My name is Vitalii Borovyk, and I’ve been in IT since 2012. For the past nine years, I’ve specialized in referral campaigns and custom integrations with third-party platforms—ranging from CRM systems to marketing automation solutions and analytics tools—while also managing bespoke requests for clients around the globe. Because my work spans both development and team leadership, I constantly look for ways to streamline routine tasks, maintain high-quality code, and deliver results on time.
1. Overview of the AI Tools I Use
In my work, I apply several AI-based solutions—both via a web interface and through APIs:
- Well suited for writing code and generating various code fragments.
- Has certain limits that can suddenly “pop up” in the middle of work, disrupting the workflow (which can be frustrating).
- Optimal for small tasks such as quick prototypes and short code blocks.
-
4о Model
- I use it for simple questions, generating short tests, drafting emails, translating texts, or adjusting tone in messages.
- Handles routine tasks well and often saves time.
-
ChatGPT o1/o3 Model
- I use it for more complex tasks: finding documentation, reading large manuals, verifying information, and searching for analytics.
- Helps when I need in-depth research or to check the correctness of extensive code fragments.
- A new model that I’m currently studying and testing.
- I use it to generate code, but for now, I’m only experimenting and do not rely on it for critical tasks.
How I Work With APIs
In addition to web interfaces, I use APIs from several AI providers to automate the process of task analysis. I have three “workers” that handle:
- Checking the completeness of task details (requirements, use cases, etc.).
- Assessing task priority.
- Assisting with estimation (workload assessment).
Right now, these “workers” use the GPT-4o model, but I plan to try connecting Deepseek due to its lower cost to see if it suits my needs. I also have a few ideas on how to use the API to automate other routine tasks related to report generation, so that the AI can generate template statistics, create summaries, and so on.
2. Help in Writing and Debugging Code
Autocompletion and Code Generation
Modern AI-based tools (for example, GitHub Copilot or the Cursor editor) can analyze context, provide suggestions, and even generate entire functions. For front-end developers, this is especially useful when working with:
- JavaScript / TypeScript: quick creation of typical constructs (debounce, throttle, form validation, asynchronous handling).
- CSS / SCSS: style autocompletion, advice on responsive design.
- Ruby code: Even though I’m a front-end developer, I sometimes need to “look under the hood” of the project. Thanks to AI, I can quickly analyze unfamiliar code and understand how certain Ruby logic works.
- Liquid language: In our campaigns, we use Liquid, and all logical blocks and variables are written in it.
Since we launch referral campaigns, we often need quick layout fixes, on-the-fly code optimization, or cleaning up old experiments (A/B tests). Autocompletion and code analysis tools help us quickly make the necessary changes without breaking the existing logic and while maintaining a consistent style.
Cursor, which I use for editing code, not only offers autocompletion but also helps refactor existing code fragments by providing clear suggestions based on an analysis of the entire project. This lets me focus on business logic rather than spending time on typical boilerplate.
Fast Debugging
AI editors and services like Claude or ChatGPT can detect potential errors—ranging from syntactic typos and unclosed brackets to logical inconsistencies. They point out where a bug might be or if a condition is incorrect, and also highlight unused variables.
Instead of spending a long time searching for problems, I get quick hints, saving me time and effort.
3. Improving Code Quality and Code Review
Automatic Analysis of Style and Quality
AI-powered static code analysis services can:
- Identify repeating fragments (violations of the DRY principle or over-DRY-ing),
- Warn about possible vulnerabilities,
- Find redundant constructs,
- Improve code readability through refactoring.
Simplifying the Review Process
As a team manager, I need to conduct code reviews efficiently. AI-based tools:
- Reduce initial review time by finding errors and suggesting improvements,
- Help set priorities—showing where the most critical issues are in the code,
- Can predict which changes might break application logic.
The final decision still rests with the human reviewer, but the risk of overlooking important details drops significantly.
4. Automating Routine Processes
Generating Documentation
One of the most disliked tasks among developers is keeping documentation up to date.
AI tools help by:
- Automatically generating documentation for Confluence,
- Creating descriptions of custom client flows or integraitons,
- Creating presentations’ structure.
5. Analyzing Campaigns for ADA Compliance
When working with referral campaigns, it’s crucial to consider accessibility requirements for people with disabilities (and complying with the Americans with Disabilities Act—ADA). This means the interface and content must meet standards that allow users with various forms of disability to comfortably interact with a web application.
AI tools can be especially helpful here:
- Automated Auditing: Machine learning–based services can scan pages and provide a report on non-compliance (for example, incorrect contrast, missing alt text for images, or improper semantic tags).
- Style and Structure Correction: AI may suggest how to improve the code to meet accessibility guidelines (WCAG 2.1, etc.). For instance, it could prompt you to add aria-label or role, or to change color schemes for better readability.
- Screen Reader Interaction Testing: AI-based test scripts can simulate how a user relying on a screen reader navigates the site, revealing navigation or labeling problems.
As a result, you can quickly refine the interface, fix key errors (e.g., incorrect markup or non-obvious menu items), and ensure the product is not only technically sound but also accessible to a wide range of users.
6. Forecasting and Strategic Planning
As a team lead, I try to look ahead:
- Trend Analysis: AI helps track new libraries, approaches, and tools that are useful for front-end development.
- Vulnerability Forecasting: Systems can identify common mistakes and patterns, indicating where risks may arise.
- Resource Optimization: AI tools help assess team workload, estimate task volumes, and automatically balance assignments based on criticality.
7. Conclusion and Recommendations
Many worry that AI will soon “take” developers’ jobs. Here’s my perspective: in the IT field, constant change is part of the profession, and those who adopt new technology first gain a competitive advantage. Therefore, any AI-related skills—be they prompt engineering or integrating various AI services—have stopped being exotic and become essential tools in every engineer’s arsenal. Improve these skills, and you’ll be able to deliver even more value to your projects, increasing quality while reducing development time. AI will only “take” the jobs of those that are slow to adopt the competitive edge that AI offers.
How to Start Using AI Technologies If Your Tech Stack Doesn’t Involve It
- Create a Test Project
Come up with a small, clearly defined task—for example, writing a Shopify plugin, a Google Chrome extension, middleware for an integration, or anything you’ve always wanted to build just for yourself.
- Formulate the Task as a Product Manager
- Write down in a notebook how you want your future product to work.
- Ask AI services (o1 or o3) to help outline the technical specification. You can phrase your request as if the AI is a “product manager at a successful company.”
- Provide as many details and clarifications as possible—the more detailed your prompt, the more precise the advice.
- If you don’t like the answer, don’t try to correct the AI directly in prompts; instead, ask it to regenerate or edit your initial prompt.
- Implement the Project Using an AI Assistant
- Use Claude, for instance, to bring your idea to life in code step by step.
- Don’t tackle a task too large, and make sure you understand every aspect you find unclear. Read, understand, and refine any code proposed by an LLM.
- If something seems doubtful or you don’t understand the purpose of a function, ask clarifying questions to your friendly AI models.
- Analyze and Repeat
- Understand how the AI generates answers and what code patterns it uses.
- Experiment with different languages and frameworks—this will broaden your horizons and improve your skill in working with various AI models.
- Reinforce your new knowledge by delving deeper into any technology you use or want to learn.
This approach will help you learn not only how to define tasks and create technical specifications for AI but also how to think more “productively” during development. As a result, you’ll quickly figure out the real value of a particular AI tool and be able to organically incorporate it into your workflow.
Most importantly, remember that AI is designed to assist—not replace—developers. Stay curious, keep learning, and always remain ready for change: then any technological innovation, including AI, will become your reliable ally rather than a threat to your career.