Information processing system, information processing method, and computer program

Abstract

We provide a system that allows users to easily obtain advertising plans with superior advertising effectiveness. [Solution] An information processing method relating to one aspect of this disclosure includes predicting the advertising effect obtained by placing advertisements on multiple advertising media under given simulation conditions. The information processing method further includes having artificial intelligence generate response sentences to user input through a user interface, and outputting the response sentences generated by the artificial intelligence through the user interface, thereby realizing a conversation with a user using artificial intelligence. The information processing method further includes setting simulation conditions through a conversation with a user using artificial intelligence, and creating an advertising placement plan for multiple advertising media based on the advertising effect predicted under those conditions.

Description

Technical Field

[0001] The present disclosure relates to an information processing system and an information processing method.

Background Art

[0002] Conventionally, a technique for selecting an advertisement publishing destination from a plurality of advertisement media so as to maximize the advertisement effect with respect to the cost has been known. For example, a technique for selecting an advertisement publishing destination for advertisements published across a plurality of advertisement media so as to maximize the reach to a target is known (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the prior art, complicated operations and specialized knowledge were required for setting various parameters necessary to realize the selection of the publishing destination. Therefore, according to one aspect of the present disclosure, it is desirable to provide a system that enables a user to obtain an advertisement publishing plan with excellent advertisement effects by a simple operation.

Means for Solving the Problems

[0005] According to one aspect of the present disclosure, an information processing system is provided. The information processing system includes a prediction unit and a conversation unit. The prediction unit is configured to predict an advertisement effect obtained by advertisement publishing for a plurality of advertisement media under given simulation conditions.

[0006] The conversation unit is configured to enable conversation with the user using artificial intelligence by having artificial intelligence generate response sentences to user input through the user interface, and outputting the response sentences generated by the artificial intelligence through the user interface.

[0007] According to one aspect of this disclosure, the conversation unit is configured to set simulation conditions through conversation with a user using artificial intelligence, and to create an advertising placement plan for multiple advertising media based on the advertising effects predicted by the prediction unit under the simulation conditions.

[0008] According to one aspect of this disclosure, the conversational unit is further configured to output advertising placement plans through a user interface. With this configured information processing system, the user can set simulation conditions by directly or indirectly communicating their advertising placement preferences to the artificial intelligence through conversation with the AI. Therefore, the user can receive advertising placement plans with superior advertising effectiveness with simple operations.

[0009] According to one aspect of this disclosure, the conversation unit is configured to create an advertising placement plan, including budget allocation to multiple advertising media, based on the advertising effects predicted by the prediction unit under simulation conditions.

[0010] With an information processing system configured in this way, users can receive advertising plans that allow them to effectively utilize their limited budget and advertise through multiple advertising media.

[0011] According to one aspect of this disclosure, the conversation unit may be configured to set advertising budgets and advertising periods as simulation conditions through conversations with a user using artificial intelligence. According to another aspect of this disclosure, the conversation unit may further be configured to set the range of budget allocations for one or more advertising media out of a plurality of advertising media as simulation conditions through conversations with a user using artificial intelligence.

[0012] According to this information processing system, it is possible to provide users with advertising placement plans that reflect their wishes in detail.

[0013] According to one aspect of this disclosure, the information processing system may include a designation information acquisition unit that acquires designation information that specifies the personality of a virtual planner through a user interface. The conversation unit may be configured to set the personality of the artificial intelligence based on the designation information, thereby causing the artificial intelligence to function as a virtual planner having the personality specified by the user, and to set simulation conditions through conversation between the virtual planner and the user.

[0014] With this configured information processing system, users can receive a variety of advertising placement plans from a variety of virtual planners by specifying the personality of the virtual planner.

[0015] According to one aspect of this disclosure, the information processing system may further include a performance acquisition unit that acquires past performance information regarding advertising effectiveness. In this case, the prediction unit may be configured to construct a prediction model for advertising effectiveness based on the performance information and to predict advertising effectiveness based on the prediction model.

[0016] According to this information processing system, advertising effectiveness can be predicted with high accuracy based on past performance data. Therefore, users can receive advertising plans that are superior in terms of advertising effectiveness.

[0017] According to one aspect of this disclosure, the information processing system may further include a record acquisition unit that acquires placement records, which are records of past advertising placements. The conversation unit may be configured to provide the placement records to artificial intelligence, instruct it to detect deviations from past advertising placements regarding simulation conditions specified by the user in the conversation, and to output an alert regarding the simulation conditions through the user interface when the artificial intelligence detects a deviation.

[0018] With this type of information processing system, users can prevent unwanted advertising plans from being created and output due to undesirable simulation conditions resulting from misinterpretations or input errors, through alert-based corrections.

[0019] According to one aspect of this disclosure, the conversation unit may be configured to selectively output a portion of the conversation content specified by the user, in accordance with instructions from the user input through the user interface, by converting it into data in a predetermined format.

[0020] With this type of information processing system, users can review the content of their conversations with artificial intelligence at a later date and use that information to help them consider their advertising placement plans.

[0021] According to one aspect of this disclosure, the conversation unit may be configured to output an advertising placement plan through the user interface, and then, when a question about the advertising placement plan is input as user input, obtain an answer to the question from artificial intelligence as a response sentence and output it through the user interface.

[0022] With an information processing system configured in this way, users can gain a deeper understanding of the provided advertising placement plan. In other words, according to one aspect of this disclosure, it is possible to provide an information processing system that is highly convenient for users.

[0023] According to one aspect of this disclosure, the prediction unit may be configured to predict the advertising effect for each of the multiple advertising media, based on the cost invested through advertising on those media.

[0024] According to one aspect of the present disclosure, simulation conditions may include an advertising budget. The prediction unit may be configured to search for the investment budget for each of a plurality of advertising media such that the sum of the advertising effects obtained by advertising placement for the plurality of advertising media satisfies a predetermined condition under the constraint conditions including the advertising budget. The conversation unit may be configured to output an advertising placement plan indicating the investment budget for each of the searched advertising media.

[0025] According to the information processing system configured as described above, the user can receive the provision of an advertising placement plan that enables advertising placement for a plurality of advertising media with a budget allocation that is excellent in terms of advertising effects.

