“How do you use AI in your daily workflow?”
This is a question, you are most certainly asked at DevOps interviews in 2026.
If you answer, “I use it to write my scripts,” you risk sounding like someone who blindly copies and pastes code. Instead, you need to frame your answer to show that your role is shifting from someone who writes every line of code to a “Principal Code Reviewer”.
Here is how to structure your response to show you use AI as a strategic tool, not a crutch.
1. Discuss System Architecture (With Caveats)
-
What to say: Explain that you use an LLM as a sounding board for high-level system design reviews. You can mention feeding it a proposed architecture to evaluate the design against established standards, like the AWS Well-Architected Framework, quickly mapping out trade-offs regarding cost, reliability, and security.
-
The crucial caveat to add: Emphasize that AI lacks business context. Tell the interviewer that if you ask AI for a reliable architecture, it often defaults to the most robust, enterprise-grade solution available in its training data, which frequently leads to severe over-engineering for smaller projects.
-
Example to share: You can mention prompting the AI to generate a comparison matrix showing technical metrics and maintenance overhead between running a custom monitoring stack on EC2 versus using a managed service. Ultimately, you make the final architectural decision based on your team’s bandwidth and budget constraints.
2. Explain Your Approach to Infrastructure as Code
-
What to say: Mention that you use AI to generate base templates for cloud environments using Terraform, OpenTofu, or Ansible. It eliminates the friction of starting from a blank page.
-
The crucial caveat to add: Highlight that AI frequently hallucinates generic resources or uses legacy directory structures, making your domain knowledge essential for ensuring the code is safe for production.
-
Example to share: Explain how you act as the reviewer when AI suggests a generic
aws_prefix_listfor a Terraform data source, stepping in and correcting it to the requiredaws_ec2_managed_prefix_list.
3. Talk About API Automation and Prototyping
-
What to say: Discuss using AI as a fast-drafting partner for custom internal utilities, such as writing Python API wrappers.
-
The crucial caveat to add: Point out that while AI is great for syntax structure, it often misses the nuanced execution context and specific payload requirements of third-party systems.
-
Example to share: Describe generating a Microsoft Graph API wrapper script where the AI provided a beautifully structured script, but you had to step in and enforce explicit requirements, such as ensuring a mandatory user ID parameter is actually passed so the API call executes successfully.
4. Frame AI as a Contextual Debugger
-
What to say: Explain how pasting error traces into an LLM transforms it into an interactive, highly-contextual debugger for CI/CD pipelines and shell scripts. It helps you locate variable validation bugs or logic errors that are easily missed by human eyes.
-
The crucial caveat to add: Note that AI might suggest a fix that technically executes but violates the framework’s intended error handling or control flow.
-
Example to share: Share a scenario where AI suggested a generic halt command for a failing script, but you knew the specific automated workflow engine required you to utilize the proper
ExecuteErrorclass to handle the failure correctly.
The Bottom Line for Your Interview To nail this question, your overarching message should be clear: The value of a modern systems engineer isn’t in writing boilerplate from memory, but in architectural strategy and verifying execution context. Embracing AI doesn’t replace you; it elevates you to the role of Principal Reviewer.
The examples mentioned in this post are from my own experience. If you are facing an interview, be prepared with examples based on your experience. Practice them before the interview so that you can answer clearly. and confidently.