AI Capabilities and Its Impact on Organisational Innovation in Malaysian SMEs: The Role of Transformational Leadership and Digital Organisational Culture

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. The introduction does not clearly explain why Malaysian SMEs are the main focus. Much of the background is quite general and could apply to SMEs in many other countries.
2. The relationships among AI capability, transformational leadership, and organisational innovation are already well established in prior research. It is not clear what new theoretical insight is gained by examining these links again in the current context.
3. The contribution section is relatively long compared to the problem formulation. Several points repeat general academic or practical implications, which makes the introduction feel unbalanced.
4. The rationale for the hypotheses is reasonable. However, the literature review and hypothesis development are overly long and somewhat repetitive. These sections could be combined to improve conciseness.
5. Table 1 (EFA results) is mentioned in the text, but the table itself does not appear in the manuscript.
6. Composite reliability values for the main constructs are not clearly reported. In particular, it is unclear whether the reported rho-a values represent composite reliability at the construct level.
7. The statistical analysis section is unnecessarily long for a standard PLS-SEM study. Several procedures described do not seem essential.
8. A methodological concern relates to the use of owners and managers as respondents for measuring transformational leadership. It is unclear whose leadership is being evaluated, which raises questions about measurement validity.
9. The discussion section is acceptable but relatively weak. Practical implications, especially those specific to SMEs, are not clearly developed.
10. Although SMEs from different industry sectors are included, industry-specific differences in AI capability are not discussed. This limits the generalisability of the findings.
11. The manuscript is considerably longer than typical articles in Sustainability and reads more like a thesis-style document. Substantial condensation is needed.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Reviewer Comments to the Authors

 

1. The structure of the first section needs improvement. Please merge "1.1. Research Problem and Questions" and "1.2. Research Contribution" into the main "1. Introduction" section. The Introduction should be a continuous narrative without these sub-headings to ensure better flow and readability. Additionally, please remove the unnecessary bold text formatting used throughout the body paragraphs; standard academic formatting should be maintained.
2. There is significant inconsistency in the reporting of decimal places throughout the manuscript. For example, Table 4 reports skewness with 3 decimal places (e.g., -0.653) but other sections or tables sometimes use 2 decimal places. Please standardize all statistical reporting (including text and tables) to 3 decimal places (or 2, depending on the journal's specific style guide) to ensure uniformity and precision.
3. The literature review requires updating. While some recent studies are cited, the field of AI and digital culture is evolving rapidly. I strongly recommend adding more high-quality journal articles published in 2024 and 2025 to strengthen the theoretical argumentation and highlight the timeliness of this research.
4. Regarding the methodology, the sampling procedure is confusing. The text states that "900 companies were randomly selected," but later mentions that "Purposive sampling was used to identify the target participants." Please clarify this multi-stage process. Specifically, explain how you moved from random company selection to purposive individual selection without introducing selection bias.
5. In terms of statistical analysis (PLS-SEM), the paper relies on self-reported data from a single source (managers/owners), which raises concerns about Common Method Bias (CMB). Although VIF values are reported, this primarily addresses multicollinearity among predictors. Please conduct and report a specific test for CMB, such as Harman’s Single Factor Test or, preferably, the Full Collinearity Assessment (where VIFs should be < 3.3 for all latent variables), to demonstrate that CMB is not a serious issue.
6. The treatment of control variables needs to be addressed. The conceptual framework (Figure 1) and the text explicitly state that "Firm size" and "Industry type" are control variables. However, the results section (specifically the structural model and path coefficients) does not report the statistical influence of these control variables on the dependent variable. Please include the results for these control variables in the path analysis tables to verify if they affect the model's outcome.
7. Table 17 (HTMT) is currently poorly formatted, making it difficult to read the variable names and their corresponding values. Please reformat this table clearly. Furthermore, ensure that the discussion explicitly states whether the HTMT values meet the threshold (usually < 0.90 or < 0.85) to establish discriminant validity.
8. References [1] through [5] in the "References" section are currently empty or missing. This is a major oversight. Please ensure all references are complete, accurate, and correspond correctly to the in-text citations.
9. Throughout the manuscript, there is a consistent typographical error where the abbreviation for Artificial Intelligence is written as "Al" (using a lowercase 'l') instead of "AI" (using an uppercase 'I'). This error appears in the Title, the Abstract, and frequently throughout the hypotheses and body text. Please perform a global check and replace all instances with the correct "AI" to ensure professional presentation.
10. Tables 5 and 9 indicate that "AI Capability" (comprising Tangible, Human Skills, and Intangible dimensions) and "Organisational Innovation" (comprising Product, Process, and Administrative dimensions) are multidimensional constructs. However, the methodology section does not clearly explain how these were modeled in SmartPLS. Please clarify in the text whether a second-order approach (such as the Disjoint Two-Stage Approach or Repeated Indicators Approach) was used to analyze these variables, as Table 23 currently treats AI Capability as a single latent construct.
11. Table 19 reports the R² value for Transformational Leadership as 0.078. This indicates that the current model explains less than 8% of the variance in this construct, which is considered weak. Please acknowledge this low explanatory power in the Discussion section as a limitation and suggest what other potential antecedents future research might explore to better explain Transformational Leadership in this context.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The submitted article addresses a highly topical and important issue concerning the impact of artificial intelligence (AI) capabilities on organizational innovativeness in the small and medium-sized enterprise (SME) sector in Malaysia.

