In recent years, AI has become an integral part of software engineering, not only as a tool for code generation but also as a partner in code review and quality improvement.
In this expert opinion, Roman Rudyi shares his team’s hands-on experience with AI-powered code reviews, what works, what doesn’t, and how to make the most of these tools. He also discusses the strengths and limitations of AI reviewers and practical use cases of GitHub Copilot.
What are the main advantages of AI code review?
The main advantage of artificial intelligence in code review lies in its speed and efficiency. A machine can analyse vast amounts of code almost instantly, detecting potential weaknesses, bugs, logical errors, or formatting issues. This significantly reduces the workload for a human reviewer, who only needs to verify and confirm the detected problems.
The second advantage is consistency. AI does not have “bad moods” and applies the same rules to every review, ensuring stable and predictable results.
The third strength is its ability to learn. AI provides suggestions based on large volumes of analysed code and shares best programming practices. This is particularly useful for junior developers who are just becoming familiar with new technologies or languages.
The fourth one is integration with the development environment. For instance, GitHub Copilot integrates seamlessly with GitHub, allowing users to apply its recommendations in just one click without leaving the platform.
The fifth is improved team collaboration. AI can generate summaries for pull requests, concisely outlining changes and organising them by file. This helps reviewers grasp the context more quickly and saves valuable time.
What are the main weaknesses of AI code review?
- Limited contextual understanding: AI does not comprehend business logic, team conventions, or specific architectural decisions, so it cannot always assess the correctness of an implementation.
- False positives: AI may flag correct code as erroneous or overlook genuine issues.
- The risk of overreliance on AI: Some less experienced developers. They may trust AI’s conclusions without proper verification, which can lead to production errors.
- Security concerns: Since AI often sends code to external servers for analysis, it may pose a cybersecurity risk. When using AI in development projects, it's essential for the entire team to strictly follow all security standards and best practices, including data anonymisation. By doing so, you can fully protect your clients’ data and maintain a high level of trust.
- Limited customisation: Each team has its own naming conventions and directory structures, which AI does not always consider. This can be partially mitigated by creating configuration files for each repository or organisation.
How does your team use GitHub Copilot in practice?
The most useful features of GitHub Copilot include summarising pull requests, detecting potential issues and explaining code in an accessible manner.
Copilot can act as a reviewer, highlighting possible logical mistakes and offering advice on patterns and best practices. However, AI does not replace human review — especially when business logic is involved.
Typically, the team first reviews what Copilot has detected automatically, then switches to chat mode to ask clarifying questions. This is especially convenient for large or complex pull requests.
The team uses the official IntelliJ plugin, which allows code reviews even before committing changes to version control. This helps identify issues locally before the code reaches the repository. AI can even generate commit messages — concise, grammatically correct, and relevant.
GitHub Copilot Chat offers three operating modes:
- Ask — for explaining existing code or exploring libraries and frameworks.
- Edit — for refactoring and generating tests in selected files.
- Agent (S) — the most advanced mode, where AI can create new files, run tests, and autonomously fix code.
How can AI help developers learn new technologies?
AI greatly simplifies getting started with a new programming language or framework — it can generate a basic project structure. However, the resulting code is average in quality: suitable for an initial prototype, but not at an expert level.
For example, a “connection flights” search method: Copilot first simplified and optimised it, then rewrote it in a functional style. This process helps developers gain insight into better coding practices.
To achieve the best results, it is crucial to formulate requests clearly, describe all possible scenarios, and provide examples of input and output data.
What role should AI play in software development?
AI is a powerful tool, but not a replacement for humans. Code should be readable by both machines and people — but it should be written primarily for humans, with clarity in mind.
AI helps enhance code readability and accelerate the development process, but its effectiveness depends on how developers ask questions, provide context, and verify the results. Ultimately, the final word must always remain with the engineer.
Related insights
The breadth of knowledge and understanding that ELEKS has within its walls allows us to leverage that expertise to make superior deliverables for our customers. When you work with ELEKS, you are working with the top 1% of the aptitude and engineering excellence of the whole country.
Right from the start, we really liked ELEKS’ commitment and engagement. They came to us with their best people to try to understand our context, our business idea, and developed the first prototype with us. They were very professional and very customer oriented. I think, without ELEKS it probably would not have been possible to have such a successful product in such a short period of time.
ELEKS has been involved in the development of a number of our consumer-facing websites and mobile applications that allow our customers to easily track their shipments, get the information they need as well as stay in touch with us. We’ve appreciated the level of ELEKS’ expertise, responsiveness and attention to details.
