System, method of execution by computer, and computer program

Abstract

Technologies are needed to streamline the analysis and utilization of open-ended responses. [Solution] The disclosed system includes a processor that performs a process of acquiring survey results, which are the results of a survey conducted on survey subjects within an organization and answered in a free-response format, and analyzing the survey results to assign one or more classifications to the survey results. The survey results include one or more sentences, and assigning the classifications may include assigning the classifications to each sentence.

Description

[[Technical Field]]

[0001] The present disclosure relates to a system, a computer-implemented method, and a computer program. [[Background Art]]

[0002] Conventionally, there is known a technique for evaluating employee engagement by conducting a survey of employees and analyzing the results. [[Prior Art Documents]] [[Patent Documents]]

[0003] [[Patent Document 1]] Japanese Unexamined Patent Application Publication No. 2022-2010 [[Summary of Invention]]

[0004] In cases such as when it is desired to collect a wide range of problem awareness from organizational members such as employees, it is conceivable to use free-form description answers instead of multiple-choice answers. However, free-form description answers have the disadvantage that, compared with multiple-choice answers, they require more time and effort to classify, analyze, and utilize the answer results.

[0005] Therefore, there is a need for a technique that improves the efficiency of analyzing and utilizing free-form description answers.

[0006] The disclosed technology may comprise a processor that executes processing including: acquiring survey results that are results of a survey conducted on survey subjects in an organization with answers provided in free-form description, and analyzing the survey results to assign one or more classifications to the survey results.

[0007] One aspect of disclosure, a computer-based method, may include obtaining survey results, which are the results of a survey conducted on survey subjects within an organization and answered in an open-ended format, and analyzing the survey results to assign one or more classifications to the survey results.

[0008] Another aspect of this disclosure, a computer program, can be used to cause a computer to perform operations that include obtaining survey results, which are the results of a survey conducted on survey subjects within an organization and answered in an open-ended format, and analyzing the survey results to assign one or more classifications to the survey results.

[0009] Further details will be described in the embodiments below. [Brief explanation of the drawing]

[0010] [Figure 1] Figure 1 shows a network configuration diagram including the survey results analysis system according to the embodiment. [Figure 2] Figure 2 is a flowchart illustrating the survey result analysis process performed by the system. [Figure 3] Figure 3 is a flowchart illustrating the problem and solution extraction process performed by the system. [Figure 4] Figure 4 is a flowchart illustrating the survey result display process performed by the system. [Figure 5] Figure 5 shows an example of a screen displaying a list of survey results. [Figure 6] Figure 6 shows an example of a filtering settings screen. [Figure 7] Figure 7 shows an example of a screen displaying the positive / negative judgment results for each sentence. [Figure 8] Figure 8 shows an example of a display screen for the problem and solution generation results. [Figure 9] Figure 9 shows an example of a screen displaying the evidence answers that form the basis of the generated task. [Modes for carrying out the invention]

[0011] <1. Overview of the system, how it is executed by the computer, and computer programs>

[0012] (1) The system according to the embodiment may include a processor that performs a process of acquiring survey results, which are the results of a survey conducted on survey subjects in an organization and answered in an open-ended format, and analyzing the survey results to assign one or more classifications to the survey results. In this case, the survey results can be efficiently reviewed even if they are in an open-ended format. Furthermore, the usability of the survey results can be improved.

[0013] (2) The survey results include one or more sentences, and assigning the classification may include assigning the classification to each sentence. In this case, the accuracy of the analysis of the survey results can be improved.

[0014] (3) Assigning classifications to the survey results may include assigning classifications using artificial intelligence from among several pre-prepared classifications. In this case, the survey results can be customized by the user, and the user can obtain the analysis results they want.

[0015] (4) Assigning classifications to the survey results may include providing the artificial intelligence with pre-defined conditions for assigning the classifications, and the artificial intelligence assigning the classifications based on those conditions. In this case, the accuracy of the analysis of the survey results can be improved.

[0016] (5) The process may further include outputting the assigned classification and the survey results in association. In this case, the visibility of the survey results can be improved and verification can be made easier.

[0017] (6) The processing may further comprise performing positive-negative determination to determine whether the survey result includes a positive opinion or a negative opinion. In this case, the content of the survey result can be easily grasped.

[0018] (7) The survey result includes one or more sentences, and performing the positive-negative determination may include performing positive-negative determination for each of the sentences. In this case, the content of the survey result can be easily grasped.

[0019] (8) The positive-negative determination may include calculating a positive-negative score obtained by digitizing a result of the positive-negative determination. In this case, the content of the survey result can be easily grasped.

[0020] (9) The positive-negative determination is performed by artificial intelligence, and the positive-negative determination may include acquiring a reason for calculation of the positive-negative score by the artificial intelligence. In this case, the accuracy of the survey result can be easily confirmed.

[0021] (10) The processing may further comprise displaying the survey results in a visually distinguishable manner based on a result of the positive-negative determination. In this case, an analysis result of the survey result can be easily grasped.

[0022] (11) The processing may further comprise determining whether notification is necessary based on a result of the positive-negative determination, and notifying information related to the survey result when it is determined that the notification is necessary. In this case, a necessary response corresponding to the survey result can be quickly grasped and performed.

[0023] (12) The positive-negative determination may further include determining that the survey result includes a neutral opinion.

