All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Dear authors, we are pleased to verify that you meet the reviewer's valuable feedback to improve your research.
Thank you for considering PeerJ Computer Science and submitting your work.
Kind regards
PCoelho
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
The manuscript has been greatly improved and now reads as a well-structured and polished paper. The introduction is detailed and clearly outlines the research gap, supported by up todate references. The figures and tables are relevant and clearly presented. The language is fluent, technical terms are used appropriately, and the flow between sections is smooth. While the introduction remains long, it provides valuable context and can be retained in its current form
The study design is solid and fits the journal’s scope. The selection of the 20 SMEs is now well explained, and the model structure, including its components and workflow, is clearly described. The paper demonstrates transparency through its use of data tables and an algorithmic explanation. The addition of the Ethical Considerations section, which addresses data privacy and bias mitigation, demonstrates an attention to responsible research practices.
The findings are clearly presented and well supported by the results. The inclusion of sensitivity analysis provides a clear understanding of how decisions vary with different parameters. The claim of higher accuracy has been suitably rephrased as preliminary evidence, with benchmarking suggested for future studies. The conclusions are realistic and consistent with the study’s objectives. The results are now convincing and meaningful for the intended audience.
The authors have made paid attention to revisions that address all previous reviewer concerns. The manuscript is now cohesive and complete in all aspects. It makes a clear and practical contribution to research on decision support and technology adoption in small and medium enterprises/ SMES
Dear authors,
Thanks a lot for your efforts to improve the manuscript.
Nevertheless, some concerns are still remaining that need to be addressed.
Like before, you are advised to critically respond to the remaining comments point by point when preparing a new version of the manuscript and while preparing for the rebuttal letter.
Kind regards,
PCoelho
-
-
-
The authors have addressed the key points raised in the first round of revision. However, the article requires thorough editing. Additionally, the introduction is excessively long and should be shortened and focused.
The manuscript has improved considerably since the first review. The introduction is now clearer, with the research gap explicitly stated and well supported by recent references. The literature review is broader and better grounded, which helps to situate the study in the wider field. The addition of a figure illustrating the proposed model strengthens the clarity of the presentation.
Tables showing the scoring process and sensitivity analysis are now included and well presented, and the supplementary data adds to the transparency of the work. The language reads smoothly, earlier typographical and formula errors have been corrected, and the overall presentation is professional.
The Experimental design is relevant and timely. The authors have clarified the selection process for the 20 SMEs, including sector and size distribution, which helps to support the credibility of the dataset. The algorithm is more fully explained and is accompanied by pseudo-code and supporting descriptions that make the approach easier to follow.
Optional: Reproducibility could be improved through the sharing of code or data repositories, the current level of detail is adequate for publication in this journal. The overall design is sound and appropriate for the research aims.
The results are logically presented. The tables and sensitivity analysis show how the model works across different settings, and the case study on cloud vendor adoption provides a practical example of application.
Some limitations remain, such as the lack of direct benchmarking against established models and the absence of open data or code. However, these can be viewed as areas for future work rather than reasons to delay publication. The conclusions are consistent with the findings and are framed realistically, with reference to further applications in other domains.
The authors have responded well to the initial feedback and the manuscript has been substantially strengthened. While there are still areas that could be further developed in future research, the current version is clear, relevant, and contributes meaningfully to the field. This revised manuscript is much stronger than the previous version. The authors have clearly worked to address the main concerns raised earlier. The paper is clear, well structured, and makes a useful contribution to the literature on decision support for SMEs.
For final polishing, I would suggest only two small improvements:
Add a short note on ethical considerations, particularly data privacy and potential bias.
Rephrase the claim of “higher accuracy” so that it is presented as preliminary evidence, with benchmarking suggested for future studies.
These are minor adjustments that do not affect the core value of the paper.
Dear authors,
You are advised to critically respond to all comments point by point when preparing an updated version of the manuscript and while preparing for the rebuttal letter. Please address all comments/suggestions provided by reviewers, considering that these should be added to the new version of the manuscript.
