I'm sure you're familiar with "vibe coding"—that creative process of building things based on a feeling, without a strict plan. It can be fun, but in a team environment, it simply doesn't work. It’s hard to scale, new people get lost, and the results are often messy.
For me, things have changed. I now focus on working with clear specifications, and I use AI as my co-pilot. Many people think using AI means you lose control, but I've found the opposite is true. When used correctly, AI is a powerful partner that helps you apply your knowledge more effectively, giving you more control, not less.
Here are the six main ways I use AI as a tech lead.
1. Supercharging My Git Workflow
I've used Git for a long time, so I'm comfortable with it. But being comfortable with a tool doesn't mean you're using it efficiently. This is where AI, especially GitHub Copilot, has made my daily Git work seamless.
Instead of typing every command by hand, I let AI handle the repetitive parts. It suggests sharp, accurate commit messages based on my changes, and most of the time they're perfect with just a few small tweaks. It can also create branch names for me based on the user story, which keeps everything consistent.
But the real game-changer is writing pull request descriptions. By connecting AI to our Azure DevOps board, it can generate a comprehensive summary of changes, including the rationale behind them, and automatically link to the corresponding user story. This frees me from tedious tasks so I can focus on what really matters: the quality of the code itself.
2. The AI-Powered Second Pair of Eyes: Code Reviews
As a tech lead, I frequently conduct code reviews for multiple teams, which means I don't always have the full context. Previously, it took a lot of time to get up to speed. Now, AI is my starting point.
I can ask it to explain changes in a pull request, identify potential problems I might have missed, or verify that code adheres to our team's coding rules. As long as our rules are written down, AI can handle a large part of the review, and it does it much faster than I ever could.
This approach enables me to focus on the most critical aspects, such as the core logic and architecture. It also helps me give better, fairer feedback. For instance, I can ask AI to "explain why we use dependency injection here." That simple request turns a standard comment into a valuable teaching moment for the whole team.
3. Staying in the Zone with AI and Azure DevOps
Connecting an AI assistant like GitHub Copilot to your project board—whether it's Azure DevOps, Jira, or GitLab—is a massive improvement for any developer's workflow. I can stay in my code editor—my focus zone—and still have a direct connection to our team's plan.
Through my editor, I can request additional information on a user story, seek help in understanding a new bug report, or even write and post comments for product owners. It effectively creates a bridge between the project plan and the actual coding work, ensuring that what I'm building is aligned with the team's goals and that everything is tracked properly.
4. Taming the CI/CD Beast
After years of building CI/CD pipelines in Azure DevOps, I’ve developed a solid understanding of how they work. But every so often, you run into a problem that’s hard to figure out.
"I remember spending two hours debugging a pipeline once, only to find it was a single misplaced character. Last week, a similar issue came up, and by pasting the error into Copilot Chat, I found the fix in less than two minutes."
My process for fixing these issues has completely changed. I just copy the strange error message from the pipeline, paste it into the chat with the AI, and since the AI can read my *.yml
file, it often finds the problem right away. It’s this teamwork—my experience guiding the AI’s powerful pattern-matching—that helps me fix problems in a fraction of the time.
5. My Go-To Brainstorming Partner
Honestly, brainstorming is the main way I've used AI from the very beginning. It feels like having a senior architect on call 24/7. I use it for all sorts of things:
- Technical ideas: "How can we build a good reusable component?" or "What are the pros and cons of different ways to build a component library?"
- Improving code: "Why is this code so slow? Give me three ideas to make it faster."
- Team skills: "Can you help me write this PR comment so it sounds helpful, not like an attack?" or "How can I help my interns understand this topic better?"
Regardless of your job, having an AI to help you think through ideas is incredibly useful. It consistently challenges my assumptions and pushes me to make better, more well-rounded decisions.
6. Why Specs Are More Important Than Ever
Using AI to build applications is powerful, but AI without clear specifications is a recipe for disaster. It’s like asking an intern to build a streaming platform in a few weeks. I often compare AI to an intern: you have to give it the right information at the right time. One single prompt is not going to create a perfect application.
The quality of the output you get from AI is a direct reflection of the quality of the input you provide. So, if managers think AI will replace developers, they first need to get much better at clearly defining their requirements.
For us developers, this is a good thing. By creating clear specifications that the team agrees on, we set the rules. We are still in control. We guide the AI to build what we planned, making sure the final product works well, is consistent, and meets the team's goals. Our jobs aren't disappearing; they're evolving.
At the end of the day, "vibe coding" is a relic of the past because modern software demands planning, teamwork, and scalability. AI, when guided by our own knowledge, is the best way to be more productive while staying in control.
How are you using AI in your role as a developer or tech lead? I'd love to hear your favorite tips in the comments below! 👇
Top comments (0)