# Vapi AI Reduces PR Merge Time by 50 Percent with In-House Tool > Vapi's in-house Pedantic Review Bot has halved PR merge times, enhancing both speed and code quality. This innovative tool redefines software review processes, allowing for more efficient engineering workflows. **Source**: vapi.ai | **Published**: 2026-10-12 | **Type**: case_study ## Key Facts - Vapi's PR merge time dropped from 130 to 65 hours, showcasing significant efficiency gains. - Pedantic Review Bot identified 363 blockers in 1,694 PRs, revealing critical vulnerabilities in code. - In-house tools outperformed paid solutions, indicating a competitive edge in custom automation. - Cost-effective review strategies led to better defect detection, enhancing overall financial performance. - Continuous prompt testing indicates a strategic shift towards data-driven decision-making in development. ## Summary \## Summary Vapi, a software engineering company, faced lengthy pull request review times averaging over 130 hours. To address this, Lat Ware developed the Pedantic Review Bot, which halved the average time to 65 hours for pull requests to be merged. This in-house solution improved efficiency and accuracy in code reviews. \## Background Vapi operates in the software engineering industry and employs a team of engineers responsible for managing code changes through pull requests. Prior to the deployment of the Pedantic Review Bot, the company relied on existing AI review tools that failed to adequately address issues within their codebase, resulting in prolonged review times. \## Challenge The primary challenge was the excessive time taken for pull requests to be reviewed and merged, which averaged over 130 hours. This delay hindered productivity and slowed down the development process. \## Solution Lat Ware built the Pedantic Review Bot to automate the review process for pull requests. The bot reviews every pull request opened by Vapi engineers, including drafts, and allows authors to request additional passes. It skips previously reviewed commits to avoid duplicate reviews. The bot provides inline comments with a summary and severity markers, ranging from green for minor issues to red for critical blockers. Additionally, a second bot, the Pedantic Architecture Review Bot, was created to identify design problems, ensuring comprehensive coverage of both code and architecture issues. \## Results The implementation of the Pedantic Review Bot reduced the average time from open to merge for pull requests from over 130 hours to 65 hours. Between September 2 and October 2, 2026, the bot reviewed 1,694 pull requests and identified 363 blocker findings across 276 of them. These findings included critical issues related to builds, tests, continuous integration, and deployment. \## Key Insights - Automating code reviews can significantly reduce review times, enhancing overall productivity. - It is essential to tailor AI tools to specific codebases for optimal performance. - False positives in automated reviews can increase the workload for engineers; thus, careful prompt design is crucial. - Running multiple lower-cost review configurations can yield better results than a single high-effort run. \## Customer Testimonial “I make decisions from data, so I enjoyed being able to run experiments that showed me how well these tools perform for us. Experiments are where the truth comes out.” — Lat Ware, Member of the Technical Staff ## Entities - **Companies**: Vapi - **Products**: Pedantic Review Bot, Pedantic Architecture Review Bot - **People**: Lat Ware ## Key Concepts Automated code review, Pull request management, AI review tools, Design flaw detection, Review efficiency, Prompt engineering, False positives in reviews, Latency budget in development ## Definitions - **Pull Request**: A pull request is a method of submitting contributions to a software project, where changes are proposed and reviewed before being merged. - **Automated Review**: Automated review refers to the use of software tools to analyze code changes for potential issues without human intervention. - **Severity Marker**: A severity marker is a label indicating the seriousness of an issue found during a code review, ranging from minor to critical. - **Latency Budget**: Latency budget is the acceptable time frame within which a task, such as code review, should be completed to maintain workflow efficiency. - **False Positive**: A false positive occurs when a review tool incorrectly identifies a non-issue as a problem that needs addressing. ## Use Cases - Reducing pull request merge time - Identifying design flaws in code - Automating code review processes - Improving code quality through inline comments - Analyzing code changes for existing issues - Experimenting with different review prompts ## Frequently Asked Questions **How did the Pedantic Review Bot improve merge times?** The Pedantic Review Bot reduced the average time from open to merge from over 130 hours to 65 hours by automating the review process and providing timely feedback. **What types of issues does the Pedantic Architecture Review Bot identify?** The Pedantic Architecture Review Bot focuses on design problems that may not be evident from individual lines of code, helping to catch high-level design pattern issues. **Why is prompt engineering important for the review bots?** Prompt engineering is crucial because the way prompts are structured can significantly affect the quality and relevance of the findings reported by the bots. **What are the benefits of using multiple review runs?** Using multiple review runs allows for a more thorough analysis, as different configurations can catch various issues that a single run might miss, leading to better overall code quality. **What should teams consider when building their own review bots?** Teams should consider their review queue's latency budget, the importance of distinguishing between pre-existing and new issues, and the potential costs associated with false positives when designing their review bots. ## Links - [Read on Welcome.AI](https://welcome.ai/content/vapi-ai-reduces-pr-merge-time-by-50-percent-with-in-house-tool) - [Original source](https://vapi.ai/blog/ai-code-review-bot) - [Vapi](https://welcome.ai/company/vapi): Featured company --- Source: Welcome.AI | https://welcome.ai/content/vapi-ai-reduces-pr-merge-time-by-50-percent-with-in-house-tool