[0024] (13) The processing may further comprise performing browsing settings for the survey result according to the classification assigned to the survey result. In this case, it becomes easy to present necessary information to a person who needs the survey result.

[0025] (14) The process may further comprise generating issues for the organization based on the survey results using artificial intelligence. In this case, issues that can be extracted from the survey results can be efficiently identified.

[0026] (15) The issues of the organization may be generated for each of the classifications. In this case, the issues that can be extracted from the survey results can be efficiently identified.

[0027] (16) The process may further include at least one of the following: presenting solutions to the organization's challenges based on the survey results, and generating and presenting solutions to the organization's challenges using artificial intelligence without relying on the survey results. In this case, solutions that can be extracted from the survey results and solutions that are not limited by the survey results can be efficiently identified.

[0028] (17) The process may further include at least one of the following: presenting the survey results that formed the basis for generating the organization's issues; and presenting the survey results that formed the basis for solutions to the organization's issues based on the survey results.

[0029] (18) A computer-based method according to the embodiment may include obtaining survey results, which are the results of a survey conducted on survey subjects in an organization and answered in a free-response format, and assigning one or more classifications to the survey results by analyzing them.

[0030] (19) The program according to the embodiment can cause a computer to perform the following operations: acquire survey results, which are the results of a survey conducted on survey subjects in an organization and answered in a free-response format; and analyze the survey results to assign one or more classifications to the survey results.

[0031] A computer program according to the embodiment may be configured to cause a computer to execute the method described above. Furthermore, a computer program according to the embodiment may cause a computer to function as the system described above. The computer program may be recorded on a computer-readable, non-temporary recording medium.

[0032] <2. Examples of systems, methods of execution by computers, and computer programs>

[0033] The embodiments will be described in more detail below with reference to the drawings.

[0034] In this embodiment, the system of this disclosure will be described using an example in which it is performed on the results of a survey conducted on employees of a company. A company is an example of an organization, and an organization may include not only companies but also groups, teams, educational and research institutions, etc., that are composed of people with a common purpose or intention. Employees are an example of survey subjects in a company, and survey subjects in a company may include all people involved with the company, such as temporary workers, part-time workers, prospective employees, interns, trainees, and apprentices.

[0035] Figure 1 shows a network configuration diagram including the survey results analysis system 1 according to the embodiment.

[0036] Survey Results Analysis System 1 (hereinafter simply referred to as "System 1") obtains the results of surveys conducted by companies on their employees as respondents as survey results. The survey may be of any type and may include all surveys that companies (organizations) conduct on their employees (members), such as employee satisfaction surveys, engagement surveys, motivation surveys, organizational satisfaction surveys, census surveys, and questionnaires for various training programs.

[0037] Furthermore, surveys can be conducted at any time or frequency, and may include various types, such as pulse surveys conducted frequently, surveys conducted regularly (monthly, quarterly, or yearly), surveys conducted after events, and surveys conducted upon retirement. In the following explanation, responses obtained from individual employees will be referred to as "responses," and the group of these responses will be referred to as "survey results."

[0038] The survey (research, questionnaire) is conducted in an open-ended format. The advantage of an open-ended survey is that employees and other respondents are not restricted by pre-defined options and can freely express their opinions and feelings in their own words. Compared to multiple-choice surveys, open-ended surveys are more effective in uncovering genuine feelings and underlying psychology, and in gathering a broader range of awareness of issues. The results of an open-ended survey can be obtained, for example, as text data representing the respondents' answers in written form.

[0039] However, with open-ended surveys, it's impossible to determine whether an individual is experiencing negative or positive emotions without reviewing the content. Furthermore, compared to multiple-choice surveys, it's difficult to quantitatively evaluate the results, making it challenging to identify trends and problems within the survey. This is especially true for companies with a large number of employees, where the sheer volume of responses makes up the survey results, requiring considerable effort and time to understand.

[0040] Therefore, the system 1 according to this embodiment can be used to perform predetermined processing on the survey results and output them as information that companies can use to efficiently utilize those survey results.

[0041] System 1 can be implemented by one or more computers. System 1 may be cloud-based, such as SaaS (Software as a Service), or on-premise. System 1 comprises a processor 10, a storage device 20 connected to the processor 10, artificial intelligence 30 (AI), and a communication unit 40.

[0042] Processor 10 is, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or another type of processor.

[0043] The storage device 20 is, for example, a primary storage device and a secondary storage device. The primary storage device is, for example, RAM (Random Access Memory). The secondary storage device is, for example, an HDD (Hard Disk Drive) or an SSD (Solid State Drive). The storage device 20 includes a computer program 21 and a database 22 (DB22).

[0044] Computer program 21 is a set of instructions that cause the processor 10 to execute various processes in order to operate the computer as system 1 according to the embodiment. The processor 10 reads and executes computer program 21 stored in storage device 20. As a result, the processor 10 executes the survey result analysis process 11 (Figure 2), the problem / solution extraction process 12 (Figure 3), and the survey result display process 13 (Figure 4), which will be described in detail later, in system 1.

[0045] DB22 records survey results 23, classification information 24, viewing settings information 25, alert condition information 26, and issue / solution generation results 27.

