Development of an AI Userbot for Engineering Design Education Using an Intent and Flow Combined Framework
Round 1
Reviewer 1 Report
This paper presents a methodology to allow engineering design students to be able to chat with potential users to improve their products but, in this case, the potential users are chatbots programmed for that purpose. The paper describes the procedure of data collection, chatbot setup using Dialogflow, and describes some experiments that were done using four students and Google Assistant.
The idea described in this paper is nice and interesting. The main weakness is the description of how the interviews were converted into intents and entities. The paper would strongly benefit from improving this point. I think it would also be interesting to elaborate a bit more on the final experiments performed by the students.
Other minor comments:
Title
- Delete the initial “The”
Section 1
- Line 49: proposed → proposes
- Line 90: aimed → aim
Section 2
- Line 93: The proposed → Proposed
- Line 94: aimed → aims; were → are
- Line 96: and used Google Assistant → delete “used”
- In general, use the present tense in the whole paragraph.
- Line 101: Intent and Entities → don’t use capital letters for Intent and Entities (applies to all the occurrences of this).
- Line 105: from 0.0 to 1.0 → just use 0 and 1
- Line 107: please use “user” and not “system user”, especially when “user” appears several times in a sentence/paragraph. We know you mean a system user, no need to specify.
- Line 108: this example could be improved.
- Line 115: As the system has increasingly utilized… → Change the tense of this sentence: As the system is increasingly utilized, the Dialogflow platform begins to
- Line 121: delete “The”
- Line 122: Change the title of this section, it is not correct.
- Line 127: interview results → results of the interviews
- Line 131: use italics instead of quotation marks here
- Line 136: same as line 127
- Line 137: I would delete “The detailed procedure … following” and just keep the sentence after that.
- Line 141: Please rewrite this sentence, it’s a bit difficult to follow
- Line 148: this paragraph can be improved as well
- Line 151: was built → are built
- Line 152: Interface (a) shows …; and interface (b) shows …
- Figure 2: It would be nice if it was translated to English so that readers can understand
- Line 155: chips → chip
- Line 156: user interface; system users …
- Line 164: delete “The”
- Line 165: the procedure of how the system is trained can be described better, it’s a bit difficult to follow.
- Line 171: receives a corrected pair
Section 3
- Line 175: establishment → setting?
- Line 176: could be called → can be called
- Line 180: delete “the” before Google Assistant
- Line 191: and suggested that the system user expressed
- Line 198: shows that a chatbot
- Line 204: Delete initial “The”.
Section 4
- Line 220: According to the data
- Line 221: The students entered
- Line 222: “To expand the direction of the discussion” → not sure what this means
Author Response
Dear Reviewer,
Thank you and other reviewers giving us positive comments on our manuscript. All changes requested from reviewers have been included in the manuscript and the response to each reviewer’s comment is listed below. The original manuscript was edited by Native English speakers of the MDPI editing service, we will have an English editing to the revised manuscript to ensure the language quality of the present manuscript to meet journal's standard.
Response to Reviewer 1 Comments
Point 1: This paper presents a methodology to allow engineering design students to be able to chat with potential users to improve their products but, in this case, the potential users are chatbots programmed for that purpose. The paper describes the procedure of data collection, chatbot setup using Dialogflow, and describes some experiments that were done using four students and Google Assistant. The idea described in this paper is nice and interesting. The main weakness is the description of how the interviews were converted into intents and entities. The paper would strongly benefit from improving this point. I think it would also be interesting to elaborate a bit more on the final experiments performed by the students.
Response: Thanks for your agreement and support. There are four steps and criteria in building userbot’s intent:
- Determining every intent of each sample question sentence based on product experience, preference, and background of the AI userbot.
- Enumerating the syntax of a question sentence with the same intent as much as possible to help the system recognize the consistent intent in different wordings of sentences (Figure 2).
- Defining entities in the sample question sentence to help the system recognize intent more accurately (Figure 2).