[0026] According to one aspect of the present disclosure, at least some of the plurality of advertising media may be advertising media capable of distributing advertisements in a plurality of distribution formats. In this case, the prediction unit may be configured to predict the advertising effect for each investment cost for each of the plurality of distribution formats with respect to at least some of the advertising media. The prediction unit may be configured to search for the investment budget for each distribution format as the investment budget for the corresponding advertising media for each of at least some of the advertising media.

[0027] According to the information processing system configured as described above, the user can receive the provision of an advertising placement plan that enables advertising placement for a plurality of advertising media and distribution formats with a budget allocation that is excellent in terms of advertising effects.

[0028] According to one aspect of the present disclosure, the conversation unit may be configured to set, as simulation conditions, the advertising budget and the range of the investment budget for each of at least one or more of the plurality of advertising media through a conversation with the user using artificial intelligence.

[0029] The prediction unit may be configured to search for the investment budget for each of the plurality of advertising media such that the sum of the advertising effects obtained by advertising placement for the plurality of advertising media satisfies a predetermined condition under the constraint conditions including the advertising budget and the range of the investment budget.

[0030] According to this information processing system, it is possible to provide advertising placement plans that reflect the user's wishes in detail.

[0031] According to one aspect of this disclosure, a computer program may be provided to cause a computer to function as a prediction unit and a conversation unit in the information processing system described above. The computer program may be recorded on a non-temporary computer-readable recording medium.

[0032] According to one aspect of this disclosure, an information processing method may be provided. The information processing method may be performed by a computer. The information processing method may include predicting the advertising effect obtained by placing advertisements on multiple advertising media under given simulation conditions.

[0033] The information processing method may include enabling conversation with a user using artificial intelligence by having artificial intelligence generate response sentences to user input through a user interface, and outputting the response sentences generated by the artificial intelligence through the user interface.

[0034] The information processing method may include setting simulation conditions through conversation with a user using artificial intelligence, and creating an advertising placement plan for multiple advertising media based on the advertising effect predicted under the simulation conditions. The information processing method may also include outputting the advertising placement plan through a user interface. According to one aspect of this disclosure, the information processing method has the same effect as the information processing system described above. [Brief explanation of the drawing]

[0035] [Figure 1] This is a block diagram representing the configuration of an information processing system. [Figure 2] This is a functional block diagram of an information processing system. [Figure 3] This diagram illustrates the structure of the service usage screen. [Figure 4]This is a flowchart illustrating the conversation-related processing performed by the processor. [Figure 5] This diagram illustrates the information provided to artificial intelligence in order to function as a media planner. [Figure 6] This figure shows an example of a conversation between a user and artificial intelligence (Part 1). [Figure 7] This is a flowchart representing the subprocesses executed by the processor. [Figure 8] This figure shows an example of a conversation between a user and artificial intelligence (part 2). [Figure 9] This is an explanatory diagram regarding the optimization process performed by the processor. [Figure 10] This diagram illustrates the configuration of a screen that displays the simulation results. [Figure 11] This diagram illustrates the detailed data from the simulation results. [Figure 12] This figure shows an example of a conversation between a user and artificial intelligence (part 3). [Figure 13] This is a flowchart illustrating the clipping-related processes performed by the processor. [Figure 14] Figure 14A is an example of the structure of the list screen related to clips, and Figure 14B is an example of the structure of clips. [Modes for carrying out the invention]

[0036] Exemplary embodiments of the present disclosure are described below with reference to the drawings. The information processing system 10 of this embodiment is configured to provide the user with an advertising placement plan (hereinafter referred to as "media plan") that includes suggestions regarding the budget allocated to multiple advertising media, using artificial intelligence (AI) 51. The information processing system 10 is configured to extract information from the user necessary to provide an appropriate media plan through natural conversation with the user using artificial intelligence 51.

[0037] As shown in Figure 1, the information processing system 10 is configured to communicate with an AI server 50 equipped with artificial intelligence 51. The information processing system 10 is further configured to communicate with a user's information terminal 70.

[0038] The information terminal 70 is, for example, a personal computer. The information processing system 10 is configured to communicate with multiple information terminals 70 that support multiple users. In Figure 1, only one information terminal 70 is shown as a simplified representation.

[0039] The information processing system 10 comprises a processor 11, memory 13, storage 15, and a communication interface 19. The communication interface 19 is configured to communicate with the AI ​​server 50 and the information terminal 70 via a communication network.

[0040] The processor 11 is configured to execute processes according to a computer program. The memory 13 functions as working memory when the processor 11 is executing processes. The memory 13 is, for example, RAM.

[0041] The storage 15 stores computer programs as well as various data used for processing performed by the processor 11. The storage 15 is, for example, a solid-state drive (SSD).

[0042] The storage 15 stores performance data D1. Performance data D1 includes performance data D11, which represents the performance of the corresponding advertisement, and performance data D12, which describes the advertising effect observed for the advertisement. Performance data D1 can be obtained, for example, from advertisers. Performance data D1 may be prepared for each advertiser for multiple advertisers.

[0043] The advertising performance data D11 is a record of past advertising placements, including information such as the placement ID, advertising period, advertising budget, advertising medium, and / or delivery menu for each advertisement. The placement ID is an ID used to identify the corresponding advertisement placement. The advertising period is the period during which the advertisement is delivered.

[0044] In this specification, "distribution" should be interpreted broadly to include distribution via communication networks and broadcasting via broadcasting stations. Hereafter, the advertising period will also be referred to as the campaign period.

[0045] An advertising medium is a medium through which advertisements are delivered. Examples of advertising mediums include e-commerce (EC) sites, search engines, news sites, and / or digital platforms that provide social networking services. Examples of advertising mediums also include broadcast media such as television and / or radio broadcasts. A delivery menu represents the method of delivering advertisements on the corresponding advertising medium. Examples of delivery methods include banner ads and / or video ads.

[0046] This section describes an example where an advertising medium can deliver advertisements using multiple distribution methods. An advertiser, or a business acted upon by an advertiser, can select their desired distribution method from among the multiple methods provided by the advertising medium and place their advertisement; in other words, they can request the advertising medium to handle the advertisement distribution. Advertising performance data D11 describes the selected distribution method as part of the distribution menu.