[0046] The survey results 23 are the results of a survey conducted among employees of a company, with responses in an open-ended format. Each response constituting the survey results 23 can be transmitted from an employee's terminal 5 and acquired via a communication unit 40. Here, the survey questions may be set by, for example, the survey administrator's terminal 6, which contains the necessary information for conducting the survey, such as the question content, timing of implementation, and target respondents, and then transmitted to each employee's terminal 5. The function of conducting such a survey may be realized by the processor 10 executing a computer program 21 stored in the storage device 20 of system 1, or it may be realized by a different computer than system 1. In other words, as long as system 1 can acquire the survey results 23, there are no particular restrictions on the method of conducting the survey.

[0047] Survey results 23 are information that associates necessary information, such as information about the employee as a respondent (name, department, employee-specific ID (identification) number, etc.) and the date and time of response, with the answers provided in a free-response format.

[0048] Classification information 24 is information provided and referenced by the processor 10 to the classification model 31 when a classification is assigned according to the content of each response in the survey results 23. Classification information 24 is entered and edited as appropriate on, for example, the surveyor's terminal 6 and stored in the database 22. Classification information 24 may include the classification to be assigned and the conditions for assigning the classification.

[0049] The classification is based on the theme to which the response belongs, such as "attendance," "work," "interpersonal relationships," "work environment," or "work-life balance." In the following explanation, the classification will simply use abstract names such as "Classification A," "Classification B," and "Classification C." Classification information 24 can be prepared in advance, for example, by accepting settings from the survey operator's terminal 6.

[0050] The conditions for assigning classifications are pre-prepared information to improve the classification accuracy of the classification model 31. These conditions include, for example, examples of words that may be included in responses to which each classification should be assigned, and definitions of each classification. The conditions may also include examples where no classification is assigned to reduce noise. Furthermore, the conditions may include a thesaurus and examples of variations in spelling.

[0051] The viewing settings information 25 is information about the persons who can view the survey results 23 stored in DB 22. The viewing settings information 25 can be prepared in advance, for example, by receiving settings from the survey person's terminal 6. The viewing settings information 25 is information necessary to set the persons who can view each response according to the classification assigned to each response. Specifically, the viewing settings information 25 is information that specifies that responses assigned to a specific classification (for example, "work-life balance") can be viewed by specific authorized persons (for example, the person in charge of analyzing the survey results 23) in a specific department (for example, the Human Resources Department). Alternatively, the viewing settings information 25 may be information that specifies the persons who can view each response according to the respondent. Specifically, the viewing settings information 25 may be information that specifies that a supervisor can view the responses of their subordinate employees.

[0052] Alert condition information 26 is information that defines the conditions for determining whether an alert notification is necessary as a result of the execution of the survey result analysis process 11. Alert condition information 26 can be prepared in advance, for example, by accepting settings from the survey person's terminal 6. Alert condition information 26 is information used to determine whether the result of the positive / negative judgment obtained in the survey result analysis process 11, which will be described later, is a result that requires sending an alert. Alert condition information 26 defines, for example, the threshold of the positive / negative score (details will be described later) that requires an alert notification, and the content of the response that requires an alert notification (for example, content related to harassment). In addition, in alert condition information 26, the recipient of the alert notification (for example, the survey person, the harassment committee, the respondent's supervisor, etc.) is associated with each condition. Alert condition information 26 may also be information that associates a classification with the recipient of an alert that will be notified when a response with the relevant classification is obtained. In this case, the response for which an alert is notified may be all responses that have been assigned the relevant classification, or it may be a response that satisfies the specified conditions.

[0053] The problem / solution generation result 27 is information about the problems and solutions generated by the problem / solution extraction model 33 in the problem / solution extraction process 12 described later (details will be provided later).

[0054] AI30 is, for example, a generative AI. AI30 may include a classification model 31, a positive / negative judgment model 32, and a problem / solution extraction model 33. The classification model 31 may be configured, for example, by giving the generative AI a prompt instructing it to classify the survey results 23. The positive / negative judgment model 32 may be configured, for example, by giving the generative AI a prompt instructing it to determine whether the survey results 23 are positive or negative. The classification model 31 and the positive / negative judgment model 32 are used in the survey result analysis process 11 described later when assigning classifications to the survey results 23 and when performing positive / negative judgments. The problem / solution extraction model 33 may be configured, for example, by giving the generative AI a prompt instructing it to extract organizational problems based on the survey results 23 and generate solutions for those problems. The problem / solution extraction model 33 is used in the problem / solution extraction process 12 when extracting and generating organizational problems based on the survey results 23, and when extracting and generating solutions for those problems.

[0055] AI30 may be implemented by a computer program 21 stored in the memory device 20 and a processor 10 that executes it, or it may be an AI owned by an external system (a system other than system 1) that is connected via the network 4 and available. If it is an AI owned by an external system, the processor 10 connects to AI30 via the network 4, gives prompts, and obtains the corresponding output.

[0056] The communication unit 40 transmits data to external devices, including terminals 5 and 6, and receives data from external devices. In other words, system 1 is connected to the surveyor's terminal 6 and the employee's terminal 5 via network 4, enabling them to communicate with each other.

[0057] Terminals 5 and 6 are, for example, personal computers, tablets, and smartphones. Terminals 5 and 6 are equipped with display devices and input devices. The display device is a display that can show display data such as input screens for free-response survey questions and their answers provided by System 1, and input screens for various setting information stored in DB22. The input device is a keyboard, mouse, touch panel, etc., that receives user operations and transmits signals based on those user operations to System 1.

