Advertisement content generation system

Abstract

To support creation of advertisement content with high advertisement effect.SOLUTION: An advertisement content generation system 30 has: presentation means 10 that presents questions about advertisement content to a user in an interactive format; and generation means 20 that acquires answers input by the user to the questions as text information, and generates advertisement content from the answers to the questions, using a learning model 21 that has learned the relation between the advertisement content and advertisement effectiveness of the advertisement content, and a generative AI server 3 that generates images based on the text information.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] This invention relates to a system capable of generating advertising content with high advertising effects.

Background Art

[0002] In an advertising distribution system, the advertising content to be distributed has a great impact on advertising effects such as the CTR (click-through rate) of the advertisement. For this reason, in the field of advertising distribution, technologies related to the replacement of advertising content have also been proposed (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, it is difficult for ordinary users without particular expertise to produce advertising content with high advertising effects.

[0005] This invention has been made in view of the above circumstances, and an object thereof is to assist in creating advertising content with high advertising effects.

Means for Solving the Problems

[0006] This invention provides an advertising content generation system having a presentation means for presenting questions regarding advertising content to a user in an interactive format, an acquisition means for acquiring, as text information, answers input by the user to the questions, a learning model that has learned the relationship between advertising content and the advertising effect of the advertising content, and a generation means for generating advertising content from the answers to the questions using an algorithm for generating advertising content based on the text information.

Advantages of the Invention

[0007] According to this invention, the creation of advertising content with a high advertising effect is supported.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3A

Figure 3B

Figure 3C

Figure 3D

Figure 4

Figure 5

Figure 6

Figure 7

Modes for Carrying Out the Invention

[0009] Hereinafter, embodiments of this invention will be described with reference to the drawings.

[0010] <First Embodiment> FIG. 1 is a block diagram showing the configuration of an advertisement distribution system 100 including an advertisement content generation system 30 according to a first embodiment of the present invention. This advertisement distribution system 100 includes an advertisement content generation system 30 and an advertisement distribution server 4. The advertisement distribution server 4 is, for example, a server that generates a web page having an advertisement placement frame. The advertisement distribution server 4 posts and distributes the advertisement content generated by the advertisement content generation system 30 in the advertisement placement frame of the web page. Here, the advertisement content may be any content that is distributed to an unspecified number of people regardless of its purpose and content. As an example of the advertisement content, there is one that includes an image (still image or moving image), but it may also consist substantially only of text information.

[0011] The advertisement content generation system 30 includes a user terminal 1, an image generation server 2, and a generation AI server 3. The user terminal 1 is a terminal that uses the advertisement content generation system 30 to generate advertisement content and requests the advertisement distribution server 4 to distribute an advertisement including this advertisement content. The user terminal 1 may be a smartphone or a personal computer. The user terminal 1 constitutes a presentation means 10 that presents questions regarding the advertisement content to the user in an interactive form.

[0012] The image generation server 2 and the generation AI server 3 acquire the answer input by the user to the question as text information, and use a learning model that has learned the relationship between the advertisement content and the advertising effect of the advertisement content, and an algorithm that generates an image based on the text information, to constitute a generation means 20 that generates an advertisement image from the answer to the question. Here, the algorithm is, for example, a so-called generation AI having a natural language processing function such as GPT, Midjouney, Bard, etc.

[0013] More specifically, the image generation server 2 includes a learning model 21 that has learned the relationship between the advertising content and the advertising effect (specifically, CTR) of the advertising content. This learning model 21 is obtained by learning the relationship between the advertising content distributed in the advertising distribution system 100 and its CTR. Further, the generation AI server 3 includes an algorithm for generating advertising content based on a prompt that is text information. The image generation server 2 in the generation means 20 uses the learning model 21 to determine feature information for improving the advertising effect of the advertising content corresponding to the answer to the question, and generates a prompt that is text information including the text information representing the feature information and the answer to the question, and provides it to the generation AI server 3. The user terminal 1 acquires the advertising content obtained from this generation AI server 3.

[0014] FIG. 2 is a sequence diagram showing an operation example of this embodiment. In this example, an application for executing a process of presenting a question sentence regarding advertising content to the user and acquiring an answer from the user, and a process of transmitting the answer to the image generation server 2 and acquiring advertising content from the generation AI server 3 is installed in the user terminal 1.

[0015] Therefore, in the operation example of FIG. 2, the user terminal 1 repeatedly performs a process of presenting a question to the user and receiving an answer from the user a plurality of times (steps S1q-1 and S1a-1 to S1q-n and S1a-n). Next, the user terminal 1 transmits the answers 1 to n received from the user to the image generation server 2 (step S2).