[0047] Performance data D12 includes historical performance information regarding advertising effectiveness, for each ad placement, including the ad placement ID and performance information describing the advertising effectiveness observed for the ad placement identified by the ad placement ID. Performance information includes, for example, cost, number of conversions (CV), and / or CPA (Cost Per Action). Cost corresponds to advertising expenses. CPA, also known as cost per customer acquisition, is defined by dividing cost by the number of conversions (CV).

[0048] In addition, storage 15 stores user-specific plan history data D2. Plan history data D2 describes the media plans created for the corresponding user. Plan history data D2 may also describe media plans for a fixed number of times in the immediate vicinity.

[0049] The processor 11 functions as a conversation unit 101, a condition setting unit 103, a simulator 105, and a model learning unit 107, as shown in Figure 2, by executing processes according to the computer program.

[0050] The conversation unit 101 is configured to receive user sentences, which are conversational sentences input by the user through the user interface 71 of the information terminal 70, and transfer them to the artificial intelligence 51, causing the artificial intelligence 51 to generate response sentences corresponding to the user sentences, and to retrieve the response sentences from the artificial intelligence 51. The user interface 71 includes a display for displaying information to the user and an input device for receiving input from the user.

[0051] The conversation unit 101 is configured to output the response sentence to the user via the user interface 71 by transmitting the response sentence generated by the artificial intelligence 51 to the information terminal 70.

[0052] In other words, the conversation unit 101 receives user messages from the information terminal 70 and sends back a response message generated using the artificial intelligence 51 to the information terminal 70, thereby realizing a conversation between the artificial intelligence 51 and the user.

[0053] Artificial intelligence 51, based on a pre-configured personality from the information processing system 10, functions as a media planner, extracting information necessary for creating a media plan from the user through conversation.

[0054] The condition setting unit 103 is configured to set simulation conditions in the simulator 105 based on a conversation between the user and the artificial intelligence 51.

[0055] The simulator 105 is configured to search for the optimal budget to allocate to multiple advertising media by predicting the expected advertising effect for advertising placements based on the simulation conditions set by the condition setting unit 103.

[0056] The model learning unit 107 is configured to learn the predictive model 109 provided to the simulator 105 based on the aforementioned performance data D1. The model learning unit 107 is configured to build and update the predictive model 109 for each combination of advertising media and distribution menu using machine learning. Hereinafter, each combination of advertising media and distribution menu will be referred to as the advertising destination.

[0057] The prediction model 109 is a mathematical model that calculates a predicted value for the advertising effect relative to the investment cost for a given advertising placement. The prediction model 109 may be constructed using a neural network, or it may be constructed by learning the design parameters of a predetermined function.

[0058] The simulator 105 is configured to predict the advertising effectiveness relative to the investment cost for each advertising platform using a pre-trained predictive model 109, and to explore the optimal budget allocation for multiple advertising platforms within the range that satisfies the simulation conditions, based on the advertising effectiveness for each platform.

[0059] To obtain a media plan using the information processing system 10 configured in this way, a user can access the information processing system 10 by operating an information terminal 70. Upon accessing the information processing system 10, the information terminal 70 undergoes user authentication and then receives the service usage screen G0 shown in Figure 3 from the information processing system 10.

[0060] The service usage screen G0 functions as a graphical user interface for receiving various operations on the media planning service provided by the information processing system 10. Figure 3 shows an exemplary configuration of the service usage screen G0.

[0061] The service usage screen G0 has an item selection area G1 and a main area G2 for displaying information about the item selected in the item selection area G1. The item selection area G1 has the following selectable items: "Home", "Upload", "Plan", "Clip", and "Result".

[0062] In Figure 3, the main area G2 displays a screen related to the item "Home". When the item "Home" is selected, the main area G2 displays an operation screen in which the user can select one of several AI planners through the user interface 71. The AI ​​planner is a virtual media planner played by the artificial intelligence 51, based on the personality settings assigned to the artificial intelligence 51.

[0063] When the item "Upload" is selected, the main area G2 displays an upload screen for the user to upload performance data D1 to the information processing system 10.

[0064] When the item "Plan" is selected, a chat screen with a user input field is displayed in the main area G2. The chat screen displays the interaction between the user and the artificial intelligence 51 in speech bubbles (see Figures 6, 8, and 12). The chat screen is used when the user desires assistance from the AI ​​planner. According to this embodiment, information necessary to provide the user with a media plan is extracted from the user through conversation between the user and the artificial intelligence 51 (AI planner) via the chat screen.

[0065] When the item "Clip" is selected, the main area G2 displays a clip of a portion of the conversation between the user and the artificial intelligence 51 in the chat screen, extracted by the user's actions (see Figures 14A and 14B).

[0066] When the "Results" option is selected, the main area G2 displays a results screen explaining the media plan created through a simulation based on information elicited from the user via the chat screen (see Figure 10). The media plan is an advertising plan that proposes an allocation budget for each advertising platform based on the advertising conditions derived from the information elicited from the user via the chat screen.

[0067] As an operation to obtain a media plan, the user can select the item "Home," and through the operation screen displayed in the main area G2, select the desired AI planner. Before or after this, the user can select the item "Upload" and provide performance data D1 to the information processing system 10 through the upload screen displayed in the main area G2.

[0068] When an AI planner is selected and performance data D1 is provided, and a new chat room is created via the chat screen, the processor 11 executes the conversation-related processing shown in Figure 4.

[0069] When the conversation-related processing shown in Figure 4 is initiated, a prompt related to initial settings is input to the artificial intelligence 51, which is the conversation partner of the user (S110). The prompt includes a description related to personality settings.

[0070] In S110, the processor 11 obtains the AI ​​planner information specified by the user in the item "Home" as the AI ​​planner specification information, and includes personality information for setting the personality of the specified AI planner in the prompt, and inputs the prompt to the artificial intelligence 51. Figure 5 shows the information provided to the artificial intelligence 51 for the initial setup of the artificial intelligence 51.

