the AI callbot system 100 .

The AI callbot system 100 may perform order menu recognition, order address recognition, etc. included in an order call through the AI engine in the API server, and may acquire affiliate store information, customer information, and the like. For reference, although the ARS and STT/TTS/Hybrid engines shown in FIG. 7 are displayed separately from the AI colbot system 100, they may correspond to configurations included in the AI colbot system 100, and the STT/TTS/ Hybrid engine can be included in a kind of AI engine. In addition, although not specifically disclosed in FIG. 7 , the AI engine may form a structure included in the recognition module 110 .

In addition, the AI callbot system 100 can use the road name address API included in the DB of the Ministry of Public Administration and Security during the order address recognition process, and can clearly support order menu recognition and order address recognition through the backup counselor service/solution 300 There, it is possible to transfer the recognized information and the like to the cloud ordering system (400).

In the cloud ordering system 400, order-related information (menu, address, etc.) is transferred to the POS store POS linkage-Agent 500 through the franchise server, and delivered to the delivery agency service linkage 600 to automatically place an order for a call order. service can be provided.

8 is a view showing the progress of the AI ordering service according to an embodiment of the present invention.

As can be seen in FIG. 8 , in the present invention, various results are derived according to the customer's intention (packaging, delivery, order cancellation, etc.). In addition, a process of requesting a menu question from the customer and recognizing the customer's vocalization by the recognition module 110 is divided into a plurality of cases and displayed.

For example, it is shown in Figure 8 that different processes are derived depending on the intention of the customer, such as menu recognition, similar menu name recognition, order cancellation, agent connection, menu not recognition, price inquiry, etc. can be checked through

The embodiments according to the present invention described above may be implemented in the form of program instructions that can be executed through various computer components and recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the computer-readable recording medium may be specially designed and configured for the present invention, or may be known and available to those skilled in the computer software field. Examples of the computer-readable recording medium include a hardware device specially configured to store and execute program instructions, such as a magnetic medium such as a hard disk. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules for carrying out the processing according to the present invention, and vice versa.

In the above, the present invention has been described with specific matters such as specific components and limited embodiments and drawings, but these are provided to help a more general understanding of the present invention, and the present invention is not limited to the above embodiments. , various modifications and variations can be devised from these descriptions by those of ordinary skill in the art to which the present invention pertains.

Accordingly, the spirit of the present invention should not be limited to the above-described embodiments, and not only the claims described below, but also all modifications equivalently or equivalently to the claims described below belong to the scope of the spirit of the present invention. will do it

100: AI callbot system
110: recognition module
200: call center infrastructure system
300: backup consulting server
400: cloud ordering system
500: POS terminal
600: delivery agency device

Claims (7)

