Check Your CI/CD Pipelines for AI-Generated Review Bottlenecks
New data shows AI tools are slowing down software delivery reviews, creating new challenges for teams managing automated build and test systems.
Key points
- CI/CD remains the most adopted DevOps practice, used by 83.4% of respondents in a JetBrains survey.
- Nearly half of organizations use AI for CI/CD tasks, but only 8% run AI steps directly in builds.
- On-premises hosting for CI/CD systems has reached a five-year high, driven by security and control needs.
Software delivery teams are facing new delays because artificial intelligence tools are creating bottlenecks during the code review process, according to a recent report. JetBrains published The State of CI/CD in 2026, which highlights how AI integration is changing workflows. While automation remains central to development, the addition of AI-generated changes is complicating the verification steps that teams rely on to ship software safely and quickly.
In plain English
Continuous Integration and Continuous Delivery, known as CI/CD, are automated systems that test and deploy code. When developers save changes, these systems automatically run tests to ensure nothing breaks. If the tests pass, the software moves closer to release. This automation allows teams to update applications frequently without manual errors. It is the backbone of modern software development, ensuring that new features are stable before they reach users.
The background
The JetBrains report found that CI/CD is the most widely used practice among DevOps teams, with 83.4% of respondents running these pipelines. This adoption rate is higher than for containerization, infrastructure as code, GitOps, and platform engineering. GitHub Actions is the leading tool for organizations, followed by GitLab CI and Jenkins. Azure DevOps Server and TeamCity also see significant use. AWS CodePipeline, Bitbucket Pipelines, and custom systems round out the landscape.
Tool choices are often driven by client demands and legacy systems. Organizations using multiple CI/CD tools cite different customer requirements and technology stacks as reasons. Departments working with specific clients may need to use tools mandated by those customers. Most developers have no plans to switch their primary CI/CD tools. Those considering a change point to scaling costs, maintenance effort, tool consolidation, and configuration complexity as key factors.
What changes now
Hosting preferences are shifting toward greater control. The survey shows that 72.8% of respondents run their primary CI/CD system on their own premises or self-managed cloud infrastructure. This marks the highest share for on-premises installations in the five years covered by the study. Teams choose this model to maintain control over hardware, networks, data, build environments, security, and costs. This trend supports a mixed deployment model rather than a full move to public cloud services.
AI usage in CI/CD is growing but remains limited in scope. Forty-eight point eight percent of surveyed organizations use AI for at least one CI/CD-related task. However, only 8% report running AI-powered steps directly within a build or test. Most AI use happens around the pipeline, helping teams review code, write or fix tests, and diagnose build failures. Other applications include maintaining pipeline configurations, quality checks, deployment assistance, and incident response.
Developers access CI/CD tools through Model Context Protocol servers and agent skills, connecting assistants to their workflows. Web interfaces remain the most common way respondents interact with these systems. Olga Bedrina, Product Marketing Manager for TeamCity, noted that self-hosting remains important for organizations needing control over infrastructure and security. She added that most AI use currently happens around the pipeline rather than inside it.
Key terms explained
Teams usually manage software delivery through automated pipelines that test code changes before deployment. When AI tools generate code or updates, human reviewers must verify these changes carefully. Check whether your organization’s review process can handle the volume of AI-generated modifications. Ensure that security checks are not skipped to speed up reviews. Monitor build times to identify if AI integration is causing unexpected delays in your release cycle.
What to do and how to stay safe: JetBrains
- Review your CI/CD pipeline configurations to ensure AI-generated changes are subject to the same rigorous testing as manual code.
- Monitor build and test durations to identify if AI integration is introducing latency or bottlenecks in your delivery process.
- Verify that your organization’s security policies apply to AI tools accessing your development environment and pipeline data.
- Evaluate whether your current hosting model provides sufficient control over the data and security risks associated with AI usage.
Step-by-step guide: Stop AI Chatbots Leaking Sensitive Business Data
General security guidance from the Vector Update newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
Which CI/CD tool is most widely adopted by organizations?
GitHub Actions leads organizational adoption, followed by GitLab CI and Jenkins, according to the JetBrains report.
How many organizations run AI steps directly within their builds?
Only 8% of surveyed organizations report running AI-powered steps directly within a build or test.
Why are more teams choosing on-premises CI/CD hosting?
Teams choose on-premises hosting to maintain control over hardware, networks, data, build environments, security, and costs.