- Creating a unified response to the sample question sentence based on the persona of the userbot. In addition to text responses, pictures, audio files, videos, and buttons are materials for the system to use to respond to system users.
We added a statement to describe the two failed intents (lines 287-289) which elaborate a bit more on the final experiments performed by the students.
We focused on describing the importance of helping students to collaborate with users in a design project, the process of developing the AI userbot system, and the feasibility of the collaboration between students and AI userbot. We realized there is a need to have more examinations on our system with more participants. We have mentioned this in the section of Future Study.
Point 2: Other minor comments.
Response: Thank you for carefully pointing out texts that need to be improved in our manuscript. We revised the manuscript according to each of your indications.
Reviewer 2 Report
This article discusses an application of used Dialogflow and Google Assistant to build a system architecture, and applied methods of persona and semi-structured interviews to develop AI virtual product users.
The paper deals with an interesting problem and show an interesting work.
It would be recommended that the authors make an effort to rewrite section 1. References to the advancements on the topic are missing, and please explain how your article builds on these.
This paper appears to be an application case as the underlying methods do not appear to be new. It is advisable to completely describe the authors' methodology. Is the method described completely new, or a better version of an existing method? Authors could describe your methodology with a new section.
Please provide a better description of the implementation.
It would be recommended that the authors make an effort to redesign your figures. There are several figures not in English. All figures in English language could make the document easier to read.
Feedback of the development is partially found. Why can your system effectively participate in student design activities in the early stage of design? Please, explain how your article builds on this.
Authors should indicate the advantages and disadvantages of your proposal compared to other
Author Response
Dear Reviewer,
Thank you and other reviewers giving us positive comments on our manuscript. All changes requested from reviewers have been included in the manuscript and the response to each reviewer’s comment is listed below.
The original manuscript was edited by Native English speakers of the MDPI editing service, we will have an English editing to the revised manuscript to ensure the language quality of the present manuscript to meet journal's standard.
Response to Reviewer 2 Comments
Point 1: This article discusses an application of used Dialogflow and Google Assistant to build a system architecture, and applied methods of persona and semi-structured interviews to develop AI virtual product users.
Response: Thanks for your agreement and support.
Point 2: The paper deals with an interesting problem and show an interesting work. It would be recommended that the authors make an effort to rewrite section 1. References to the advancements on the topic are missing, and please explain how your article builds on these.
Response: We added a paragraph (lines 96-102) to describe current AI research in educational application and indicate the importance of our research in engineering design education field.
Point 3: This paper appears to be an application case as the underlying methods do not appear to be new. It is advisable to completely describe the authors' methodology. Is the method described completely new, or a better version of an existing method? Authors could describe your methodology with a new section. Please provide a better description of the implementation.
Response: We added a section of “Method” as suggested. The innovation of this study is that we collected users’ information with persona and semi-structured interviews, and built them into the AI userbot system using Dialogflow and Google Assistant.
Point 4: It would be recommended that the authors make an effort to redesign your figures. There are several figures not in English. All figures in English language could make the document easier to read.
Response: We revised figures 2, 4, 5, and 6, all the texts in figures are translated into English.
Point 5: Feedback of the development is partially found. Why can your system effectively participate in student design activities in the early stage of design? Please, explain how your article builds on this.
Response: Students talked to AI userbots through oral real-time dialogs to explore real users' product preferences, use experience, and needs, and eventually the students collaborated with the userbots to generate design solutions in the early stage of design. Detailed discussions present at the section 4 in the manuscript.
Point 6: Authors should indicate the advantages and disadvantages of your proposal compared to other.
Response: We have discussed the pros and cons of collaboration between students and the AI userbots in the section 4 of the manuscript. To our knowledge, there is no other chatbot system developed for engineering design education. AI userbot system is the unique educational chatbot system developed for engineering design education. Consequently, we added a paragraph (lines 96-102) to describe current AI research in educational application and indicate the importance and the originality of our research in engineering design education field. Moreover, we also added another paragraph in the section 5 to express the need of future related research helping us to find the advantages and backdraws of our system compared to other related AI educational chatbots.