Evaluation of the individual sections:

Abstract
The abstract concisely presents the research objective, the methodology employed (SMART-PLS), and the key findings concerning the role of transformational leadership and digital organizational culture. It clearly indicates which hypotheses were supported and which were rejected. However, it lacks an explicit indication of the study’s unique theoretical or practical contribution that would distinguish it from the existing literature at the initial stage of reading.

Introduction

The authors situate the study within the economic context of Malaysia, identifying SMEs as the “backbone” of the national economy. The research gap is well defined, particularly by highlighting the scarcity of studies on the mediating role of transformational leadership in the AI–innovation relationship in this region. Nevertheless, this section could more extensively address ethical issues related to AI implementation (such as data privacy or algorithmic bias), which are highly relevant in the context of sustainable development.

Literature Review
This section is very extensive and multidimensional, covering definitions of transformational leadership, organizational culture, and the technical aspects of AI (Machine Learning, NLP). The hypotheses are logically derived from TOE (Technology–Organization–Environment) theory and organizational culture theory. Nevertheless, although the review is broad, it is at times overly descriptive. Greater emphasis could be placed on a critical analysis of contradictory findings in prior studies, which would better justify the need to test moderating variables.

Research Methodology
Rigorous procedures for validating the research instruments were applied, including EFA, the KMO measure, and Bartlett’s test of sphericity. However, the major shortcoming concerns the representativeness of the sample, as well as inconsistencies in the authors’ statements regarding generalization. Although the limitations section acknowledges that the results cannot be generalized, the methodology section states that the selected number of 900 firms (the initial sample) was sufficient to ensure adequate statistical representativeness and to allow generalization to the entire target population. While the authors undertook steps to enhance the reliability of the study (random sampling from an official database), the geographical concentration of the sample and reliance on voluntary online responses suggest that the sample is more illustrative of selected regions than fully representative of the entire SME sector. Therefore, I recommend revising the conclusions to more clearly emphasize that the findings primarily apply to the surveyed regions and to the specific digital context in which the responding managers operate.

In addition, the current version of the manuscript contains a structural error involving duplicated section titles. Both subsection 3.2 and the main Section 4 are titled “Results.” I therefore suggest renaming Section 3.2, as it does not present data analysis results but rather describes the instruments used to measure the variables. A proposed revised title for Section 3.2 is: “Research Instruments and Operationalization.”

Results
The statistical analysis is comprehensive and includes reliability (Cronbach’s alpha), convergent validity (AVE), and discriminant validity (the Fornell–Larcker criterion). The use of the f² (effect size) and Q² (predictive relevance) indicators adds methodological depth to the study.
Nevertheless, the results show that the effect sizes (f²) for most relationships are classified as “small” (values ranging from 0.036 to 0.092). The authors should devote greater attention to interpreting what such small effects mean in practical, business terms for SMEs.

Discussion

The authors make an interesting attempt to explain the rejection of Hypothesis H5, suggesting that the impact of AI on leaders may be independent of digital culture. The findings are appropriately discussed in relation to the existing literature (e.g., the works of Alshehri and Mohib). However, the practical recommendations remain rather general (e.g., “invest in AI”). More concrete guidance is needed on how resource-constrained SMEs can develop “AI capabilities” without incurring substantial technological costs.

 

Conclusions

In its current form, Section 5 (Conclusions) is too concise and does not fully reflect the significance of the research conducted. In the manuscript, this section consists of only one short paragraph that largely reiterates the main findings regarding the impact of AI on innovativeness and the mediating and moderating roles of the remaining variables.

I suggest that the authors expand this section along the following lines in order to give the paper a more synthetic and mature character:

• Rather than restating which hypotheses were supported, the conclusions should synthesize insights into how technology (AI), human leadership, and organizational culture interact to form a coherent innovation ecosystem in SMEs.
• The authors should clearly indicate how their findings address the research questions posed in the introduction, particularly those concerning the specific challenges faced by Malaysian SMEs, such as limited financial and technical resources.
• The conclusions should place stronger emphasis on how the study enriches the TOE (Technology–Organization–Environment) framework and organizational culture theory in the context of emerging economies.
• This section should conclude with a reflection on how building AI capabilities and transformational leadership contributes to the long-term stability and competitive advantage of SMEs in the digital era.
• Attention should also be paid to a numbering error: the manuscript contains two sections labeled as Section 5 (“Discussion” and “Conclusions”). The conclusions should be presented as a separate sixth section.

