Orchestrating AI-assisted code remediation : socio-technical bottlenecks in a large industrial repository
Bexell, Andreas LU
; Gullstrand Heander, Lo
LU
; Söderberg, Emma
LU
and Eldh, Sigrid
(2026)
- Abstract
- Background: Code degradation in large, long-lived codebases is costly to remediate through manual refactoring and opportunistic clean-ups. LLM-based coding assistants can perform mechanical remediation at scale, but their impact on industrial workflows is underexplored. Objective: We investigate how massive AI-assisted code remediation affects build-on-commit continuous integration (CI), code review, and team coordination in a large industrial repository, and which socio-technical bottlenecks constrain such remediation when source editing becomes cheap through AI assistance.
Method: We report on a 15-day exploratory single-case field study in which an experienced developer used a command-line AI coding buddy to remediate... (More) - Background: Code degradation in large, long-lived codebases is costly to remediate through manual refactoring and opportunistic clean-ups. LLM-based coding assistants can perform mechanical remediation at scale, but their impact on industrial workflows is underexplored. Objective: We investigate how massive AI-assisted code remediation affects build-on-commit continuous integration (CI), code review, and team coordination in a large industrial repository, and which socio-technical bottlenecks constrain such remediation when source editing becomes cheap through AI assistance.
Method: We report on a 15-day exploratory single-case field study in which an experienced developer used a command-line AI coding buddy to remediate widespread issues in a closed-source industrial C++ repository. We triangulate Gerrit metadata with a developer diary and team chat, analyzed through descriptive statistics and qualitative coding. Results: AI-assisted remediation rapidly generated hundreds of commits touching thousands of lines, saturating CI and reviewer attention. Naïve per-file commits overloaded build-on-commit CI; Switching to directory-based batching and capping the number of files per change restored throughput, but still required explicit review solicitation, negotiation of acceptable commit granularity, and iterative follow-up to resolve build and static-analysis failures.
Conclusion: When mechanical editing is cheap, CI capacity, review effort, and change orchestration become primary bottlenecks. Sustainable AI-assisted remediation in very large repositories requires deliberate control of commit, review, and CI batch granularity and treating semantic change sets, such as ``fix all instances of warning X'', as first-class units of work that can be sliced differently for developers, reviewers, and CI. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/ff6b2ed2-bc9f-4c52-89c5-bd421d2fc105
- author
- Bexell, Andreas
LU
; Gullstrand Heander, Lo
LU
; Söderberg, Emma
LU
and Eldh, Sigrid
- organization
- publishing date
- 2026-09-24
- type
- Working paper/Preprint
- publication status
- submitted
- subject
- pages
- 16 pages
- DOI
- 10.48550/arXiv.2609.29172
- language
- English
- LU publication?
- yes
- id
- ff6b2ed2-bc9f-4c52-89c5-bd421d2fc105
- date added to LUP
- 2026-10-01 15:15:16
- date last changed
- 2026-10-05 13:14:23
@misc{ff6b2ed2-bc9f-4c52-89c5-bd421d2fc105,
abstract = {{Background: Code degradation in large, long-lived codebases is costly to remediate through manual refactoring and opportunistic clean-ups. LLM-based coding assistants can perform mechanical remediation at scale, but their impact on industrial workflows is underexplored. Objective: We investigate how massive AI-assisted code remediation affects build-on-commit continuous integration (CI), code review, and team coordination in a large industrial repository, and which socio-technical bottlenecks constrain such remediation when source editing becomes cheap through AI assistance.<br/><br/>Method: We report on a 15-day exploratory single-case field study in which an experienced developer used a command-line AI coding buddy to remediate widespread issues in a closed-source industrial C++ repository. We triangulate Gerrit metadata with a developer diary and team chat, analyzed through descriptive statistics and qualitative coding. Results: AI-assisted remediation rapidly generated hundreds of commits touching thousands of lines, saturating CI and reviewer attention. Naïve per-file commits overloaded build-on-commit CI; Switching to directory-based batching and capping the number of files per change restored throughput, but still required explicit review solicitation, negotiation of acceptable commit granularity, and iterative follow-up to resolve build and static-analysis failures. <br/><br/>Conclusion: When mechanical editing is cheap, CI capacity, review effort, and change orchestration become primary bottlenecks. Sustainable AI-assisted remediation in very large repositories requires deliberate control of commit, review, and CI batch granularity and treating semantic change sets, such as ``fix all instances of warning X'', as first-class units of work that can be sliced differently for developers, reviewers, and CI.}},
author = {{Bexell, Andreas and Gullstrand Heander, Lo and Söderberg, Emma and Eldh, Sigrid}},
language = {{eng}},
month = {{09}},
note = {{Preprint}},
title = {{Orchestrating AI-assisted code remediation : socio-technical bottlenecks in a large industrial repository}},
url = {{http://dx.doi.org/10.48550/arXiv.2609.29172}},
doi = {{10.48550/arXiv.2609.29172}},
year = {{2026}},
}