[0016] FIG. 3 shows an example of a screen when the above-described application is executed, in which a process of presenting questions to the user and receiving answers from the user is performed. In the example of this figure, sequential question posing and answer input are performed in an interactive format between a computer (an avatar called "AI Assistant", etc.) and the user, and the question content and the answer content are displayed in time series. The object IF is used for the user to input the answer content in text or voice. The objects D1, D3, D5 indicate the content of the speech (question sentences) by the AI assistant, and the objects D2, D5 indicate the answer content input by the user.

[0017] When the image generation server 2 receives these answers 1 to n, it extracts the feature information of the advertisement content estimated to have a high CTR using the learning model 21 (step S11). Next, the image generation server 2 generates a prompt including this feature information (step S12) and transmits the prompt to the generation AI server 3 (step S13). At the same time, as shown in FIG. 3B, a message indicating that the user terminal 1 is waiting because an image is being generated is displayed on the user terminal 1.

[0018] FIG. 3C illustrates the prompt 210 transmitted to this generation AI server 3. In this prompt 210, the feature information other than the feature information 211 is the feature information corresponding to the answers 1 to n. And the "ratio of the image and the character part is **~** and the overall tone is calm" of the feature information 211 is the advertisement content corresponding to the answers 1 to n and is the feature information of the advertisement content estimated to have a high CTR. In other words, the prompt 210 is obtained by adding a prompt (prompt 2) indicating the learning result obtained from the learning model 21 possessed by the image generation server to a prompt (prompt 1) generated based only on the answer content input by the user. Note that the prompt 210 does not necessarily need to be simply concatenated or inserted with the prompt 1 and the prompt 2. In short, it is sufficient that the knowledge (learning result) obtained from the learning model 21 is reflected in the prompt 210 in some form.

[0019] The generation AI server 3 generates a plurality of advertising contents according to this prompt 210 (step S20), and transmits the plurality of advertising contents to the user terminal 1 (step S21). FIG. 3D shows an example of a screen displayed on the user terminal 1 at this time. The generated images IM1, IM2, and IM3 are displayed, and the user is prompted to select one of them.

[0020] The user terminal 1 displays the plurality of advertising contents (step S31). The user selects one advertising content from the plurality of advertising contents (step S32). In this way, the user obtains advertising content with a high CTR. Note that only one advertising content acquired from the generation AI server 3 and maintained by the user may be used. In this case, the process of step S32 is omitted.

[0021] As described above, according to the present embodiment, the user can easily obtain advertising content with a high CTR.

[0022] <Second Embodiment> FIG. 4 is a sequence diagram showing an operation example of an advertising content generation system according to the second embodiment of the present invention. In the present embodiment, the processes of steps S11 to S13 and S21 in the first embodiment (FIG. 2) are replaced with the processes of steps S16, S22, S17, and S18. The other processes are the same as those in the first embodiment.

[0023] When the image generation server 2 receives an answer to a question from the user terminal 1, it generates a prompt for generating advertising content based on the answer (step S16), and transmits this prompt to the generation AI server 3 (step S13).

[0024] The generation AI server 3 generates a plurality of advertising contents according to this prompt (step S20), and transmits the plurality of advertising contents to the image generation server 2 (step S22).

[0025] The image generation server 2 uses the learning model 21 to identify one or more advertising contents estimated to have a high CTR among the plurality of received advertising contents (step S17), and transmits the identified advertising contents to the user terminal 1. The processing in steps S31 and S32 after that is the same as that in the first embodiment.

[0026] FIG. 5 illustrates a prompt 220 transmitted from the image generation server 2 to the generation AI server 3 in step S13. The prompt 210 (see FIG. 3) in the first embodiment included the feature information 211 of the advertising content that can obtain a high CTR. In contrast, the prompt 220 in this embodiment does not include what corresponds to this feature information 211. In this embodiment, the process for obtaining the advertising content that can obtain a high CTR is executed in step S17.

[0027] Also in this embodiment, the same effects as those in the first embodiment can be obtained.

[0028] <Other Embodiments> As described above, the first and second embodiments of the present invention have been described, but other embodiments are also conceivable for the present invention. For example, the processing in steps S1q-1 and S1a―1 to S1q-n and S1a-n in the first or second embodiment may be replaced with the processing shown in FIG. 6.

[0029] In the first embodiment, the user terminal 1 constituted the presentation means 10. In contrast, in this embodiment, the user terminal 1 and the image generation server 2 constitute the presentation means 10.