[0071] At this time, the artificial intelligence 51 can be provided with personality information, including the gender of the AI ​​planner, tone of voice, planning characteristics, and knowledge of advertising media. Furthermore, the artificial intelligence 51 can be instructed through prompts to set the parameter values ​​necessary for creating a media plan in conversation with the user, so as to act as a media planner, and to notify the user to run the simulator 105 as soon as the preparation for creating the media plan is complete.

[0072] Examples of planning characteristics include "planning freely without being bound by past performance," "steady planning from existing menus in accordance with advertising trends," and "focusing on making the most effective use of the budget to the fullest extent." For example, if "steady planning from existing menus in accordance with advertising trends" is selected, personality information such as "planning so that the CPA falls within the range of 0.8 to 1.2 times the average of past advertising" is provided to the artificial intelligence 51.

[0073] In addition, the processor 11 may include information about the parameters necessary for creating a media plan, as well as information about the types of optimization modes available to the simulator 105 (details described later), as media-related knowledge, in the above prompt.

[0074] Furthermore, processor 11 provides artificial intelligence 51 with performance data D1 so that the AI ​​planner can perform planning based on past advertising campaigns. As mentioned above, performance data D1 includes records of past advertising campaigns and records of the advertising effectiveness obtained therefrom. In addition, processor 11 can provide artificial intelligence 51 with plan history data D2, which explains past media plans created for the user.

[0075] In addition, the processor 11 includes alert instruction information in the prompt and inputs the prompt to the artificial intelligence 51. The alert instruction information is information that instructs the artificial intelligence 51 to detect deviations from past advertising performance regarding the advertising conditions specified by the user during the conversation between the user and the artificial intelligence 51, and to notify the information processing system 10 of such detections. For example, the processor 11 can be instructed to detect deviations from past performance regarding the advertising budget, target number of conversions, and target CPA.

[0076] In the subsequent S120, the processor 11 sends an initial message to the information terminal 70 as a signal to start the conversation, causing the information terminal 70 to display the initial message. The initial message is output to the chat screen that the information terminal 70 displays through the user interface 71.

[0077] For example, the processor 11 can instruct the artificial intelligence 51 to start a conversation inquiring about the campaign period with the user, obtain the initial message from the artificial intelligence 51, and display it on the information terminal 70 via a chat screen.

[0078] Subsequently, the processor 11 receives conversational text from the user, i.e., user text, input through the user interface 71 from the information terminal 70, and executes the process of inputting the user text to the artificial intelligence 51 (S130). Furthermore, the processor 11 obtains a response text to the user text from the artificial intelligence 51 and executes the process of outputting the response text to the chat screen through the user interface 71 (S140).

[0079] Figure 6 shows an example of a conversation. In the example shown in Figure 6, the initial message displayed on the chat screen is, "Please tell me the settings required for the media plan. First, please tell me the campaign start date."

[0080] In this example, in response to the initial message, the user inputs information specifying the start date of the campaign period, i.e., the start date of ad delivery. Processor 11 can input this user statement, which includes the information specifying the start date, to artificial intelligence 51 (S130).

[0081] The processor 11 then receives a message from the artificial intelligence 51 as a response, stating, "The start date is [Month] [Day]. Next, please tell me the end date of the campaign." The processor 11 then sends the received message to the information terminal 70, which can then display the corresponding message on the chat screen (S140).

[0082] When there is no change operation by the AI ​​planner (No in S150), the processor 11 repeatedly performs the following processes until it determines in S160 that the conversation has been terminated: inputting the user's sentence to the artificial intelligence 51 (S130) and displaying the response sentence from the artificial intelligence 51 through the user interface 71 (S140).

[0083] Furthermore, in S140, the processor 11 monitors the conversation between the user and the artificial intelligence 51 and performs subprocessing in accordance with the conversation, as shown in Figure 7.

[0084] For example, if the processor 11 receives a parameter value related to the creation of a media plan (Yes in S210), it updates the simulation conditions stored in memory 13 to hold the parameter value as a simulation condition (S220). Alternatively, the processor 11 receives notification from the artificial intelligence 51 that a parameter value has been set and updates the simulation conditions.

[0085] In this embodiment, it is necessary to set a campaign period in order to create a media plan. The parameter values ​​related to the creation of the media plan include a value that defines the campaign period.

[0086] As will be explained in more detail later, when the optimization mode is CV maximization mode, the parameter values ​​related to media plan creation include a value that defines the advertising budget as an optimization metric. When the optimization mode is cost minimization mode, the parameter values ​​related to media plan creation include a value that defines the target number of conversions as an optimization metric. When the optimization mode is CPA maintenance mode, the parameter values ​​related to media plan creation include a value that defines the target CPA as an optimization metric.

[0087] Furthermore, in this embodiment, the range of possible budget allocation can be arbitrarily set as a constraint for each advertising destination as a simulation condition. That is, budget constraints can be set individually for each of any one or more advertising destinations. When the budget allocation is set to zero, the corresponding combination of advertising media and distribution menu is excluded from the advertising destinations. The range of the budget allocation may be set by explicit specification by the user, or by the user's agreement to the AI ​​planner's suggestions.

[0088] When parameter values ​​related to the advertising budget, campaign period, or budget range are set, the processor 11 can store the set parameter values ​​as simulation conditions (S220).

[0089] In addition, in this embodiment, when creating a media plan, the user can select one type of optimization mode from among several types of optimization modes. When an optimization mode is set, the processor 11 determines that parameter values ​​related to the creation of the media plan have been set (Yes in S210) and can store the set optimization mode as a simulation condition (S220).

[0090] In the example shown in Figure 6, after obtaining information on the start and end dates of the campaign period, the conversation between the user and the artificial intelligence 51 progresses to a stage where the artificial intelligence 51 inquires about the optimization mode desired by the user. In this example, the user inquires with the artificial intelligence 51 through the chat screen about the differences between several selectable optimization modes, receives an explanation of each mode, and then selects "Maximize Conversions" as the optimization mode.