Reviewer 3 Report
The topic of the paper is important, and relevant for engineering design education field. However, the manuscript has several critical flaws that largely diminishes the quality of this study. Some of them are lack of development and elaboration, empirical evidence, transparency, as well as problems with structure, method and scientific rigor. These are commented as follows:
Introduction section: The quotation should be interpreted in the text. If possible, avoid quotations.
The topic about the importance to define target users and understand their needs, and its relation to design education is underdeveloped.
Section 1.1 Chatbots is rather technical and should be moved after Section 1.2 dealing with Engineering Participatory Design. In this section, the connection between open and closed-domain chatbots with intent- or flow-based issues is unclear. What is better or worse for each situation?
It is unclear how students can benefit from chatbots, and how these can replace real users.
The use of chatbots in design, and design collaboration in particular, is underdeveloped. A separate section must be included after the section dealing with Chatbots. It must present key prior studies and discuss their relation to the research goal.
Section 1.2. Engineering Participatory Design: Authors must discuss key studies illustrating how participatory design allows designers and product users to jointly participate in the construction of design problems, and how this help understand user needs.
Section 2.1. The Proposed Architecture System: You say that AI userbots were built for different design themes. Explain what themes you mean?
Explain in further detail what is natural language understanding (NLU), and include references.
At this stage, a comment should be made about how the system related to virtual users is represented, what this data-base consists of, and from where it is obtained or retrieved.
A Method section should be included, and explanations provided with regard to the methodological approach followed in this paper, including all the information concerned with the semi-structured interviews, and the ‘persona’ / ‘persona design’ which appears out-of-the-blue. Procedure followed and sample information for both students and users should be included as well.
The focus of this study is mainly set on the building of userbots. However, no information is provided regarding how the empirical data collected in the interviews was analyzed to this end. How the content was coded and analyzed? Who analyzed it? What level of agreement was reached between the researchers? What themes pop-up? Moreover, no empirical evidence is provided regarded the data analyzed, and examples of the outcomes.
Section 2.1. The Userbot Build: how the 137 Intents were created. What was the criteria, and what the relation to the information obtained from the interviews?
Why only six women were included as participants in the interviews? Normally eight is the minimum. No background information is offered for the participants. Likewise, why only four and not at least eight participated in the study?
In figure 3, the dialog process map is far than clear, and should be interpreted in the text.
Since figures 4 and 5 are not depicted in English, the connection with their corresponding parts in the text is uncertain.
Section 4.2 Results: Results are speculative provided that they are not based on empirical evidence of any kind, and no examples are presented at all. Readers must trust authors comments as a sole evidence.
There is no Discussion section attempting to discuss the results, and connecting them to existing literature in any way. Hence, the importance, impact, and contribution of this study in the field remains uncertain, speculative and unsupported by empirical results.
Is the userbot AI based system able to learn anything in order to improve its performance in the future?
It would be useful if authors could provide internet link where it could be possible for the audience to try this system.
The references used in this study should be expanded to include relevant and updated bibliographical sources of existing sections be developed further, and the background sections to be included.
Other issues:
Where you say: “ In response to this, several studies have emphasized the importance of having students cooperate with product users in design education [4].” the opposite is true: "to have product users cooperating with students"
Since this is an English based journal, do not include other foreign language in the text or in the figures that the audience cannot read. (e.g., Figures 2,4,5, and 6)
Author Response
Dear Reviewer,
Thank you and other reviewers giving us positive comments on our manuscript. All changes requested from reviewers have been included in the manuscript and the response to each reviewer’s comment is listed below. The original manuscript was edited by Native English speakers of the MDPI editing service, we will have an English editing to the revised manuscript to ensure the language quality of the present manuscript to meet journal's standard.
Response to Reviewer 3 Comments
Point 1: Introduction section: The quotation should be interpreted in the text. If possible, avoid quotations.
