How AI Has Helped Me at Work

Most conversations about artificial intelligence focus on code generation. However, after using it daily for the past several months, I have come to the conclusion that this has not been its greatest impact on my work. The real transformation has happened in how I approach problems, plan solutions, understand complex systems, and execute tasks that used to consume a significant amount of time.

AI Forces Me to Plan Solutions Better

One of the most unexpected changes has been the increased importance of planning. When a task is poorly defined, the results tend to be poor. In contrast, when the requirements, expected behavior, edge cases, and acceptance criteria are clearly specified, the quality of the responses improves considerably.

For this reason, I have ended up spending more time analyzing problems before starting to implement a solution. I frequently use a Spec-Driven Development (SDD) approach, where the specification becomes the central piece of the process. The clearer and more detailed the specification, the better the results I get.

Paradoxically, AI has forced me to think more before writing less code.

AI Writes a Large Part of My Code

It has also significantly reduced the amount of code I write manually. TypeScript, GraphQL queries, utility functions, endpoints, and parts of backend logic can now be developed much faster than before.

This does not mean the work disappears. It is still necessary to make technical decisions, validate implementations, and ensure that the solution correctly addresses the project requirements. However, many repetitive or mechanical tasks require less time than before.

As a result, I can focus more on solution design and less on writing repetitive code.

AI Has Made Debugging Easier and Less Stressful

Finding the root cause of a bug has always been one of the most frustrating parts of software development. A seemingly simple problem can end up consuming hours of investigation.

AI has considerably reduced that time. It can analyze error messages, identify possible root causes, explain unexpected behavior, and help rule out hypotheses quickly.

It does not always find the correct answer, but it usually accelerates the investigation process significantly. As a consequence, solving problems has become faster and much less stressful.

AI Makes It Easier to Understand Old and Complex Projects

Probably one of the benefits I value most is the ability to understand legacy projects more quickly.

In many cases I find myself working on older technologies or systems with complex business logic that has evolved over years. It used to be common to spend long hours reading files, following dependencies, and trying to mentally reconstruct how a feature worked.

Today I can use AI to analyze large chunks of code, explain relationships between modules, summarize responsibilities, and help me understand much more quickly why a particular implementation exists.

The need to understand the system still exists, but the time required to gain context has been enormously reduced.

AI Has Helped Me Automate Testing and Accessibility

Another area where I have gained great benefits is automation.

Tools like Playwright, combined with AI, make it easier to generate automated tests and reduce the effort required to validate functionality. I have also used AI to perform accessibility reviews, interpret WCAG criteria, and detect issues before they reach QA.

More recently, tools like Figma MCP have made it possible to automate visual validations and pixel-perfect checks that previously required much more extensive manual reviews.

This not only saves time but also improves the consistency of the development process.

Conclusion

If I had to summarize the impact of AI in a single sentence, I would say that my work has shifted from being primarily focused on writing code to being focused on designing solutions.

I still code every day, but now I spend much more time defining problems, creating specifications, making technical decisions, validating implementations, and ensuring the quality of the final result. Many of the tasks that used to consume a large part of my time — research, debugging, understanding complex systems, and testing — can now be resolved much more efficiently.

In my experience, the biggest advantage of AI is not that it writes code. It is that it allows me to spend less time on implementation and more time on architecture, design, and problem solving.