Buy vs build: Why WRITER bought their AI code review tool
30% time savings
Leadership productivity
Unified engineering culture
Cross-team visibility
Overview
For Kyrylo Buha, Member of Technical Staff at WRITER, managing code quality across an AI-native company requires both technical excellence and operational efficiency. WRITER is where the leading enterprises orchestrate AI-powered work. With WRITER’s end-to-end platform, teams can build, activate, and supervise AI agents that are grounded in their company’s data and fueled by WRITER’s enterprise-grade LLMs.
Kyrylo’s role spans both the software development cycle and people management, making efficient code review processes critical for team productivity. Integrating AI into the code reviews just made sense for their team. At first, they considered using their own LLMs for code reviews, Kyrylo explained. However, they quickly realized that building an in-house solution would be too time-consuming and resource-intensive.
That led Kyrylo to CodeRabbit. What started as an experiment with his frontend teams quickly expanded across the company, with CodeRabbit now being used on more than 37 repositories. Multiple backend teams also adopted the tool after seeing the frontend team’s success.
Challenge: Scaling AI-native development with efficient code reviews
Before CodeRabbit, WRITER’s development process followed a standard approach with one reviewer per pull request, escalating to multiple reviewers only for core and essential changes. However, several challenges emerged as the company scaled its AI-powered development.
- Allocating resources across the core product and internal tooling
While WRITER had the technical capability to build their own code review AI, resource allocation became a critical consideration. “We realized that it would be time-consuming and we didn’t have the capacity to build something in-house,” Kyrylo noted. The team needed to focus its AI expertise on its core product rather than internal tooling.
- Maintaining consistency across multiple teams and projects
With developers working across frontend, backend, and full-stack projects, maintaining coding standards and style consistency became increasingly challenging. WRITER had developed internal style guides distributed across Confluence and Notion pages, but ensuring adherence required significant manual oversight.
- Optimizing review processes for high-volume development
Processing approximately 50 pull requests per day across their active repositories, WRITER needed a solution that could maintain consistency while accelerating their development velocity. “We are kind of AI evangelists. So we believe that AI can significantly improve performance,” Kyrylo explained, emphasizing their commitment to leveraging AI for operational efficiency.
Why WRITER loves CodeRabbit
Catching a regex UUID validation bug before production
CodeRabbit’s contextual understanding quickly earned the trust of WRITER’s engineering team. It consistently flags issues that humans tend to overlook, Kyrylo shared, such as tricky if/else branches, negative conditions, and overly permissive validations. One standout example was a UUID check. CodeRabbit identified a hidden problem in a developer's utility function, even though unit tests covered it. The developer's regex was too broad, allowing non-UUID strings to slip through and risking corrupted data in production.CodeRabbit spotted the gap and suggested a stricter pattern with proper anchors, version, and variant constraints, preventing a subtle bug from shipping.
In this case, it enforced a more precise UUID check and saved us from a production issue.
- Kyrylo Buha, Member of Technical Staff
By tightening validation at review time, CodeRabbit helped WRITER reduce error risk, raise code standards and move faster with confidence.
Unifying style guides with path-based instructions
WRITER unified their scattered internal documentation from Confluence/Notion guidance into a single CodeRabbit style guide, using File Patterns to target guideline files, CodeRabbit’s Learnings (local/global/auto) to control scope, and path-based instructions to tailor rules by directory (e.g., services/payments/**).
With our style guides in CodeRabbit, newcomers get instant feedback that matches how we code, and over 70% of those suggestions are accepted.
- Kyrylo Buha, Member of Technical Staff
Cross-team collaboration through JIRA integration
"CodeRabbit’s JIRA integration delivered unexpected benefits by surfacing connections between work streams across different teams." Kyrylo explained.
What I find really useful is that it can track the PRs across multiple projects in Jira. So, when you make the change, CodeRabbit tracks the changes in other projects on other boards.
- Kyrylo Buha, Member of Technical Staff
This feature helped prevent duplicate work and improved coordination between the teams.
Visualization with sequence diagrams for more technical reviews
CodeRabbit’s sequence diagrams feature became an instant hit with WRITER’s developers, particularly for understanding complex changes. “The sequence diagram gives us more clarity around the changes that were performed within the PR,” Kyrylo noted. This feature proved especially valuable for large pull requests, providing developers with comprehensive insights without requiring local testing. “CodeRabbit’s sequence diagrams let reviewers grasp a PR’s behavior without checking out the branch, allowing clearer, faster reviews when local runs aren’t practical.”
Productivity boost inside VS Code and Cursor
WRITER uses the CodeRabbit IDE extension for seamless integration into their day-to-day workflows. The extension allows engineers to trigger reviews, resolve comments inline, and access context-aware Learnings right where they code — eliminating context switching between tools. “What we find useful is the extension for Cursor. It’s amazing to be able to run the PR checks and fix them prior to raising the PR for a wider audience,” Kyrylo shared.
Results: Measurable improvements across the development lifecycle
30% time savings for team leads: Kyrylo reports cutting code-review time by 30% thanks to CodeRabbit’s targeted comments, which let him skip already covered areas and focus on higher value work: “Personally, my code review time is down around 30%. When I see useful comments, I don’t focus on those areas; I cover other areas and rely on CodeRabbit to have my back.” The savings are especially valuable in his dual role managing two teams: “The savings are huge for my dev/manager role, I can spend more time on management, coding, and strategic product improvements," Kyrylo shared.
Rapid organic adoption across the organization: CodeRabbit’s ease of integration facilitated rapid expansion beyond Kyrylo’s initial frontend teams. “It’s really easy to integrate. That’s why you don’t miss anything. You just enable the repository and you instantly start getting the feedback,” he noted.
CodeRabbit = Effortless AI integration for AI-native development
Before CodeRabbit
- •One human reviewer per PR -> Limited coverage
- •Style guide scattered across Confluence & Notion
- •Slow review cycles for large PRs
- •Hard to maintain consistency across 37+ repos
After CodeRabbit
- 30% faster code reviews for team leads
- Regex and conditional bugs caught early
- Unified style guide inside CodeRabbit
- Sequence diagrams & Jira integration improve context and collaboration
For Kyrylo and his team at WRITER, CodeRabbit represents the ideal balance of powerful AI capabilities with minimal overhead. “It’s a fast way to introduce AI code reviews with a quick setup and roll out across projects and teams, getting value from day one.”
What sets CodeRabbit apart is its ability to complement its existing development culture while adding immediate value. As Kyrylo emphasized, “What I love about CodeRabbit is its easy-to-use nature, a critical factor for a company that needs to maintain focus on their core AI product development.”
The platform’s success at WRITER demonstrates how even companies building AI solutions themselves can benefit from purpose-built tools that integrate seamlessly into their workflows. By handling routine code quality checks and providing enhanced visibility into complex changes, CodeRabbit allows WRITER’s developers to focus on what they do best: building innovative AI solutions.
San Francisco, United States
https://www.writer.com/Team size
500+ employees
Languages
JavaScript, Python
Challenge
WRITER Scaling AI-native development efficiently while maintaining consistent code quality and review velocity across multiple teams and repositories.