Response: We removed the quotation as suggested.
Point 2: The topic about the importance to define target users and understand their needs, and its relation to design education is underdeveloped.
Response: We re-arranged the content of the whole section 1 and added a paragraph (lines 96-102) to describe current AI research in educational application and indicate the importance of our research in engineering design education field.
Point 3: Section 1.1 Chatbots is rather technical and should be moved after Section 1.2 dealing with Engineering Participatory Design. In this section, the connection between open and closed-domain chatbots with intent- or flow-based issues is unclear. What is better or worse for each situation?
Response: We re-arranged the two sections as suggested, and we have a description regarding the connection between open and closed-domain chatbots with intent- or flow-based dialogues which is on the “Proposed Architecture System” section.
Point 4: It is unclear how students can benefit from chatbots, and how these can replace real users.
Response: We have stated the difficulty of inviting real users to participate in students’ design process because of time, budget, and location (lines 38-39), and indicate the importance of our research in using virtual users to replace real user in engineering design education. The benefits and backdraws of our AI userbot system was also discussed on the section 5.
Point 5: The use of chatbots in design, and design collaboration in particular, is underdeveloped. A separate section must be included after the section dealing with Chatbots. It must present key prior studies and discuss their relation to the research goal.
Response: Thank reviewer’s suggestion. We added a paragraph (91-95) as suggested.
Point 6: Section 1.2. Engineering Participatory Design: Authors must discuss key studies illustrating how participatory design allows designers and product users to jointly participate in the construction of design problems, and how this help understand user needs.
Response: We added a paragraph to present more information related to participatory design (59-63).
Point 7: Section 2.1. The Proposed Architecture System: You say that AI userbots were built for different design themes. Explain what themes you mean?
Response: We revised the description from design theme to design project for clarifying that instructors from all over the world can use this system with their familiar language to build chatbots for different design projects in helping students’ design with virtual chatbots.
Point 8: Explain in further detail what is natural language understanding (NLU), and include references.
Response: We added an Explanation of NLU (lines 114-117) with two references in the manuscript.
Point 9: At this stage, a comment should be made about how the system related to virtual users is represented, what this data-base consists of, and from where it is obtained or retrieved.
Response: We have described related information on the section 2.1 and 2.2.
Point 10: A Method section should be included, and explanations provided with regard to the methodological approach followed in this paper, including all the information concerned with the semi-structured interviews, and the ‘persona’ / ‘persona design’ which appears out-of-the-blue. Procedure followed and sample information for both students and users should be included as well.
Response: We added a section of “Method” and re-arranged the entire content of the manuscript as suggested.
Point 11: The focus of this study is mainly set on the building of userbots. However, no information is provided regarding how the empirical data collected in the interviews was analyzed to this end. How the content was coded and analyzed? Who analyzed it? What level of agreement was reached between the researchers? What themes pop-up? Moreover, no empirical evidence is provided regarded the data analyzed, and examples of the outcomes.
Response: We have described related information on the section 2.1 and 2.2.
Point 12: Section 2.1. The Userbot Build: how the 137 Intents were created. What was the criteria, and what the relation to the information obtained from the interviews?
Response: There are four steps and criteria in building a userbot:
- Determining every Intent of each sample question sentence based on product experience, preference, and background of the AI userbot.
- Enumerating as much as possible the syntax of a question sentence with the same Intent, to help the system recognize the consistent intent in different wordings of sentences (Figure 2).
- Defining Entities in the sample question sentence to help the system recognize Intent more accurately (Figure 2).
- Creating a unified response to the sample question sentence based on the persona of the userbot. In addition to text responses, pictures, audio files, videos, and buttons are materials for the system to use to respond to system users.
Point 13: Why only six women were included as participants in the interviews? Normally eight is the minimum. No background information is offered for the participants. Likewise, why only four and not at least eight participated in the study?