[0091] In this embodiment, the following three optimization modes are provided. • Maximize Conversion Mode: This optimization mode explores budget allocations for each advertising platform to maximize the total number of conversions across all platforms, while staying within the specified advertising budget. • Cost Minimization Mode: This optimization mode explores the optimal budget allocation for each advertising platform to minimize the advertising budget while still achieving the target number of conversions. • CPA Maintenance Mode: This mode explores the optimal budget allocation for each advertising platform to maximize the advertising budget while keeping the CPA below the target value (i.e., target CPA).

[0092] According to this embodiment, once an optimization mode is specified by the user, a conversation continues to set the parameter values ​​necessary for creating a media plan in the corresponding optimization mode according to the user's wishes. In the example shown in Figure 8, when the CV maximization mode is specified as the optimization mode, the artificial intelligence 51 continues the conversation to extract information about the advertising budget.

[0093] In the example shown in Figure 8, the user enters the wrong number of digits for the advertising budget in response to an inquiry about the advertising budget. In this embodiment, the artificial intelligence 51 detects the discrepancy between the advertising budget provided by the user and past advertising records, based on prior alert instructions, and issues an alert. The conditions for determining a discrepancy may be specified in advance to the artificial intelligence 51, or it may be left to the judgment of the artificial intelligence 51.

[0094] When the processor 11 receives an alert notification from the artificial intelligence 51, it alerts the user by outputting a message via the chat screen using a dedicated alert bubble, asking whether there are any input errors regarding the advertising conditions received from the artificial intelligence 51.

[0095] In S140, the processor 11 monitors the conversation between the user and the artificial intelligence 51. When it receives an alert notification from the artificial intelligence 51 (Yes in S230), as shown in Figure 7, it executes a process to output a message from the artificial intelligence 51 to the chat screen in the form of an alert, as described above (S240), inquiring whether there are any input errors regarding the posting conditions. The alert-specific speech bubble display is achieved, for example, by displaying the outline of the speech bubble in a different color (e.g., red) than when there is no alert.

[0096] In addition, when the processor 11 obtains all the parameter values ​​necessary for creating a media plan through conversation between the user and the artificial intelligence 51 and the conditions for executing the optimization process are met (Yes in S250), it executes the optimization process using the simulator 105 in accordance with the notification from the artificial intelligence 51 (S260).

[0097] In the optimization process, the processor 11 searches for the optimal media plan, i.e., the budget allocated to multiple advertising outlets, under the set simulation conditions. By executing the optimization process, the processor 11 functions as a simulator 105 and a model learning unit 107, as shown in Figure 9.

[0098] In the optimization process, the processor 11 obtains the latest performance data D1 by reading it from the storage 15, and learns a predictive model 109 for each advertising location based on this performance data D1. For advertising locations without a history of advertising, a standard predictive model not based on performance data D1 can be prepared as the predictive model 109 for the corresponding advertising location.

[0099] As described above, the predictive model 109 for a single advertising platform is a mathematical model that calculates a predicted value for the advertising effectiveness relative to the investment cost for that platform. Advertising effectiveness is, for example, the number of conversions. Figure 9 shows a conceptual diagram of the predictive model 109, illustrating the curve of change in advertising effectiveness relative to investment cost.

[0100] In the optimization process, the processor 11 can input the investment cost to these prediction models 109 and repeat the operation of obtaining advertising effectiveness from the prediction models 109 while adjusting the investment cost.

[0101] When the optimization mode is set to CV maximization, the processor 11 searches for the budget allocation for each advertising placement that maximizes the sum of the advertising effects of each placement, within the range of the specified advertising budget limit, by repeating this operation. In this way, the processor 11 searches for the optimal media plan. The sum of advertising effects is the sum of the predicted values ​​of advertising effects output by the prediction model 109 for multiple advertising placements, and corresponds to the predicted value of advertising effect obtained by advertising on multiple placements. By calculating the sum of the predicted values, the processor 11 can predict the overall advertising effect and search for the optimal media plan.

[0102] If the budget allocated to each advertising platform is defined as a constraint, the processor 11 can optimize the media plan within the limits that satisfy the constraint. As mentioned above, the constraint may be specified by the user or set by the AI ​​planner.

[0103] For example, a conservative AI planner can set constraints to prioritize allocating budget to advertising platforms where the advertiser has previously placed ads, and to avoid allocating budget to platforms where the advertiser has not. An aggressive AI planner can set constraints to allocate budget to advertising platforms where the advertiser has not previously placed ads. Furthermore, the AI ​​planner can also add constraints regarding upper and / or lower limits on CPA.

[0104] When the optimization mode is set to cost minimization mode, the processor 11 can find the optimal media plan by searching for the budget allocated to each advertising platform that minimizes the advertising budget while ensuring that the sum of the advertising effects of each platform does not fall below the target number of conversions.

[0105] When the optimization mode is set to CPA maintenance mode, the processor 11 can find the optimal media plan by searching for the budget allocation for each advertising platform that maximizes the advertising budget within a range that does not exceed the target CPA.

[0106] By performing the optimization process in this manner, the processor 11 can create an optimized media plan according to the optimization mode and the set advertising conditions. The advertising conditions referred to here are the simulation conditions other than the optimization mode, such as the advertising budget, campaign period, budget range for each advertising platform, target number of conversions, and target CPA.

[0107] Subsequently, the processor 11 provides the artificial intelligence 51 with information about the created media plan (S270). At this time, the processor 11 can instruct the artificial intelligence 51 via prompts to answer questions about the media plan based on the provided media plan information.

[0108] In S270, the processor 11 further outputs the media plan created above as an optimized result to the service usage screen G0 of the information terminal 70. Specifically, the media plan is output to the results screen corresponding to the "Result" item on the service usage screen G0 via the user interface 71.

[0109] Figure 10 shows an example of how the media plan is displayed on the results screen. In this example, the results screen displays bar graphs explaining the budget (cost) for each advertising platform for multiple platforms where the budget is not zero. In Figure 10, the multiple advertising platforms are depicted as Platforms M1, M2, M3, M4, M5, M6, and M7.

[0110] The vertical axis of the bar graph represents the budget (cost). In the upper right corner of the results screen, there are buttons B1, B2, and B3 that allow you to switch the vertical axis of the bar graph between "Cost," "CV," and "CPA." In other words, the results screen is configured to display the budget, estimated number of conversions, and estimated CPA for each advertising platform.