Kind regards,
PCoelho
The paper addresses a timely and relevant topic on human-AI collaboration in SMEs and has strong potential to contribute to digital transformation and decision support literature. The proposed HAI-HDM model is promising, but the manuscript requires clearer identification of the research gap, stronger theoretical grounding with updated references, and improved structure (citations, figures, and tables):
Research Gap: The introduction does not clearly identify the research gap. It is important to explicitly highlight what prior studies have not addressed and how this study aims to fill that gap.
Literature Support: The literature reviewed in the introduction is very limited. Please expand it by incorporating more recent and relevant studies to strengthen the theoretical foundation.
Formatting: Avoid using double dashes (“--”), as they resemble ChatGPT outputs and may appear unprofessional. Instead, use proper punctuation such as commas, semicolons, or parentheses.
Tables: All tables must be embedded within the text at the appropriate locations rather than placed separately at the end of the manuscript.
Model Visualization: Include a figure illustrating the Human-AI Hybrid Decision Model (HAI-HDM). A visual representation of the framework will improve clarity and allow readers to easily understand the model’s structure and functioning.
Citations: Several paragraphs are written without appropriate citations. Ensure that every key claim, argument, or theoretical statement is backed by relevant references.
Overall, the paper is well written and easy to follow. The structure is clear, and the flow from the introduction, through the literature review, to the model and case study makes sense. The language is professional, though a few places could be tightened to avoid repetition. For example, the introduction repeats the limitations of AI (lack of context, explainability) and human judgment (biases, slow processing) in several paragraphs—this could be made more concise.
The literature review does a good job of setting the stage, drawing on both the classical models (DOI, TAM, TOE) and newer work on AI in SMEs. The references are up to date, though adding one or two more recent surveys on SME technology adoption would help.
The model acronym (HAI-HDM) is introduced more than once. After the first definition, it would be clearer to stick to the abbreviation. I also noticed a few typographical issues—most notably in the final score formula, where it reads “(12³)” instead of “(1–γ).” Minor formatting errors, such as this, should be corrected before publication.
The proposed Human-AI Hybrid Decision Model is an interesting concept that aligns well with the journal’s scope. I especially liked the inclusion of explainable AI techniques (SHAP and LIME), which are very relevant for SMEs that may be skeptical of black-box models. The adaptive weighting between human expertise and AI confidence is also a novel and practical idea.
That said, there are a few gaps in the design. The case study utilizes data from 20 SMEs; however, it’s unclear how these companies were selected or whether they are representative. Providing more details here would help the reader assess the reliability of the findings. Also, while the pseudo-code is helpful, there isn’t enough information to reproduce the study—no code, no clear description of preprocessing, and no mention of cross-validation. Sharing the dataset or at least describing it in more detail would strengthen the paper considerably.
Finally, the paper makes a strong claim about achieving “higher accuracy” than baseline models, but doesn’t provide any comparative metrics. Even a small benchmarking exercise against something like TAM or AHP would make the evaluation more convincing.
The findings are logical and well-aligned with the research questions. The case study is realistic and shows how the model could be applied in practice—for example, the Azure recommendation makes sense given the cost, compliance, and performance balance. The sensitivity analysis is also a good touch, as it demonstrates that the results hold up when the AI-human weighting changes.
However, the evaluation is still simulation-based, and the dataset is small. This makes it hard to generalize beyond the example given. The authors could strengthen their conclusions by either piloting the model in a real SME setting or by expanding the case study to cover more domains. Additionally, while the sensitivity analysis is helpful, the absence of confidence intervals or statistical measures leaves the reader uncertain about the robustness of the results.
• Trim down the introduction to avoid repeating the same points.
• Provide much more detail on the dataset and methods, and ideally share the code and data.
• Add a proper comparative evaluation against baseline approaches.
• Fix minor errors in formulas and tables.
• Include a short note on ethical considerations—especially data privacy and bias in AI recommendations.
The paper has a lot of potential. The idea is strong and the framework is well thought-out, but the lack of methodological transparency and rigorous validation means it’s not quite ready for acceptance yet.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.