[0058] Next, the details of the processing performed in System 1 in this embodiment will be described. First, System 1 executes the processing when it obtains the survey results 23 as a group of responses from the employee's terminal 5 and when it becomes possible to execute the survey result analysis processing 11. As mentioned above, the survey questions can be anything as long as they are in a free-response format. For example, the survey questions could be: "Please write anything you think freely," "Please write down any issues or areas for improvement you feel about your team," or "Please write down the strengths of your team." System 1 stores each acquired response in DB 22 as the survey results 23.

[0059] Figure 2 is a flowchart illustrating the survey result analysis process 11 performed by System 1. This survey result analysis process 11 may be performed as needed, for example, when a response is received from Terminal 5, or it may be performed periodically (for example, at midnight every day, every Saturday, etc.) on the accumulated survey results 23.

[0060] In step S1, the processor 10 of system 1 assigns one or more classifications to the obtained survey results 23 using the classification assignment model 31. The processor 10 instructs the classification assignment model 31 to assign a classification to each sentence (sentence) of the survey results 23. The processor 10 provides the classification assignment model 31 with classification information 24 along with the survey results 23. The processor 10 sends a prompt to the classification assignment model 31, for example, "Assign the specified classification to each sentence in the survey results 23." As a result, the classification assignment model 31 assigns one or more appropriate classifications to each sentence from among the predefined classifications ("Classification A", "Classification B", "Classification C", etc.) based on the classification information 24. Furthermore, the classification assignment accuracy may be improved by the classification assignment model 31 assigning classifications based on conditions for assigning classifications.

[0061] The processor 10 retrieves the classification assigned by the classification assignment model 31, associates the classification with each sentence of the response, and stores it as the survey result 23.

[0062] In step S2, the processor 10 performs a positive / negative judgment using the negative / positive judgment model 32. The negative / positive judgment determines whether the survey results 23 contain positive or negative opinions. The negative / positive judgment may also include a neutral judgment where the opinion is neither negative nor positive. Furthermore, the negative / positive judgment may also include calculating a negative / positive score, which is a numerical representation of the result of the negative / positive judgment, indicating the degree of negativity and positivity. In addition, the negative / positive judgment may also include obtaining the reason for calculating the negative / positive score. The processor 10 instructs the negative / positive judgment model 32 to perform these negative / positive judgments for each sentence (each response) in the survey results 23.

[0063] The processor 10 obtains the results determined by the positive / negative sentiment determination model 32 and stores the result of the positive / negative sentiment determination, the positive / negative sentiment score, and the reason for its calculation for each response sentence as the survey result 23.

[0064] Furthermore, in step S3, the processor 10 instructs the positivity judgment model 32 to calculate a positivity score for each response as a whole (all sentences). The positivity score calculated for each response as a whole (overall positivity score) may be, for example, the sum of the positivity scores assigned to each sentence, or it may be calculated by weighting the positivity scores assigned to each sentence under predetermined conditions, or it may be newly calculated for the entire sentence without using the positivity scores for each sentence.

[0065] The processor 10 obtains the results determined by the negative / positive judgment model 32, associates the overall negative / positive score with each response, and stores it as the survey results 23.

[0066] In step S4, processor 10 configures the viewing settings for the survey results 23. For example, processor 10 configures the viewing settings for each response in the survey results 23 according to the classification assigned to each response and the attributes of the respondent. Processor 10 refers to the viewing setting information 25 in DB 22 and configures the viewing settings. The free-response format survey results 23 may contain personal information or sensitive content. Therefore, configuring viewing settings has the advantage of increasing employees' sense of security regarding their responses and making it easier to obtain frank answers. In addition, configuring viewing settings according to the assigned classification has the advantage of making it easier to present the necessary information in the survey results 23 to those who need it.

[0067] In step S5, the processor 10 refers to the assigned classification, the positive / negative judgment result, and the alert condition information 26 to determine whether or not to issue an alert. The processor 10 may use either the positive / negative score or the overall positive / negative score as the positive / negative judgment result to make the determination.

[0068] In step S6, if the processor 10 determines that an alert is necessary based on the result of step S5 (YES in step S6), in step S7, it notifies the target notification terminals 5 and 6 of the alert. The processor 10 refers to the alert condition information 26 and determines the recipients of the alert. For example, if there is an employee whose negative / positive score is lower than the threshold, the processor 10 notifies the survey person or the respondent's supervisor as an alert, along with the response (sentence) from the survey result 23. As another example, if the processor 10 obtains a response that may constitute a harassment case, it notifies the harassment response department, such as a harassment committee, as an alert, along with the response from the survey result 23. As yet another example, if the processor 10 obtains a negative response related to attendance, it notifies the labor relations department, such as a labor relations department, as an alert, along with the response from the survey result 23. As another example, when processor 10 receives a negative response regarding evaluation or reward, it notifies the relevant department, such as the department responsible for making evaluation and reward decisions, as an alert, along with the survey results 23.

[0069] Furthermore, the processor 10 may refer to the alert condition information 26 and notify the associated notification recipient as an alert, along with the survey result 23, that it has obtained a response with the relevant classification.

[0070] The processor 10 may also send a notification that does not include the response in question, but instead states that it has obtained the response that warrants an alert and prompts the user to confirm that response. In this case, the processor 10 will also apply the viewing settings performed in step S4 to the alert recipient, and the recipient may become a viewer.