Response: After interviewing six women and analyzing their responses, we found many repeated information among their responses, and therefore we decided six women were enough. The information collected from the six women could be classified into three different virtual product users with different personalities, product experience, and preferences. Therefore, three AI userbots for self-defense products were established based on the interview results. The AI userbot will learn automatically and add more Intent gradually. Furthermore, we focused on describing the importance of helping students to collaborate with users in a design project, the process of developing the AIuserbot system, and the feasibility of the collaboration between students and AI userbot. We realized there is a need to have more examinations on the our system with more participants. We have mentioned this in the section of Future Study.
Point 14: In figure 3, the dialog process map is far than clear, and should be interpreted in the text.
Response: We added statement (lines 182-183) to describe dialog process map.
Point 15: Since figures 4 and 5 are not depicted in English, the connection with their corresponding parts in the text is uncertain.
Response: We revised figures 4, 5, all the texts in figures are translated into English.
Point 16: Section 4.2 Results: Results are speculative provided that they are not based on empirical evidence of any kind, and no examples are presented at all. Readers must trust authors comments as a sole evidence.
Response: We-arranged the entire sections of the manuscript, and we added references in the texts as the empirical evidence regarding the discussions on the outcomes of the present study. Please refer to section of “Discussion and Conclusions”.
Point 17: There is no Discussion section attempting to discuss the results, and connecting them to existing literature in any way. Hence, the importance, impact, and contribution of this study in the field remains uncertain, speculative and unsupported by empirical results.
Response: We-arranged the entire sections of the manuscript, and the discussions were appeared in section 4. Moreover, we added 10 references in revised manuscript to reinforce the empirical evidence to our study.
Point 18: Is the userbot AI based system able to learn anything in order to improve its performance in the future?
Response: YES. It is the most important function and attribute using AI technology of Dialogflow flow. The AI userbot will learn automatically and add more Intent gradually.
Point 19: It would be useful if authors could provide internet link where it could be possible for the audience to try this system.
Response: We are waiting Google’s approval for publishing our system. Currently, the system is a beta version; only invited users can access the system to communicate with AI-userbot. Please provide me the email address and I will send the invitation.
Point 20: The references used in this study should be expanded to include relevant and updated bibliographical sources of existing sections be developed further, and the background sections to be included.
Response: We added 10 references in revised manuscript to reinforce the empirical evidence to our study.
Point 21: Other issues: Where you say: “ In response to this, several studies have emphasized the importance of having students cooperate with product users in design education [4].” the opposite is true: "to have product users cooperating with students"
Response: We revised the discrepancy accordingly.
Point 22: Since this is an English based journal, do not include other foreign language in the text or in the figures that the audience cannot read. (e.g., Figures 2,4,5, and 6)
Response: We revised figures 2, 4, 5, and 6, all the texts in figures are translated into English.
Reviewer 4 Report
This manuscript is an article related to chatbots and ΑΙ for participatory design in engineering design education and is adequate for an applied journal of sustained academic excellence like Applied Sciences.
This work is an interesting article for participatory design using AI userbots and I found it extremely interesting and valuable for the scientific community. My comments to strengthen your work and make it clearer and more robust are:
In subsection 1.1 please make your engagement with the literature better especially with applications of chatbots in education and industry (lines 67-68).
In subsection 2.1 please give more details about used technologies (Dialogflow, Google Assistant) based on literature.
In subsection 3.3 please make clearer the roles of Ring and Ting userbots for the reader.
In the title of section 4 you mean Evaluation (the implementation was presented in section 3).
In section 4 please give more details in subsection 4.1. Also, in subsection 4.2 give more details about the results (How did you use the recorded videos, which of Intents failed?) and work on figure 6 to make the sketch resolution better and add English text for the description of the components on the design/sketch.
Finally, in all figures provide for userbots messages the English translation for the international readers.
Author Response
Dear Reviewer,
Thank you and other reviewers giving us positive comments on our manuscript. All changes requested from reviewers have been included in the manuscript and the response to each reviewer’s comment is listed below. The original manuscript was edited by Native English speakers of the MDPI editing service, we will have an English editing to the revised manuscript to ensure the language quality of the present manuscript to meet journal's standard.
