AI-Driven Personalization in Marketing Administration: Qualitative Insights from European Professionals

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The research presents an original perspective on the issue of AI implementation in marketing management. The challenges were diagnosed based on an in-depth research using qualitative methods. The work supports the general discussion around the potential of AI-human collaboration from the perspective of senior managers. The directions of future research were diagnosed correctly.   

Author Response

The research presents an original perspective on the issue of AI implementation in marketing management. The challenges were diagnosed based on an in-depth research using qualitative methods. The work supports the general discussion around the potential of AI-human collaboration from the perspective of senior managers. The directions of future research were diagnosed correctly.

Dear Reviewer,

Thank you for your constructive evaluation and for highlighting the manuscript’s strengths. The final version of the paper has been uploaded to the platform.

Reviewer 2 Report

Comments and Suggestions for Authors

General. The study is interesting since it uses phenomenological qualitative research to understand individual perspectives by exploring the meaning of lived experiences in the high-tech environment, such as AI. It also employs a dual framework: phenomenological inquiry and ANT (actor-network theory).

Abstract.  The abstract describes the research objective and the methodology but does not report the results.

Sampling. The purposive sample comprises 36 senior marketing executives from sectors including tourism, fintech, professional services, and digital media. There is no sufficient reason why only these 4 sectors represent the heterogeneity.

Literature gap. The second paragraph of section 2 ends with a statement identifying a literature gap. However, this conclusion is premature because no in-depth literature review or meta-analysis is cited as supporting evidence.

Limitations of the research. There is no discussion on this issue.

Comments on the Quality of English Language

NA

Author Response

Dear Reviewer,

Please refer to the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Author(s),

I found the study to be timely, relevant, and thoughtfully designed. By combining interpretive phenomenology with Actor-Network Theory (ANT) and drawing on rich qualitative data from 36 senior executives across diverse European service sectors, the paper offers meaningful insights into how AI-driven personalisation is reshaping marketing administration. The manuscript clearly has the potential to advance qualitative and socio-technical debates on AI in marketing.

That said, I believe the paper would benefit from revisions to enhance its quality:

Introduction and Theoretical Background

1. While the introduction provides a comprehensive overview of AI-driven personalisation and its benefits, the specific theoretical and empirical gap addressed by this study could be articulated more explicitly. The authors are encouraged to clearly distinguish how their interpretive phenomenological and ANT-based approach advances existing qualitative research beyond identifying ethical concerns, for example by specifying what new insights into marketing administration practices, governance, or decision-making networks emerge from their analysis.
2. The integration of ANT is well motivated; however, the paper would benefit from a more explicit explanation of why ANT is particularly suitable for examining AI-driven personalisation in marketing administration compared to alternative socio-technical frameworks (e.g., sociomateriality or institutional theory). Clarifying how ANT concepts (e.g., translation, enrolment, stabilisation) guide data collection and analysis would improve theoretical coherence.
3. The term “marketing administration” (p. 2, line 68) is central to the study but remains somewhat broad. The authors could enhance clarity by specifying the administrative functions under examination (e.g., campaign governance, CRM oversight, compliance, interdepartmental coordination). This would help readers better understand the empirical boundaries of the study and how AI-driven personalisation operates within these organisational contexts.
4. Although ethical concerns and regulatory frameworks (GDPR and the EU AI Act) are acknowledged, their role within the ANT networks could be more systematically theorised. The authors may consider treating regulations and ethical norms as non-human actors that shape network stability and resistance, thereby deepening the analysis of governance, trust, and accountability in AI-enabled marketing practices.

Materials and Methods

The methodological pairing of interpretive phenomenology and ANT is conceptually coherent and well motivated for socio-technical inquiry; however, methodological execution requires clarification:

1. Define the analytic approach precisely (e.g., Braun and Clarke’s reflexive thematic analysis with ANT sensitizing concepts) and report coding procedures, codebook development, intercoder training, and reliability metric with statistics, confidence intervals, and the unit of analysis.
2. Replace “theoretical saturation” with a more appropriate construct for phenomenology (e.g., information power) and justify sample size via role heterogeneity, case richness, and analytic aims.
3. Provide ethics/IRB approval details, consent process specifics, anonymization, storage, and GDPR-compliant research handling of transcripts and metadata.
4. Explain language/translation approach for interviews across countries (interpreters, back-translation, or English-only) and consequent interpretive limitations.

Results

1. Inclusion of verbatim quotes for each theme (2–3 per theme, across sectors) to evidence claims, along with disconfirming/negative cases to avoid confirmation bias.
2. Greater sectoral granularity: Where do fintech vs tourism differ on fairness guardrails? Are UK vs EU mainland organizations differentially responsive to AI Act readiness?
3. Thick vignettes on “override moments,” DPIA delays, and risk-tiering decisions that concretize co-agency and compliance-by-design beyond summary paraphrases.
4. Clearer reporting of the 91% agreement: the number of coders, initial independent agreement before consensus, statistic type, and how disagreements were adjudicated.

Discussion

1. Incorporate uplift modeling and causality-aware personalization. Your Theme 1 (co-agency) and Theme 3 (fairness/opacity) intersect with constrained uplift and ranking-oriented evaluation; acknowledging the fragility to preprocessing and distribution shift (DUMOM) would strengthen your Infrastructure pillar.
2. Connect Theme 3 and Responsibility pillar to frameworks like COFFEE (e.g., Wang et al. (2022); https://doi.org/10.48550/arXiv.2210.15500), which operationalizes counterfactual fairness for personalized text explanations, relevant to explainable marketing content and to balancing quality-fairness trade-offs.
3. Situate your observations about generative tools with surveys on personalized image generation and generative storytelling in marketing to ground the “Capability” pillar in concrete modality-specific challenges (IP, safety, bias amplification).
4. Position AI-MARC relative to NIST AI RMF and ISO/IEC 42001, and to operational frameworks like AI TIPS 2.0. A mapping table (even in appendix) showing complementarities and what AI-MARC uniquely contributes (e.g., marketing-specific role archetypes, consent instrumentation) would bolster originality and utility. What is uniquely contributed for marketing versus general AI governance?
5. Nuance the rights-based framing with comparative governance analysis, which argues EU AI governance is managerial; this could enrich your Theme 2 discussion on how organizations experience regulation as risk containment rather than purely rights realization.

Limitations & Clarity

1. How did you mitigate the risk of social desirability or self-exoneration in senior executives’ narratives, especially around ethics committees and guardrails (e.g., triangulation with documents, incident logs, or audit records)?
2. Some titles/phrases (“insight ontologists,” “generative semioticians”) are unconventional for job roles/tools. Are these verbatim labels, metaphors, or analytic constructs? Please clarify to avoid misinterpretation.

These revisions would, in my view, significantly enhance the rigor and impact of the study.

Author Response

Dear Reviewer,

Please refer to the attachment.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Author(s),

Thank you for submitting the revised version of your manuscript entitled “AI-Driven Personalisation in Marketing Administration: Qualitative Insights from European Professionals.”

I appreciate the effort you have made to address my previous comments and concerns. The revisions have clearly strengthened the manuscript, and the overall quality has improved noticeably. The paper is now more coherent and refined, which enhances its contribution to the literature.

Thank you again for your careful revisions and responsiveness to the feedback provided.