[0071] In step S4, the processor 10 made the determination of whether to configure the viewing settings and whether to notify alerts in step S5, by referring to the viewing setting information 25 and alert condition information 26 in DB22. However, the processor 10 may also have the AI ​​30 determine who can view the information and configure the viewing settings, or it may have the AI ​​30 determine whether to notify alerts and who to notify alerts, and then notify alerts, depending on the survey results 23 and the results obtained from the analysis process therefor (classification, positive / negative judgment results).

[0072] After step S7, and if it is determined that no alert is needed (NO in step S6), the processor 10 terminates processing.

[0073] Next, we will explain the issue / solution extraction process 12, which extracts issues and solutions from the survey results 23. Figure 3 is a flowchart illustrating the issue / solution extraction process 12 executed by System 1. This issue / solution extraction process 12 is executed, for example, immediately after the completion of the survey result analysis process 11 (Figure 2).

[0074] In step S11, the processor 10 generates issues for the company (organization) based on the survey results 23 obtained by the issue / solution extraction model 33. The processor 10 instructs the issue / solution extraction model 33 to generate issues for each classification assigned to the survey results 23. For example, the processor 10 sends a prompt to the issue / solution extraction model 33 such as, "Please extract issues from the survey results 23 that are assigned 'Category A'." The processor 10 may also instruct the model 33 to generate issue titles and detailed issue descriptions as issues. Furthermore, the processor 10 may also instruct the model 33 to generate relative evaluations of the issue's frequency and severity in relation to other issues.

[0075] For example, the title of the generated task might be "Fatigue and Stress Due to Long Working Hours," and the detailed content of the generated task might be "Average monthly overtime exceeds 40 hours, resulting in high levels of employee fatigue."

[0076] The processor 10 stores the issues generated from the issue / solution extraction model 33 in the DB 22 as issue / solution generation results 27. At this time, the processor 10 also obtains information from the issue / solution extraction model 33 regarding the survey results 23 that formed the basis for generating the issues, and stores the survey results 23 and the generated issues in association.

[0077] In step S12, the processor 10 generates solutions based on the survey results 23 for the issues obtained in step S11, using the issue / solution extraction model 33. The processor 10 may, for example, instruct the model 33 to generate solutions from the survey results 23 that have been assigned the same classification as the obtained issues. The processor 10 may send a prompt to the issue / solution extraction model 33, for example, "Please extract solutions corresponding to the generated issues from the survey results that have been assigned the category 'Classification A'." The processor 10 may also instruct the model 33 to generate a title for the solution and a detailed description of the solution. There is no particular limit to the number of solutions that can be generated; there may be one or more.

[0078] Furthermore, if the responses obtained as survey results 23 include solutions, the solutions based on survey results 23 may be obtained by directly extracting them from the responses, rather than being generated by the problem / solution extraction model 33. Extracting solutions included in the responses may be done, for example, by the surveyor via the surveyor's terminal 6, or the problem / solution extraction model 33 may be used to directly extract the relevant parts.

[0079] For example, the title of the solution based on the generated survey results 23 is "Expansion of the flextime system," and the content of the generated detailed solution is "Shorten core time and enable more flexible working hours."

[0080] The processor 10 stores the solutions generated from the problem / solution extraction model 33 in the DB 22 as problem / solution generation results 27. At this time, the processor 10 also obtains information about the survey results 23 that formed the basis for generating the solutions from the problem / solution extraction model 33, and stores the survey results 23 and the generated solutions in association.

[0081] In step S13, the processor 10 generates solutions to the issues obtained in step S11 that are not based on the survey results 23, using the issue / solution extraction model 33. Here, while solutions "based" on the survey results 23 are generated in step S12, solutions "not based" on the survey results 23 are generated in step S13. This is because, compared to extracting solutions based solely on the survey results 23, it is expected that solutions based on new perspectives that transcend the boundaries and culture of the company will be generated.

[0082] In step S13, the problem / solution extraction model 33 is not entirely excluded from referring to the survey results 23. For example, it is permitted to generate solutions based on other information that the problem / solution extraction model 33 can access while referring to the survey results 23. In contrast, in step S12, it is preferable that the problem / solution extraction model 33 generates solutions by referring only to the survey results 23.

[0083] The processor 10 sends a prompt to the problem / solution extraction model 33, for example, "Please propose solutions to the generated problems." The processor 10 may also instruct the model to generate a solution title and a detailed description of the solution. There is no particular limit to the number of solutions that can be generated; it may be one or more.

[0084] The processor 10 stores the solutions generated from the problem / solution extraction model 33 in DB 22 as problem / solution generation results 27.

[0085] Next, we will describe the survey result display process 13, which displays the survey results 23 and the issue / solution generation results 27 stored in DB22. Figure 4 is a flowchart illustrating the survey result display process 13 executed by System 1. This survey result display process 13 is executed after the survey result analysis process 11 (Figure 2) and the issue / solution extraction process 12 (Figure 3) are completed, and is requested from the terminals 5 and 6 of the survey person or the person who can view the results (hereinafter sometimes simply referred to as "viewer").

[0086] In step S21, the processor 10 determines whether it has received an instruction from terminals 5 and 6 to display the survey results 23. If the processor 10 determines that it has not received an instruction to display the survey results 23 (NO in step S21), it waits until it receives an instruction. On the other hand, if the processor 10 determines that it has received an instruction to display the survey results 23 (YES in step S21), in step S22, it displays the required answers included in the survey results 23 on terminals 5 and 6. In other words, the processor 10 displays the survey results 23 to be viewed according to the person who instructed it to display the survey results 23.

