Tell me about a time you used LLMs to improve your efficiency at work
Try This Question Yourself
Practice with feedback and follow-up questions
What is this question about
Interviewers are usually not just checking whether you have used AI tools. They want to see whether you can identify a real work bottleneck, apply a new tool with sound judgment, and translate that into credible impact. Strong answers also show restraint: you understood where LLMs helped, where they did not, and how you verified quality rather than treating the model as magic.
Key Insights
- You should make the problem feel real before you make the tool feel impressive. 'I used ChatGPT a lot' is weak; 'I was spending hours on repetitive investigation, drafting, or test-writing, so I tried a focused workflow' is much stronger.
- You need to show judgment, not just enthusiasm. Good answers explain what you trusted the LLM to do, what you still verified yourself, and any privacy, correctness, or security boundaries you respected.
- Efficiency alone is not enough if it only made your own life easier in a vague way. Tie the usage to a meaningful outcome such as faster delivery, better coverage, clearer communication, or more time spent on higher-value work.
What interviewers probe atlevel
Top Priority
A strong junior answer shows you did not blindly trust the model—you used it as an assistant and checked the work yourself.
Good examples
🟢I used the generated test ideas as a draft, but I ran them, fixed incorrect assumptions, and removed a couple that didn't match our actual behavior.
🟢I asked it for an explanation of the library and then verified the relevant parts in our internal docs and code before changing anything.
Bad examples
🔴I let it generate the tests and then I submitted them since they looked correct.
🔴I mostly trusted the answer because it explained the code path clearly, and it turned out to be good enough.
Weak answers outsource thinking; strong answers use the LLM to accelerate work while keeping responsibility for correctness.
Valuable
Example answers atlevel
Great answers
On a small backend task, I had to add several unit tests around input validation, and I noticed I was spending a lot of time thinking through repetitive edge cases. I used an LLM to generate a list of likely boundary conditions and a rough test structure, then I checked each one against the actual code and our existing test patterns. A few suggestions were wrong for our implementation, so I removed those and rewrote the assertions myself. That cut the test-writing part from a few hours to about one hour, and because I had extra time, I also added one case around empty input that I might have missed otherwise. The main thing I learned was that it was much more useful for brainstorming and drafting than for final code I could trust directly.
At my last job, I spent a lot of time turning meeting notes and customer emails into clean status updates for my manager and the sales team. I started using an LLM to summarize long threads, pull out action items, and draft a first pass of the update in our usual format. I still read everything carefully and edited the tone and details myself, but it saved me from rewriting the same information over and over. It probably cut that work in half, and it also made my updates more consistent because I used the same structure every time. What I liked most was that it freed me up to spend more time following up on the actual customer issues instead of just documenting them.
Poor answers
I use LLMs pretty regularly now because they make coding faster. One time I had to write tests, so I asked it to generate them and then I pasted them into my branch with a couple of small edits. It saved me a lot of time and let me finish the task quickly. Since then I've relied on it whenever I want to speed up development.
Question Timeline
See when this question was last asked and where, including any notes left by other candidates.
Early June, 2026
I was asked to give examples solely from my work experience. Since LLMs became mainstream only in 2023, I was allowed to answer this question with ideas for prospective improvements alongside non-LLM improvements I have made at work.
Hello Interview Premium
Your account is free and you can post anonymously if you choose.