Expanding this section would allow readers to better understand why the study matters and what its overarching conclusions are, beyond the statistical results alone.

 

In addition, the manuscript contains minor issues that may suggest haste in its preparation, e.g. inconsistency in abbreviations - while the correct abbreviation NLP for Natural Language Processing is used in the text, the list of abbreviations at the end of the article incorrectly refers to it as “NLH.”

 

Comments on the Quality of English Language

Although the text is generally understandable and written in an academic style, it requires thorough language and editorial revision. The manuscript contains numerous typographical and grammatical errors, as well as inconsistencies that reduce the overall professionalism of the publication.

Below I provide detailed comments on the quality of the English language used in the manuscript:

1. Typographical and punctuation errors (Typos)
   The text contains numerous typographical errors, suggesting a lack of final author proofreading:

• In the abstract, the phrase “analysing tha data” appears instead of “the data.”
• In Section 3.3, the table heading contains an error: “igital organisational culture” (missing the letter “D”).
• In the discussion section, the authors wrote “the results fond that” instead of “found.”
• In the list of abbreviations, Natural Language Processing is labeled as “NLH,” whereas the main text correctly uses “NLP.”

2. Grammatical and syntactic errors
   There are problems with subject–verb agreement and incorrect pronoun usage:

• In Section 1 (Introduction), the sentence “As many of the economies… employ small and medium enterprises (SMEs) as its backbone” uses the singular pronoun “its,” which does not agree with the plural noun “economies.”
• Errors in verb agreement appear, for example: “the efficiency… largely depend on” instead of “depends on.”
• Awkward and incorrect constructions occur, such as “which is mean the study model is reliable” or “which is mean the model have predictive ability.” The correct forms are “which means” and “the model has.”

3. Repetition and redundancy
   In Section 2.2 (Organisational Culture), an entire sentence is duplicated verbatim:
   “Culture therefore helps to avoid fragmentation, conflict and tension.”
   Such copy-and-paste errors should be strictly eliminated.
4. Inappropriate word choice (Diction)
   Some terms are used in an incorrect or non-standard way:

• In the description of the article structure, the authors write “The reminders of the research structure…” whereas the correct term is “remainder.”
• The abbreviation “PC” is used for Computer Vision, which is unconventional; the standard abbreviation in the literature is “CV.”

5. Formatting and consistency

• Headings: As noted earlier in the review, both Section 3.2 and Section 4 are titled “Results,” which disrupts the logical structure of the manuscript.
• Hyphenation: There is inconsistency in the use of hyphens in compound adjectives, for example “vision oriented” (missing hyphen) versus “risk-taking perspective” (correct usage).

Overall, the manuscript requires professional English language editing, preferably conducted by a native speaker or a qualified academic or business/medical language editor.

 

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Major Issues That Need to Be Addressed 

Overall, large parts of the manuscript read as general background rather than theory-driven argumentation. 

1. The manuscript is overly long and lacks focus The literature review and hypothesis development sections remain too detailed and repetitive. Many parts read as general background rather than theory-driven argumentation, and similar definitions and ideas appear multiple times across sections. This weakens the flow of the paper and gives it a thesis-like structure. What is needed here is not minor editing but substantial condensation. The authors should streamline these sections, remove non-essential material, and focus more clearly on the logic that leads directly to the hypotheses. 

2. Concerns about the measurement of transformational leadership persist I continue to have reservations about the way transformational leadership is measured. Using owners and managers to assess their own leadership raises questions about measurement validity, particularly given the well-known limitations of self-reported leadership measures. Although the revised version provides some clarification, the justification is still not fully convincing. The authors should either strengthen the methodological rationale for this approach with clearer theoretical support or discuss its implications and limitations more openly. Without this, it is difficult to place full confidence in the reported findings.

Author Response

Please see the attachment.    

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for making the appropriate corrections to my comments.

The manuscript has been much improved and is in a nice condition now.

I considered that the modifications made improve the quality of the manuscript.

Author Response

Thank you for making the appropriate corrections to my comments.

The manuscript has been much improved and is in a nice condition now.

I considered that the modifications made improve the quality of the manuscript.

Response 1: Thank you for your valuable comment. We greatly appreciate your opinion of the current version. We repeat our thanks.

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors

1. The authors have added a justification for using self-reported transformational leadership measures and expanded the limitations section. This is appreciated. However, please clarify more explicitly that the findings reflect leaders’ self-perceptions rather than follower-rated leadership, and briefly discuss how this may affect interpretation of the results.

2. Some parts remain descriptive. A final round of tightening is recommended to remove remaining redundancy and strengthen the theory-driven flow toward the hypotheses.

3. Please ensure consistency and clarity throughout the manuscript in this final revision.

Author Response

Please see the attachment

Author Response File: Author Response.pdf