[0087] Here, Figure 5 shows an example of a screen 50 displaying a list of survey results 23. The processor 10 displays the respondent's (employee's) name field 52, the overall positive / negative score field 53, and the classification field 54 associated with each response, corresponding to the response field 51 which displays the responses obtained from each employee.

[0088] On the list display screen 50, the displayed answers can be filtered according to the entered conditions. For example, the answers displayed in the answer field 51 can be filtered by the classification displayed in the classification field 54. For example, by pressing the filtering button 55, the filtering settings screen 60 shown in Figure 6 is displayed on terminals 5 and 6. The filtering settings screen 60 can be displayed overlaid on the list display screen 50, for example, as a pop-up window.

[0089] The filtering settings screen 60 has a list of categories 62 that can be selected using checkboxes 61, and a search window 63 for searching for the desired category. Viewers can set the filtering by selecting the category assigned to the answer they want to view via the checkboxes 61. When the OK button 65 or the Cancel button 66 is pressed, the filtering settings screen 60 is closed.

[0090] Furthermore, the list display screen 50 allows for sorting of the displayed responses. For example, the sorting item field 57 allows sorting by overall positive / negative score, either from lowest to highest or from highest to lowest. In this way, the list display screen 50 allows for filtering by classification and sorting by overall positive / negative score, thereby improving the viewability of the survey results 23 for viewers and enabling efficient verification of even free-response answers.

[0091] Furthermore, the processor 10 visually distinguishes and displays the survey results 23 based on the positive / negative judgment. For example, as shown in Figure 5, in the answer field 51, the sentences included in the answers are visually distinguished according to the positive / negative judgment result.

[0092] For example, processor 10 will underline sentences that it determines to be positive opinions with a single underline 58 or display them in green. Also, processor 10 will underline sentences that it determines to be negative opinions with a double wavy line 59 or display them in red.

[0093] Furthermore, processor 10 displays in green the classification in classification column 54 assigned to sentences judged to be positive opinions, and the overall negative / positive score in overall negative / positive score column 53 assigned to sentences judged to be more positive than neutral. Processor 10 also displays in red the classification in classification column 54 assigned to sentences judged to be negative opinions, and the overall negative / positive score in overall negative / positive score column 53 assigned to sentences judged to be more negative than neutral.

[0094] Furthermore, the processor 10 displays sentences, classifications, and overall positive / negative scores that are determined to be neutral in a different manner from sentences, classifications, and overall positive / negative scores that are determined to be positive or negative, for example, in black or without an underline.

[0095] The displayed survey results 23 are shown in a selectable format for each sentence. The processor 10 is configured to accept instructions from the viewer to display the positive / negative sentiment score and the reason for calculating the score, which are the positive / negative sentiment judgment results for the selected sentence.

[0096] In step S23, the processor 10 determines whether it has received an instruction to display the positive / negative judgment result for each sentence. If the processor 10 determines that it has received an instruction to display the positive / negative judgment result (YES in step S23), in step S24, it displays the positive / negative score and the reason for calculating the positive / negative score for each sentence on terminals 5 and 6.

[0097] Here, Figure 7 shows an example of a screen 70 displaying the positive / negative judgment results for each sentence. The processor 10 displays an answer field 71 that displays the selected sentence (answer) and a classification field 72 that displays the classification assigned to that sentence on the positive / negative judgment result display screen 70. The processor 10 also displays a score field 73 that shows the positive / negative score calculated for that sentence and a reason field 74 that displays the reason for calculating the positive / negative score. The positive / negative judgment result display screen 70 can be displayed overlaid on the list display screen 50, for example, as a pop-up window.

[0098] For example, if "To be honest, I've been doing a lot of routine work lately and don't feel very fulfilled" is selected in the list display screen 50 of Figure 5, the negative / positive score for each sentence, calculated in step S2 of the survey result analysis process 11 (Figure 2), is read from the survey results 23 in DB 22. The negative / positive score is displayed in the score column 73, along with a numerical value on an evaluation axis that shows the degree of negativity and positivity, for example. In addition, the reason for calculating the negative / positive score obtained in the same step S2 is read from the survey results 23. The reason is displayed in the reason column 74 as, "It is thought that the employee is experiencing stress and dissatisfaction because they have been doing a lot of routine work lately and do not feel fulfilled."

[0099] After step S24, or if it is determined that no instruction to display the positive / negative judgment result has been received (NO in step S23), in step S25, the processor 10 determines whether or not it has received an instruction to terminate the display of the survey results on the list display screen 50 and the positive / negative judgment result display screen 70. An instruction to terminate the display of the display screens 50 and 70 is received, for example, by pressing a button on the display screens 50 and 70 that is provided to input an instruction to close the screen.

[0100] If the processor 10 determines that it has not received an instruction to end the display (NO in step S25), in step S26 it determines whether it has received an instruction to present the problem / solution generation result 27 by displaying it. If the processor 10 determines that it has not received an instruction to display the problem / solution generation result 27 (NO in step S26), it returns to step S25 and waits until it receives an instruction to end the display of the list display screen 50 or the negative / positive judgment result display screen 70, or until it receives an instruction to display the problem / solution generation result 27.