[0111] The results screen further displays the campaign period, optimization mode type, and other constraints. Examples of other constraints include advertising budget, budget range per ad placement, target number of conversions, and target CPA. In addition, a download button B4 is placed on the results screen.

[0112] The user can obtain detailed data regarding the simulation results and media plan in the form of a data file in a predetermined format by pressing the download button B4 through the user interface 71 of the information terminal 70.

[0113] When the download button B4 is pressed, the processor 11 sends detailed data to the information terminal 70, for example, in the form of a spreadsheet. Figure 11 conceptually shows an example of detailed data that the user can obtain through download.

[0114] The detailed data includes the simulation name, simulation conditions, simulation results, and detailed information for each advertising platform. The detailed data describes the simulation conditions, including the optimization mode, budget minimum, budget maximum, campaign start date, end date, and other constraints.

[0115] The detailed data describes the estimated budget, estimated number of conversions (estimated CV), and estimated cost per acquisition (CPA) for the optimal media plan as a simulation result. The estimated budget, estimated number of conversions, and estimated CPA are estimates of the advertising budget, number of conversions, and CPA for the entire advertising platform, assuming advertising is placed according to the media plan.

[0116] The detailed data, categorized by advertising platform, includes information on the proposed investment cost (i.e., investment budget) for each platform, as well as the expected number of conversions (CV) and cost per acquisition (CPA) for that investment cost.

[0117] After receiving such a media plan, users can select the "Plan" item on the service usage screen G0 and ask the artificial intelligence 51 questions about the created media plan.

[0118] In the example shown in Figure 12, the user asks, "Why has the budget been reduced for one of the advertising outlets in the created media plan?" The artificial intelligence 51 generates an answer to the question based on the information in the given media plan and its knowledge as a media planner. In S130, the processor 11 sends the question, and in S140, it retrieves the answer to the question from the artificial intelligence 51 and can output it to the chat screen of the information terminal 70.

[0119] In addition, as shown in the example in Figure 12, the user asks, "Can you tell me about recent CPA trends?" The artificial intelligence 51 generates an answer to the question based on the information of the given media plan, the performance data D1, and its knowledge as a media planner. The processor 11 can send the question in S130 and output the answer to the chat screen of the information terminal 70 in S140.

[0120] In this embodiment, the artificial intelligence 51 is given the personality of a media planner, and further, by providing it with performance data D1 and media plan information, the artificial intelligence 51 is made capable of answering various questions from users regarding advertising placement.

[0121] In addition, in this embodiment, a clip button B5 is attached to the speech bubble displaying the response text of the artificial intelligence 51 on the chat screen. The processor 11 repeatedly executes the clip-related processing shown in Figure 13, and when the clip button B5 is pressed through the user interface 71 (Yes in S310), the response text of the artificial intelligence 51 associated with the pressed clip button B5 is saved as a clip in the storage 15, along with the previous user text (S320). Here, "clip" means a snippet of the conversation, and specifically corresponds to an extraction of a part of the conversation between the user and the artificial intelligence 51.

[0122] In S320, the processor 11 not only extracts a portion of the conversation from the chat screen, but also organizes the conversation content into a specific format and outputs it to a file. That is, the processor 11 selectively converts a portion of the conversation content specified by the user into data in a predetermined format and outputs it to a file. The storage 15 stores the data file as a clip in a specific file format (e.g., presentation file format).

[0123] Users can view and download saved clips by selecting the "Clips" item on the service usage screen G0. Figure 14A shows the configuration of the list screen displayed in the main area G2 when the "Clips" item is selected.

[0124] When the item "Clip" is selected, the processor 11 reads the corresponding clips stored in the storage 15 and can display a list screen with thumbnails of the stored clips arranged in the main area G2 of the service usage screen G0 via the user interface 71.

[0125] Subsequently, the processor 11 accepts an operation on the list screen, and when a clip is selected, it can display the clip through the user interface 71, as shown in Figure 14B. In the example shown in Figure 14B, a clip is displayed in which the user's statement is written as the title, and the response from the artificial intelligence 51 is summarized and written in bullet points.

[0126] In addition, if the user and artificial intelligence 51 are having a conversation via the chat screen and the "Home" operation screen is opened, and a different AI planner than the currently set AI planner is selected, the processor 11 will perform a conversation handover process.

[0127] In other words, when processor 11 detects that a change operation has been made to the AI ​​planner during the conversation (Yes in S150), it executes a handover process that includes assigning the newly selected AI planner's personality to artificial intelligence 51, as shown in Figure 4 (S155). Here, similar to the process in S110, the newly selected AI planner's personality can be assigned to artificial intelligence 51 by sending a prompt with personality information to artificial intelligence 51.

[0128] At this point, the processor 11 prompts the artificial intelligence 51 with the necessary information so that the new AI planner can continue the conversation by inheriting information about the previous conversation history, simulation results, and simulation condition settings from the previous AI planner, thereby setting up the new AI planner in the artificial intelligence 51 (S155).

[0129] After setting up the new AI planner, processor 11 adds a notification message to the chat screen indicating the change in planner (S155). Subsequently, processor 11 repeatedly executes the processes from S130 to S160, as before, to enable a conversation between the user and the artificial intelligence 51.

[0130] According to the information processing system 1 of this embodiment described above, artificial intelligence 51 can be used to extract information necessary for creating a media plan from the user in a natural conversation, create a media plan, and provide it to the user. Therefore, according to this embodiment, a system can be provided that allows the user to obtain a media plan with excellent advertising effectiveness through simple operations.

[0131] In particular, in this embodiment, users can use different AI planners with different personalities and receive media plans from the AI ​​planners that are conservative or challenging, depending on the personality of the planner.

[0132] In addition, in this embodiment, a predictive model 109 for predicting the advertising effect relative to the investment cost is generated for each advertising platform using machine learning with past performance data D1. The processor 11 uses this predictive model 109 for each advertising platform to predict the number of conversions and / or CPA as the overall advertising effect across multiple advertising platforms when a budget is allocated to multiple platforms, and generates an optimal media plan. Therefore, according to this embodiment, accurate advertising effect predictions can be made, and an appropriate media plan can be generated as a result.