Response to Reviewer 4 Comments
Point 1: This manuscript is an article related to chatbots and ΑΙ for participatory design in engineering design education and is adequate for an applied journal of sustained academic excellence like Applied Sciences. This work is an interesting article for participatory design using AI userbots and I found it extremely interesting and valuable for the scientific community.
Response: Thanks for your agreement and support.
Point 2: In subsection 1.1 please make your engagement with the literature better especially with applications of chatbots in education and industry (lines 67-68).
Response: A new paragraph added to lines 96-102, as suggested.
Point 3: In subsection 2.1 please give more details about used technologies (Dialogflow, Google Assistant) based on literature.
Response: A new paragraph and two new references were added to lines 114-117, as suggested.
Point 4: In subsection 3.3 please make clearer the roles of Ring and Ting userbots for the reader.
Response: We replaced the Figure 5. All the texts in Figure 5 are translated into English which makes clearer the roles of Ring and Ting userbots.
Point 5: In the title of section 4 you mean Evaluation (the implementation was presented in section 3).
Response: Thanks for your correction.
Point 6: In section 4 please give more details in subsection 4.1. Also, in subsection 4.2 give more details about the results (How did you use the recorded videos, which of Intents failed?) and work on figure 6 to make the sketch resolution better and add English text for the description of the components on the design/sketch.
Response: We focused on the description of the development of the AI userbot system and we decided not to include the analysis of the observational content from recording videos. We added a statement to describe the two failed Intents (lines 287-289). A better resolution of Figure 6 was added to the manuscript.
Point 7: Finally, in all figures provide for userbots messages the English translation for the international readers.
Response: We revised figures 2, 4, 5, and 6, all the texts in figures are translated into English.
Reviewer 5 Report
The paper, titled “The development of an AI Userbot for engineering design education using an intent and flow combined framework” by Y.-H. Chien and C.-K. Yao, presents the use of Dialogflow and Google Assistant to build a system architecture in student design activities.
The introduction is well-written. Unfortunately, the English style of the other paragraphs needs significant improvement. Some sentences are not clear. Paragraph 2.1 requires more details. Some parts of Figures 2, 4, 5, and 6 must be translated into English.
Can the chatbot, built in Taiwan's language, be used in other foreign languages as it is? Does it require “adjustments” to be used in other countries?
Author Response
Dear Reviewer,
Thank you and other reviewers giving us positive comments on our manuscript. All changes requested from reviewers have been included in the manuscript and the response to each reviewer’s comment is listed below. The original manuscript was edited by Native English speakers of the MDPI editing service, we will have an English editing to the revised manuscript to ensure the language quality of the present manuscript to meet journal's standard.
Response to Reviewer 5 Comments
Point 1: The paper, titled “The development of an AI Userbot for engineering design education using an intent and flow combined framework” by Y.-H. Chien and C.-K. Yao, presents the use of Dialogflow and Google Assistant to build a system architecture in student design activities. The introduction is well-written. Unfortunately, the English style of the other paragraphs needs significant improvement. Some sentences are not clear. Paragraph 2.1 requires more details. Some parts of Figures 2, 4, 5, and 6 must be translated into English.
Response: we added 10 references in revised manuscript to reinforce the empirical evidence to our study. And all texts in the Figures are translated into English, as suggested.
Point 2: Can the chatbot, built in Taiwan's language, be used in other foreign languages as it is? Does it require “adjustments” to be used in other countries?
Response: We used Dialogflow to build the AI userbot system. Future instructors from all over the world can use this system with their familiar language to build chatbot for different design projects in helping students’ design with virtual chatbots.
Round 2
Reviewer 3 Report
The manuscript was improved and now is much more consistent, developed and clearer than before. However, I still found the following points raised in the previous review as overlooked and needing further clarification and development:
Point 10: Explanations must be provided with regard to the methodological approach followed in this paper, including all the information concerned with the semi-structured interviews, Procedure followed and sample information for both students and users should be included as well.