[0030] In this embodiment, when the answer 1 to the first question 1 is input to the user terminal 1 (step S1a-1), the user terminal 1 transmits the answer 1 to the image generation server 2 (step S2a-1). When the image generation server 2 receives the answer 1, it extracts the feature information necessary for generating the advertising content expected to obtain a high CTR using the learning model 21 (step S41).

[0031] Next, the image generation server 2 determines the next question (in this case, question 2) based on the extracted feature information (step S42), and transmits the next question (question 2) to the user terminal 1 (step S2q-2).

[0032] The user terminal 1 displays question 2 (step S1q-2), receives answer 2 in response thereto (step S1a-2), and transmits answer 2 to the image generation server 2 (step S2a-2).

[0033] The image generation server 2 determines whether or not the cumulative answer content (in this case, answer 1 + answer 2) satisfies the condition for expecting a CTR of a certain level or higher (step S43). In the case of this example, it is a determination as to whether or not a CTR of a certain level or higher can be obtained from the advertisement content having the feature information corresponding to answer 1 + answer 2.

[0034] If the determination result in step S43 is "NO", the image generation server 2 extracts the feature information necessary for generating the advertisement content expected to obtain a high CTR, determines the next question based on the extracted feature information, and repeats the processes after step S2q-2 described above.

[0035] On the other hand, if the determination result in step S43 is "YES", the image generation server 2 completes the hearing of sending questions and receiving answers (step S45), and proceeds to the process of step S16 in the first embodiment (FIG. 4), for example. The subsequent processes are the same as those in the first embodiment.

[0036] FIG. 7 is a diagram showing examples of questions and answers that occur in the present embodiment. In the present embodiment, depending on the content of the answer, the next question content (such as a question of the type that further delves into the answer or a question from another perspective) changes. In this example, in the promotion of sushi, since there is a learning result that the importance of including detailed information on the business format is the highest in increasing the CTR, a question that delves deeper into the previous question (a question regarding the business format) is generated. On the other hand, in the case of questions related to personnel recruitment, since there is a learning result that the content of the message is likely to affect the CTR, a question for determining the content of the message (recruitment content) is generated as the next question.

[0037] In short, in the information processing system according to the present invention, which is composed of one or more processors or devices, a step of presenting a question regarding advertising content to a user in an interactive format, obtaining the user's answer to the question as text information, using a learning model that has learned the relationship between the advertising content and the advertising effect of the advertising content, and an algorithm for generating advertising content based on the text information, a step of generating advertising content from the answer to the question may be executed. Here, when obtaining the user's answer, instead of the user performing character input or voice input using an input device as in the above-described embodiment, a plurality of options for answering the question may be displayed together with each question, and the user may obtain the answer by designating one option. In short, it is sufficient that the user's answer can ultimately be obtained as text information. According to the present embodiment, since the next question to be presented is determined based on the content of the answer obtained for one question sentence using a learning model, characteristic information for generating advertising content with a high CTR can be efficiently searched for.

Explanation of Reference Numerals

[0038] 100... Advertising distribution system, 30... Advertising content generation system, 4... Advertising distribution server, 1... User terminal, 2... Image generation server, 3... Generation AI server, 10... Presentation means, 20... Generation means.

Claims

1. Presentation means for presenting questions about advertising content to a user in an interactive format, Generation means for generating advertising content from the answer to the question using a learning model that has learned the relationship between advertising content and the advertising effect of the advertising content, and an algorithm for generating advertising content based on text information, by obtaining the user's answer to the question as text information An advertising content generation system having.

2. The generation means uses the learning model to determine feature information for improving the advertising effect of the advertising content corresponding to the answer, generates text information including the text information representing the feature information and the answer, and provides the generated text information to the algorithm to obtain advertising content from the algorithm. The advertising content generation system according to claim 1.

3. The generation means generates text information based on the answer to the question, provides the generated text information to the algorithm, obtains a plurality of advertising contents from the algorithm, and selects one or more advertising contents from the obtained plurality of advertising contents. The advertising content generation system according to claim 1.

4. The presentation means uses the learning model to determine the next question to be presented based on the content of the answer obtained for one question. The advertising content generation system according to claim 1.

5. On a computer, A step of presenting questions about advertising content to a user in an interactive format, A step of obtaining the answer input by the user to the question as text information, and generating advertising content from the answer to the question using a learning model that has learned the relationship between advertising content and the advertising effect of the advertising content, and an algorithm for generating advertising content based on text information A program for causing the computer to execute.

Images