[0133] According to this embodiment, the user can specify a desired optimization mode from among several types of optimization modes, including a CV maximization mode, a cost minimization mode, and a CPA maintenance mode, and receive a media plan searched according to the specified optimization mode. Therefore, this information processing system 10 can provide media plans that meet the various needs of the user.

[0134] According to this embodiment, another advantage is that an alert can be issued to the user when the advertising conditions set deviate significantly from past advertising records due to a user input error. In addition, the clipping function of the information processing system 10 can improve user convenience when considering advertising placements, including the provided media plan.

[0135] [Other embodiments] This disclosure is not limited to the embodiments described above, and various other embodiments are possible. For example, the illustrated service usage screen G0 is merely an example and can take various forms. In the above embodiment, advertising effectiveness was predicted and the optimal media plan was searched for on a per-combination basis of advertising media and distribution menus, but the distribution menu may not be considered, and advertising effectiveness may be predicted and the optimal media plan searched for on a per-advertising media basis. Only a portion of the multiple advertising media may be advertising media that can deliver ads using multiple distribution methods.

[0136] In addition, while the above embodiment uses conversion rate (CV) and cost per acquisition (CPA) as indicators of advertising effectiveness, advertising effectiveness may be quantified using other indicators such as page views (PV). Advertising effectiveness can be evaluated using various indicators other than those exemplified.

[0137] The information processing system 10 described above may have a function to output a report that allows the user to compare multiple media plans that have been created. For example, the report may be structured with images explaining the media plans arranged in parallel, as shown in Figure 10. Alternatively, the report may be output in the form of a spreadsheet, in which case the spreadsheet may be structured with detailed data, as shown in Figure 11, listed item by item for each of the multiple media plans.

[0138] The function of one component in the above embodiment may be distributed among multiple components. The functions of multiple components may be integrated into one component. Some parts of the configuration of the above embodiment may be omitted. At least some parts of the configuration of the above embodiment may be added to or replaced by the configuration of other above embodiments. Any aspect of the technical concept specified by the wording of the claims constitutes an embodiment of the present disclosure.

[0139] [Technical Concept Disclosed in This Specified Specification] This specification can be understood to disclose the following technical concepts: [Item 1] A prediction unit configured to predict the advertising effect obtained by placing advertisements on multiple advertising media under given simulation conditions, A conversation unit is configured to enable conversation with a user using artificial intelligence by having artificial intelligence generate response sentences to user input through a user interface, and outputting the response sentences generated by the artificial intelligence through the user interface. Equipped with, The conversation unit sets the simulation conditions through conversation with the user using the artificial intelligence, the prediction unit creates an advertising placement plan for the multiple advertising media based on the advertising effect predicted under the simulation conditions, and outputs the advertising placement plan through the user interface. [Item 2] The information processing system described in item 1, The conversation unit is an information processing system that creates an advertising placement plan, including budget allocation to the multiple advertising media, based on the advertising effect predicted by the prediction unit under the simulation conditions. [Item 3] An information processing system as described in item 1 or item 2, The conversation unit is an information processing system that, through conversation with the user using the artificial intelligence, sets the advertising budget, advertising period, and the range of the budget allocated to one or more of the multiple advertising media as the simulation conditions. [Item 4] An information processing system described in any one of items 1 to 3, The specified information acquisition unit acquires specified information that specifies the personality of the virtual planner through the aforementioned user interface. Furthermore, The conversation unit is an information processing system that sets the personality of the artificial intelligence based on the specified information, thereby causing the artificial intelligence to function as a virtual planner having the personality specified by the user, and sets the simulation conditions through a conversation between the virtual planner and the user. [Item 5] An information processing system described in any one of items 1 to 4, Performance Acquisition Unit for acquiring past performance information regarding the aforementioned advertising effectiveness Furthermore, The prediction unit is an information processing system configured to construct a prediction model for the advertising effect based on the performance information and to predict the advertising effect based on the prediction model. [Item 6] An information processing system described in any one of items 1 to 5, Records acquisition department that acquires advertising records, which are records of past advertising placements. Furthermore, The aforementioned conversation section is, The artificial intelligence is provided with the advertising record and instructed to detect deviations from past advertising placements regarding the simulation conditions specified by the user in the conversation. When the artificial intelligence detects a discrepancy, it outputs an alert regarding the simulation conditions through the user interface. An information processing system configured in such a way. [Item 7] An information processing system described in any one of items 1 to 6, The conversation unit is an information processing system configured to convert a portion of the conversation content specified by the user, from the conversation content with the user, into data in a predetermined format and output it, in accordance with instructions from the user input through the user interface. [Item 8] An information processing system described in any one of items 1 to 7, The conversation unit is configured to output the advertising placement plan through the user interface, and then, when a question regarding the advertising placement plan is input as user input, to obtain the answer to the question as a response sentence from the artificial intelligence and output it through the user interface. [Item 9] The information processing system described in item 1, The aforementioned simulation conditions include an advertising budget. The prediction unit, For each of the aforementioned multiple advertising media, predict the advertising effect for each investment cost obtained from placing advertisements on the advertising media. Under the constraints including the aforementioned advertising budget, the system searches for the budget allocated to each of the aforementioned advertising media such that the sum of the advertising effects obtained from advertising on each of the aforementioned advertising media satisfies predetermined conditions. It is configured in such a way, The aforementioned conversation unit is an information processing system that outputs an advertising placement plan showing the budget allocated to each of the searched advertising media. [Item 10] The information processing system described in item 9, At least some of the aforementioned multiple advertising media are advertising media capable of delivering advertisements in multiple delivery formats, The prediction unit, With respect to at least some of the aforementioned advertising media, predict the advertising effect for each of the aforementioned multiple distribution formats based on the aforementioned investment cost. With respect to at least some of the aforementioned advertising media, for each advertising media, the budget allocated to the corresponding advertising media is searched for based on the distribution format. An information processing system configured in such a way. [Item 11] The information processing system described in item 9, The conversation unit is configured to set the advertising budget and the range of the investment budget for at least one of the multiple advertising media as simulation conditions through conversation with the user using the artificial intelligence. The prediction unit is an information processing system configured to search for an investment budget for each of the multiple advertising media such that the sum of the advertising effects obtained by placing advertisements on each of the multiple advertising media satisfies predetermined conditions, under constraints including the range of the advertising budget and the investment budget. [Item 12] A computer program for causing a computer to function as the prediction unit and the conversation unit in the information processing system described in any one of items 1 to 11. [Item 13] A method of information processing performed by a computer, Under given simulation conditions, predict the advertising effect obtained by placing advertisements on multiple advertising media, The system enables conversation with a user using artificial intelligence by having artificial intelligence generate response sentences to user input through a user interface, and outputting the response sentences generated by the artificial intelligence through the user interface. The process involves setting the simulation conditions through a conversation with the user using the artificial intelligence, and creating an advertising plan for the multiple advertising media based on the predicted advertising effect under the simulation conditions by making the prediction. The aforementioned advertising placement plan is output through the aforementioned user interface, Information processing methods including [Explanation of Symbols]

