Understanding AI Assistance in Professional Software Engineering - Replication Package
Authors/Creators
-
1.
IT University of Copenhagen
- 2. IT University Of Copenhagen
Description
Dataset Overview
The dataset consists of qualitative data from a semi-structured interview study with 13 senior software developers(each with more than five years of experience and daily use of AI coding assistants), recruited from seven companies spanning finance, shipping, biotech, media, and environmental sectors (large international Danish enterprises to a US startup). Interviews were conducted online between March and April 2026, audio-recorded with participant consent, transcribed, and fully de-identified.
The study explored how senior developers integrate AI assistance (e.g., GitHub Copilot, ChatGPT, Claude, Gemini, Cursor) into their development environments and workflows. Thematic analysis (Braun & Clarke) yielded four themes: AI Collaboration, AI Integration, Mental Models, and Workflow.
Contents
| File | Description |
|---|---|
Interview_P01.md – Interview_P13.md |
De-identified transcripts of the 13 semi-structured interviews (Markdown) |
Pre_interview_questionnaire.csv |
Pre-interview questionnaire responses: demographics, programming experience, and AI tool usage per participant |
Interview_guide.pdf |
The semi-structured interview guide, organized in three sections: (a) AI familiarity and background, (b) coding workflows via recent task walkthroughs, (c) AI assistance — when it is used, how outputs are handled and validated |
Code_transcript.xlsx |
Coding spreadsheet from the thematic analysis: codes (semantic and latent) derived from participant responses, with traceability to interview sections |
Notes on Data Handling
- All names and personal details were removed from the transcripts to maintain participant confidentiality.
- Transcripts were reviewed against the recordings for accuracy, and a cleaning pass removed speech disfluencies (e.g., filler words) while preserving natural speech flow.
- Questionnaire email addresses are redacted in the CSV.
Citation
Please cite the paper when using this dataset:
M. Y. Hristova, R. G. Scaramuzza, C.-E. Popovici, P. Burelli, and P. Tell, "Understanding AI Assistance in Professional Software Engineering: An Interview Study of Development Practices and Workflows," In: Product-Focused Software Process Improvement. PROFES 2026.
Files
Files
(75.6 kB)
| Name | Size | Download all |
|---|---|---|
|
md5:a20f663864b20129ef4e1adefeaac65d
|
75.6 kB | Download |
Additional details
Dates
- Available
-
2026-09-22