Making the Invisible Visible: Understanding the Mismatch Between Organizational Goals and Worker Experiences in AI Adoption

Referee: [—] Methods section: The abstract and summary describe an interview-based study but provide no details on sample size, recruitment procedures, participant selection criteria, interview protocol, data analysis methods (e.g., thematic analysis steps or coding process), or steps taken to address bias and ensure trustworthiness. These omissions are load-bearing because the central claims about barriers and the need for worker-centered approaches rest entirely on the validity of this primary data.

Authors: We agree that the Methods section currently lacks the necessary detail and transparency. In the revised manuscript, we will expand this section to fully describe the sample size, recruitment procedures (via professional networks and targeted outreach in the three sectors), participant selection criteria (professionals with daily AI system interaction), the semi-structured interview protocol, the thematic analysis approach (including coding steps), and measures taken to ensure trustworthiness such as reflexive practices and validation techniques. These additions will directly address the load-bearing nature of the empirical claims. revision: yes

Referee: [—] Findings and Discussion sections: The generalizability of the four barriers and the adaptation strategies to 'AI adoption' broadly is asserted without evidence of thematic saturation, justification for the choice of only three sectors, or discussion of potential sector-specific confounders. The skeptic's concern about over-extrapolation from a finite, opaque sample is not addressed, weakening the claim that these are the primary reasons for adoption failures.

Authors: We accept that the manuscript should more explicitly address scope and limitations. As a qualitative study, we do not claim statistical generalizability but rather transferability of insights. We will add discussion of thematic saturation, justification for selecting healthcare, finance, and management as sectors with prominent AI adoption, acknowledgment of potential sector-specific factors (e.g., regulatory differences), and a dedicated limitations section addressing sample constraints and risks of over-extrapolation. Claims will be qualified to reflect the study context while noting resonance with existing literature on socio-technical barriers. revision: yes

Referee: [—] Discussion section: The adaptation strategies at individual, task, and organizational levels are presented as solutions derived from the interviews, but the manuscript does not explicitly map specific interview themes or quotes to each strategy or provide evidence that they would effectively close the identified gaps. This makes the prescriptive recommendations appear post hoc rather than empirically grounded.

Authors: We acknowledge that the link between data and recommendations could be made more explicit. In the revision, we will add clear mappings in the Discussion, using a structured format (e.g., table or subsection) to connect each adaptation strategy to specific interview themes and representative quotes. We will also elaborate on how the strategies emerged from the data to address the identified gaps, ensuring the recommendations are presented as empirically derived rather than post hoc. revision: yes