A method for recognizing a delivery order, comprising:
In the state that the recognition module is included in the AI callbot system, and the recognition module for recognizing the delivery order is learned,
(a) the call center infrastructure system, receiving the order call from the consumer's terminal;
(b) the AI callbot system, receiving the order call from the call center infrastructure system, and using the recognition module to check the delivery order included in the order call;
(c) registering, by the AI callbot system, an order book including the delivery order in the cloud order system;
(d) the in-store POS terminal receiving the order form from the cloud order system, and displaying the information included in the order form on the screen for a preset time so that the in-store person in charge confirms the order; and
(e) if there is no order confirmation for the preset time, the order call is automatically canceled, and when the order is confirmed for the preset time, the POS terminal prints the order through an in-store printer transmitting the information included in the order form to the delivery agency device;
including,
In a state in which the first parameter and the second parameter are included in the recognition module,
Before step (a),
The AI callbot system is
The first character for the test representing the food menu is passed through the recognition module, the calculated result value is compared with the first character for the correct answer to derive a first difference value, and the first parameter is determined based on the first difference value update,
A second character for a test indicating a quantity is passed through the recognition module, the calculated result value is compared with a second character for a correct answer to derive a second difference value, and the second parameter is updated based on the second difference value and
The AI callbot system is
By passing the second character for the test representing a value of 4 or less through the recognition module, and comparing the calculated result value with the second character for the correct answer, the second difference value is derived, and based on the second difference value update the second parameter;
In a state in which the first parameter included in the recognition module is updated,
The AI callbot system is
A plurality of test 1-1 characters representing a specific food menu are passed through the recognition module to obtain a plurality of prediction 1-1 characters, and each coordinate according to the frequency of the plurality of prediction 1-1 characters The value is calculated, and the 1-1 character for a specific prediction having the highest frequency is set as the 1-1 character for a specific correct answer, and based on this, the 1-1 character for a specific prediction within a preset distance is set as the specific Judging by the 1-1 character for the correct answer,
When the order call is canceled, the AI call bot system selects a specific store that sells the same menu as the order menu or a menu corresponding to a higher concept among other stores located a certain distance from the delivery address of the consumer to the consumer's terminal If it is determined that the consumer is a black consumer, the specific store is not recommended to the consumer's terminal,
The recommended order is
C = A/x + By (however, A and B are constants, C is the recommended score, x is the distance from the delivery address to the store, y is the store’s evaluation grade)
Method, characterized in that calculated based on the recommendation score (C) determined by the above formula. delete delete delete According to claim 1,
In step (b),
When any one food menu can be classified into a detailed category, and the detailed category is called an option,
The AI callbot system uses the recognition module to recognize the food menu, the option, and the quantity based on the order expression included in the order call, respectively, to confirm the delivery order. According to claim 1,
In step (b),
The AI callbot system is
When it is determined that the delivery order is confirmed, the delivery order is transmitted to the consumer's terminal and verified, and based on this, the order book including the delivery order is registered in the cloud order system,
When it is determined that the delivery order is not confirmed, a visual order screen is displayed on the consumer's terminal, and the delivery order is obtained based on an input value from the consumer's terminal. In the system for recognizing a delivery order,
In the state that the recognition module is included in the AI callbot system, and the recognition module for recognizing the delivery order is learned,
a call center infrastructure system that receives an order call from a consumer's terminal;
The AI callbot system for receiving the order call from the call center infrastructure system, confirming the delivery order included in the order call using the recognition module, and registering the order book including the delivery order in the cloud order system; and
Receives the order form from the cloud order system, displays the information included in the order form on the screen for a preset time so that the store manager confirms the order, and if there is no order confirmation for the preset time, the order call an in-store POS terminal that automatically cancels, and when an order is confirmed during the preset time, prints the order form through an in-store printing device, and transmits information included in the order form to a delivery agency device;
including,
In a state in which the first parameter and the second parameter are included in the recognition module,
The AI callbot system is
The first character for the test representing the food menu is passed through the recognition module, the calculated result value is compared with the first character for the correct answer to derive a first difference value, and the first parameter is determined based on the first difference value update,
A second character for a test indicating a quantity is passed through the recognition module, the calculated result value is compared with a second character for a correct answer to derive a second difference value, and the second parameter is updated based on the second difference value and
The AI callbot system is
By passing the second character for the test representing a value of 4 or less through the recognition module, and comparing the calculated result value with the second character for the correct answer, the second difference value is derived, and based on the second difference value update the second parameter;
In a state in which the first parameter included in the recognition module is updated,
The AI callbot system is
A plurality of test 1-1 characters representing a specific food menu are passed through the recognition module to obtain a plurality of prediction 1-1 characters, and each coordinate according to the frequency of the plurality of prediction 1-1 characters The value is calculated, and the 1-1 character for a specific prediction having the highest frequency is set as the 1-1 character for a specific correct answer, and based on this, the 1-1 character for a specific prediction within a preset distance is set as the specific Judging by the 1-1 character for the correct answer,
When the order call is canceled, the AI call bot system selects a specific store that sells the same menu as the order menu or a menu corresponding to a higher concept among other stores located a certain distance from the delivery address of the consumer to the consumer's terminal If it is determined that the consumer is a black consumer, the specific store is not recommended to the consumer's terminal,
The recommended order is
C = A/x + By (however, A and B are constants, C is the recommended score, x is the distance from the delivery address to the store, y is the store’s evaluation grade)
System, characterized in that calculated based on the recommendation score (C) determined by the above formula.

Images