[0101] If the processor 10 determines that it has received an instruction to display the problem / solution generation result 27 (YES in step S26), in step S27, it presents the problem / solution generation result by displaying it on terminals 5 and 6.

[0102] Here, Figure 8 shows an example of the display screen 80 for the problem / solution generation result 27.

[0103] On the issue / solution display screen 80, for example, when receiving an instruction to display the issue / solution generation results 27, the system may also accept a selection of the category in which to display the issues / solutions. For example, if "Category G" is selected, as shown in Figure 8, the processor 10 reads the issues / solutions generated for "Category G" from the issue / solution generation results 27 in DB 22 and displays them in the issue column 81. The processor 10 also reads the generated solutions from the issue / solution generation results 27 and displays them in the solutions columns 82 and 83. Furthermore, if the issue / solution extraction process 12 also generates relative evaluations of the frequency and severity of issues in relation to other issues, that information may be displayed in the issue column 81.

[0104] For example, in the task section 81, the title of the task generated for "Category G" is displayed as "Fatigue and Stress Due to Long Working Hours," and the detailed task description is "Average monthly overtime exceeds 40 hours, resulting in high levels of employee fatigue."

[0105] In the Solution section 82, solutions to the problem based on the survey results 23 are displayed. In the Solution section 82, the title of the solution, "Expansion of the Flex-Time System," and the detailed content of the solution, "Shorten core time to enable more flexible working hours," are displayed. If there are multiple solutions (for example, three) based on the survey results 23, the Solution section 82 will be created and displayed according to the number of solutions. In the Solution section 83, solutions not based on the survey results 23 are displayed in a similar manner. In this case, in order to distinguish between solutions generated based on the survey results 23 and solutions generated not based on the survey results 23, it is preferable that each Solution section 82 be labeled "Solution from an employee" and the Solution section 83 be labeled "Suggestion from AI."

[0106] In this case, it is preferable that the processor 10 allows the viewer to confirm the answers that formed the basis for the displayed issues and solutions from the survey results 23. For example, the processor 10 places buttons 85 in the issue column 81 and each solution column 82 and 83 of the issue / solution display screen 80 to receive instructions to display the answers that formed the basis. Note that the solutions displayed in the solution column 83 may be solutions generated by the issue / solution extraction model 33 that are not based solely on the survey results 23. Therefore, if the solutions displayed in the solution column 83 are generated not based on the survey results 23, the buttons 85 do not need to be placed.

[0107] In step S28, the processor 10 determines whether it has received an instruction to display the supporting answer on the problem / solution display screen 80. If the processor 10 determines that it has received an instruction to display the supporting answer (YES in step S28), in step S29, it presents the supporting answer for the problem or solution for which it received an instruction to display.

[0108] Here, Figure 9 shows an example of a screen 90 displaying the evidence answers that form the basis of the generated issues. The processor 10 reads the answers that form the basis of the generated issues from the survey results 23 in DB22 and displays the corresponding answers. The evidence answer display screen 90 can be displayed overlaid on the issue / solution display screen 80, for example, as a pop-up window. The answers are displayed as they are, for example, "I'm always working overtime and it's nothing but painful" or "I've been working long hours every day recently, so I'm not getting any rest." If there are multiple corresponding answers, the processor 10 displays all of them.

[0109] Similarly, if the processor receives an instruction to display the answers that underlie the generated solution, it reads the answers that underlie the generated solution from the survey results 23 in DB22 and displays the corresponding answers.

[0110] Furthermore, the answers that formed the basis for the problems and solutions may not only be displayed on the evidence answer display screen 90 after receiving the instruction to display them, but may also be displayed on the problem / solution display screen 80 together with the problem / solution generation results 27.

[0111] After step S29, and if it is determined in step S28 that the instruction to display the supporting answer has not been received (NO in step S28), in step S30, the processor 10 determines whether or not it has received an instruction to terminate the display of the problem / solution on the problem / solution display screen 80 and the supporting answer display screen 90. An instruction to terminate the display of the display screens 80 and 90 is received, for example, by pressing a button on the display screens 80 and 90 that is provided to input an instruction to close the screen. If the processor 10 determines that it has not received an instruction to terminate the display (NO in step S30), it waits until it receives an instruction to terminate the display of the problem / solution display screen 80 and the supporting answer display screen 90. On the other hand, if the processor 10 determines that it has received an instruction to terminate the display (YES in step S25, NO in step S30), the processor 10 terminates processing.

[0112] The present invention is not limited to the above embodiments, and various modifications are possible.

[0113] For example, although the classification and positive / negative sentiment determination were explained using an example where they are performed on a sentence-by-sentence basis, they may also be applied to the entire response (all sentences). Furthermore, solutions based on the survey results and solutions not based on the survey results may be generated and displayed, or at least one of them may be displayed. In addition, the classification model 31 and the positive / negative sentiment determination model 32 may be implemented using a text analysis algorithm that does not use AI 30. [Explanation of Symbols]