Point 11: The focus of this study is mainly set on the building of userbots. However, no information is provided regarding how the empirical data collected in the interviews was analyzed to this end. How the content was coded and analyzed? Who analyzed it? What level of agreement was reached between the researchers? What themes emerged that were later on used in the system? Moreover, no empirical evidence is provided regarding the data analyzed from the interviews, and examples of the outcomes from the interviews. This is basic to understand the relationship between data obtained from the interviews and it was used to develop the Userbot system.
Author Response
Response to Reviewer’s Comments
Point 1: Explanations must be provided with regard to the methodological approach followed in this paper, including all the information concerned with the semi-structured interviews, Procedure followed and sample information for both students and users should be included as well.
Response 1: Thank reviewer for carefully reviewing the revised manuscript and pointing out what still needs to be explained. We add statements to the manuscript according to your suggestions. Please refer to the red texts in the manuscript of pages 4, 8, 9, and 11.
Point 2: The focus of this study is mainly set on the building of userbots. However, no information is provided regarding how the empirical data collected in the interviews was analyzed to this end. How the content was coded and analyzed? Who analyzed it? What level of agreement was reached between the researchers? What themes emerged that were later on used in the system? Moreover, no empirical evidence is provided regarding the data analyzed from the interviews, and examples of the outcomes from the interviews. This is basic to understand the relationship between data obtained from the interviews and it was used to develop the Userbot system.
Response 2: We add an example to explain how did we collect empirical data in the interviews and how did the data (“questions to users” and “users’ response”) become intent and entity. For instance, when a designer asked a user “Have you ever used self-defense products?”, we considered the intent of this question was “the use experience”, and the entity was “use” and “self-defense product.” The entire process of establishing intent and entity was different from semantic analysis and coding. The correctness of intent and entity will be gradually improved by training chatbot with Dialogflow and by automatic learning mechanism with AI technology.
Currently we have only built the AI userbots for the project of female self-defense product design. We will add product design projects for users whom students are not familiar with, need to pay more attention to, and have to collaborate with, such as pregnant women, elderly, or disabled.
We will have an English editing to the final version of manuscript to ensure the language quality of the present manuscript to meet journal's standard.
Reviewer 4 Report
I am satisfied with the new text, the authors tackled my comments.
Author Response
Response to Reviewer's Comments
Point 1: I am satisfied with the new text, the authors tackled my comments.
Response: Thanks for your agreement and support.
Reviewer 5 Report
The authors have improved their paper according to the reviewers' comments.
Author Response
Response to Reviewer's Comments
Point 1: The authors have improved their paper according to the reviewers' comments.
Response: Thanks for your agreement and support.
Round 3
Reviewer 3 Report
With the last additions the quality of the paper was enhanced. However one more issue is pending.
Authors say "In total, 137 intents were established; they are 45, 48, and 44 intents for the three AI userbots, 177 respectively. In the entities section, 58 total keyword-categories, corresponding to 20, 22, and 16 178 keyword categories respectively, were established as well." For the sake of transparency and clarity, authors should include a table, which can also be located in the Appendix, indicating what keyword-categories were identified from the participants (rows), and what keyword-categories correspond to what participants (columns). These will also enable to see what keyword-categories were more frequent.
After dealing with this issue, the paper can be suitable for publication.
Author Response
Response to Reviewer’s Comments
Point 1: Authors say "In total, 137 intents were established; they are 45, 48, and 44 intents for the three AI userbots, 177 respectively. In the entities section, 58 total keyword-categories, corresponding to 20, 22, and 16 178 keyword categories respectively, were established as well." For the sake of transparency and clarity, authors should include a table, which can also be located in the Appendix, indicating what keyword-categories were identified from the participants (rows), and what keyword-categories correspond to what participants (columns). These will also enable to see what keyword-categories were more frequent.
Response 01: An appendix is added to the manuscript, as suggested. Thank you.