[0140] 10...Information processing system, 11...Processor, 13...Memory, 15...Storage, 19...Communication interface, 50...AI server, 51...Artificial intelligence, 70...Information terminal, 71...User interface, 101...Conversation unit, 103...Condition setting unit, 105...Simulator, 107...Model learning unit, 109...Predictive model, G0...Service usage screen, G1...Item selection area, G2...Main area.

Claims

1. A prediction unit configured to predict the advertising effect obtained by placing advertisements on multiple advertising media under given simulation conditions, A conversation unit is configured to enable conversation with a user using artificial intelligence by having artificial intelligence generate response sentences to user input through a user interface, and outputting the response sentences generated by the artificial intelligence through the user interface. Equipped with, The conversation unit sets the simulation conditions through conversation with the user using the artificial intelligence, the prediction unit creates an advertising placement plan for the multiple advertising media based on the advertising effect predicted under the simulation conditions, and outputs the advertising placement plan through the user interface.

2. The information processing system according to claim 1, The conversation unit is an information processing system that creates an advertising placement plan, including budget allocation to the multiple advertising media, based on the advertising effect predicted by the prediction unit under the simulation conditions.

3. The information processing system according to claim 1, The conversation unit is an information processing system that, through conversation with the user using the artificial intelligence, sets the advertising budget, advertising period, and the range of the budget allocated to one or more of the multiple advertising media as the simulation conditions.

4. The information processing system according to claim 1, The specified information acquisition unit acquires specified information that specifies the personality of the virtual planner through the aforementioned user interface. Furthermore, The conversation unit is an information processing system that sets the personality of the artificial intelligence based on the specified information, thereby causing the artificial intelligence to function as a virtual planner having the personality specified by the user, and sets the simulation conditions through a conversation between the virtual planner and the user.

5. The information processing system according to claim 1, Performance Acquisition Unit for acquiring past performance information regarding the aforementioned advertising effectiveness Furthermore, The prediction unit is an information processing system configured to construct a prediction model for the advertising effect based on the performance information and to predict the advertising effect based on the prediction model.

6. The information processing system according to claim 1, Records acquisition department that acquires advertising records, which are records of past advertising placements. Furthermore, The aforementioned conversation section is, The artificial intelligence is provided with the advertising record and instructed to detect deviations from past advertising placements regarding the simulation conditions specified by the user in the conversation. When the artificial intelligence detects a discrepancy, it outputs an alert regarding the simulation conditions through the user interface. An information processing system configured in such a way.

7. The information processing system according to claim 1, The conversation unit is an information processing system configured to selectively output, in accordance with instructions from the user input through the user interface, a portion of the conversation content with the user that is specified by the user, by converting it into data in a predetermined format.

8. The information processing system according to claim 1, The conversation unit is configured to output the advertising placement plan through the user interface, and then, when a question regarding the advertising placement plan is input as user input, to obtain the answer to the question as a response sentence from the artificial intelligence and output it through the user interface.

9. The information processing system according to claim 1, The aforementioned simulation conditions include an advertising budget. The prediction unit, For each of the aforementioned multiple advertising media, predict the advertising effect for each investment cost obtained from placing advertisements on the advertising media. Under the constraints including the aforementioned advertising budget, the system searches for the budget allocated to each of the aforementioned advertising media such that the sum of the advertising effects obtained from advertising on each of the aforementioned advertising media satisfies predetermined conditions. It is configured in such a way, The aforementioned conversation unit is an information processing system that outputs an advertising placement plan showing the budget allocated to each of the searched advertising media.

10. The information processing system according to claim 9, At least some of the aforementioned multiple advertising media are advertising media capable of delivering advertisements in multiple delivery formats, The prediction unit, With respect to at least some of the aforementioned advertising media, predict the advertising effect for each of the aforementioned multiple distribution formats based on the aforementioned investment cost. With respect to at least some of the aforementioned advertising media, for each advertising media, the budget allocated to the corresponding advertising media is searched for based on the distribution format. An information processing system configured in such a way.

11. The information processing system according to claim 9, The conversation unit is configured to set the advertising budget and the range of the investment budget for at least one of the multiple advertising media as simulation conditions through conversation with the user using the artificial intelligence. The prediction unit is an information processing system configured to search for an investment budget for each of the multiple advertising media such that the sum of the advertising effects obtained by placing advertisements on each of the multiple advertising media satisfies predetermined conditions, under constraints including the range of the advertising budget and the investment budget.

12. A computer program for causing a computer to function as the prediction unit and the conversation unit in the information processing system according to any one of claims 1 to 11.

13. A method of information processing performed by a computer, Under given simulation conditions, predict the advertising effect obtained by placing advertisements on multiple advertising media, The system enables conversation with a user using artificial intelligence by having artificial intelligence generate response sentences to user input through a user interface, and outputting the response sentences generated by the artificial intelligence through the user interface. The process involves setting the simulation conditions through a conversation with the user using the artificial intelligence, and creating an advertising plan for the multiple advertising media based on the predicted advertising effect under the simulation conditions by making the prediction. The aforementioned advertising placement plan is output through the aforementioned user interface, Information processing methods including

Images