[0114] 1: Survey results analysis system (system) 4: Network 5: Terminal 6: Terminal 10: Processor 11: Survey result analysis processing 12: Problem / Solution Extraction Process 13: Survey result display processing 20: Storage device 21: Computer Programs 22: Database (DB) 23: Survey Results 24:Classification information 25: Browsing settings information 26: Alert Condition Information 27: Issue / solution generation results 30:Artificial intelligence (AI) 31: Classification Model 32: Negative / Positive Determination Model 33: Problem / Solution Extraction Model 40: Communication Unit 50: List display screen 51:Answer column 52: Name field 53: Overall Negative / Positive Score Column 54:Classification field 55: Filtering button 57: Sort item column 58 :Single underline 59: Double wavy line 60: Filtering settings screen 61: Checkbox 62: List 63: Search box 65: OK button 66: Cancel button 70: Negative / positive judgment result display screen 71:Answer column 72:Classification field 73: Score column 74:Reason field 80: Issue / solution display screen 81: Assignment column 82: Solution section 83: Solution section 85: Button 90: Screen displaying the basis for the answer

Claims

1. We obtain survey results, which are the results of a survey conducted among survey participants within an organization, with responses in an open-ended format. By analyzing the survey results, one or more classifications are assigned to the survey results. Artificial intelligence is used to generate the organization's challenges based on the survey results, The system generates and presents, using artificial intelligence, at least one of the solutions to the organization's challenges based on the survey results, and at least one of the solutions to the organization's challenges not based on the survey results. A system equipped with a processor that performs a process that includes the following.

2. The survey results include one or more sentences, Assigning the aforementioned classification includes assigning the aforementioned classification to each of the aforementioned sentences. The system according to claim 1.

3. Assigning classifications to the survey results includes assigning the classifications using artificial intelligence from among a pre-prepared set of classifications. The system according to claim 1.

4. Assigning classifications to the survey results includes providing the artificial intelligence with pre-defined conditions for assigning the classifications, and the artificial intelligence assigning the classifications based on those conditions. The system according to claim 3.

5. The process further comprises outputting the assigned classification and the survey results in association. The system according to claim 1.

6. The process further comprises performing a positive / negative determination to determine whether the survey results include positive or negative opinions. The system according to claim 1.

7. The survey results include one or more sentences, The aforementioned positive / negative determination includes performing a positive / negative determination for each sentence. The system according to claim 6.

8. The aforementioned positive / negative determination includes calculating a positive / negative score, which is a numerical representation of the result of the positive / negative determination. The system according to claim 6.

9. The aforementioned positive / negative determination is performed by artificial intelligence. The aforementioned positive / negative determination includes obtaining the reason for calculating the positive / negative score using the artificial intelligence. The system according to claim 8.

10. The process further comprises visually distinguishing and displaying the survey results based on the results of the positive / negative determination. The system according to claim 6.

11. The system according to claim 6, further comprising: determining whether notification is necessary based on the result of the negative / positive determination; and, if it is determined that notification is necessary, providing information regarding the survey results.

12. The aforementioned positive / negative determination further includes determining whether the survey results include neutral opinions. The system according to any one of claims 6 to 11.

13. The process further comprises configuring the viewing settings for the survey results according to the classification assigned to the survey results. The system according to claim 1.

14. The challenges of the aforementioned organization are generated for each of the aforementioned classifications. The system according to claim 1.

15. The process further comprises at least one of the following: presenting the survey results that formed the basis for generating the organization's issues, and presenting the survey results that formed the basis for solutions to the organization's issues based on the survey results. The system according to claim 1.

16. Obtain survey results, which are the results of a survey conducted on survey subjects within an organization and answered in a free-response format, By analyzing the survey results, one or more classifications are assigned to the survey results. A negative / positive determination is made to determine whether the survey results include positive or negative opinions. Based on the results of the aforementioned positive / negative assessment, a determination is made as to whether notification is necessary. If it is determined that notification is necessary, information regarding the survey results is provided. A system equipped with a processor that performs a process that includes the following.

17. We obtain survey results, which are the results of a survey conducted among survey participants within an organization, with responses in an open-ended format. By analyzing the survey results, one or more classifications are assigned to the survey results. Artificial intelligence is used to generate the organization's challenges based on the survey results, The system generates and presents, using artificial intelligence, at least one of the solutions to the organization's challenges based on the survey results, and at least one of the solutions to the organization's challenges not based on the survey results. A method performed by a computer that includes the following:

18. Obtain survey results, which are the results of a survey conducted on survey subjects within an organization and answered in a free-response format, By analyzing the survey results, one or more classifications are assigned to the survey results. A negative / positive determination is made to determine whether the survey results include positive or negative opinions. Based on the results of the aforementioned positive / negative assessment, a determination is made as to whether notification is necessary. If it is determined that notification is necessary, information regarding the survey results is provided. A method performed by a computer that includes the following:

19. We obtain survey results, which are the results of a survey conducted among survey participants within an organization, with responses in an open-ended format. By analyzing the survey results, one or more classifications are assigned to the survey results. Artificial intelligence is used to generate the organization's challenges based on the survey results, The system generates and presents, using artificial intelligence, at least one of the solutions to the organization's challenges based on the survey results, and at least one of the solutions to the organization's challenges not based on the survey results. A computer program that causes a computer to perform an action that includes a certain condition.

20. Obtain survey results, which are the results of a survey conducted on survey subjects within an organization and answered in a free-response format, By analyzing the survey results, one or more classifications are assigned to the survey results. A negative / positive determination is made to determine whether the survey results include positive or negative opinions. Based on the results of the aforementioned positive / negative assessment, a determination is made as to whether notification is necessary. If it is determined that notification is necessary, information regarding the survey results is provided. A computer program that causes a computer to perform an action that includes a certain condition.

Images