Enriching Human–AI Collaboration: The Ontological Service Framework Leveraging Large Language Models for Value Creation in Conversational AI

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Overall 
The manuscript demonstrates strong originality and practical value. With moderate revisions to improve clarity, polish the English, and expand the validation discussion, it has clear potential for publication in Knowledge (MDPI).

Areas for Improvement

1. Clarify Novelty and Contribution – Strengthen the introduction and discussion by explicitly differentiating this work from existing ontology-based or LLM-only systems. A short paragraph summarizing the unique integration would make the contribution more visible.

2. Figures and Presentation – Ensure figure labels and references are consistent and legible (especially Figures 2–5).

3. Ethical and Practical Considerations – Consider adding a brief discussion of data privacy, LLM bias, and clinical accountability in AI-supported healthcare systems.

Comments on the Quality of English Language

Language and Structure – The manuscript would benefit from careful language editing to improve clarity, reduce redundancy, and unify terminology across sections.

 

Author Response

We have updated the manuscript and addressed the comments. Please find the attached file for the detailed feedback. 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The introduction is too much. In particular, since the theoretical background is being written, I think it will be enough to reduce the amount by two-thirds. Please summarize the introduction a little more by deleting parts that are not appropriate in the logical context.

The research method was properly structured.

In the results of the study, figures 5, 6, and 7 are less readable. 
In addition, the results are presented with 10 chapters from 4.1 to 4.10. If not all of the contents are necessary, it seems appropriate to derive the results only from the hypotheses and research topics mentioned in the necessity of training. Explaining too much content is not a good study.

The discussion was properly organized, but it is regrettable that the topics are not distinguished. It seems much more readable to develop the discussion by deriving about three topics of discussion.

Overall Opinion: The research results were drawn on meaningful topics, but there was a lot of difficulty in understanding the results because too much content was listed. Please reduce the number of pages by deleting unnecessary parts. Thank you for your hard work.

Author Response

We have addressed the valuable feedback. Our response is attached in the file. 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

General Assessment
The manuscript addresses a highly relevant and timely topic in the intersection of ontology-driven approaches and large language models (LLMs) for conversational agents in healthcare. The authors present a well-structured study, grounded in Design Science Research Methodology (DSRM) and customized ontology engineering methods, supported by a case study from the Karolinska University Hospital.

The contribution is original, methodologically sound, and of interest to both researchers and practitioners in digital health and AI-assisted clinical workflows.

Strengths

• Innovative integration of ontologies and LLMs within the proposed Service-oriented Human-AI Collaborative Framework (SHAICF).

• Rigorous methodological approach with detailed elaboration of ontology development and evaluation.

• Strong practical relevance, demonstrated through the case study and prototype (MediBot).

• Clear structure and logical flow of arguments, with results linked to research questions.

Points for Improvement

1. Clarity and readability: Some sections (especially methodology) are overly detailed and stylistically complex. Consider simplifying the language for broader readership.

2. Empirical validation: While the technical evaluation is solid, inclusion of preliminary pilot data from clinical settings would further strengthen the contribution.

3. Discussion of limitations: Expand the reflection on practical implementation challenges (legal, ethical, interoperability issues).

4. Figures: Some visual materials could be simplified to emphasize the key messages more clearly. I suggest either providing more concise and targeted explanations for each figure, or considering the removal of those that do not substantially contribute to the manuscript’s main arguments.

Author Response

We have addressed the feedback and updated as mentioned in the attached file

Author Response File: Author Response.pdf