--- title: Roblox’s approach to tripling AI code review acceptance... description: Roblox turns historical code review feedback into 'exemplar' rules that guide AI agents, pushing AI-generated review comment acceptance to nearly 70%, compared... canonical: https://daily.dev/posts/roblox-s-approach-to-tripling-ai-code-review-acceptance-rates-ew09cjypd twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Roblox’s approach to tripling AI code review acceptance rates | daily.dev og:description: Roblox turns historical code review feedback into 'exemplar' rules that guide AI agents, pushing AI-generated review comment acceptance to nearly 70%, compared... og:url: https://daily.dev/posts/roblox-s-approach-to-tripling-ai-code-review-acceptance-rates-ew09cjypd og:image: https://api.daily.dev/og/posts/ew09CJYpd.png og:image:alt: Roblox’s approach to tripling AI code review acceptance rates og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Roblox’s approach to tripling AI code review acceptance rates **[Engineering Enablement](https://daily.dev/sources/engineering_enablement)** · 5 min read · 2 upvotes · 0 comments ## Summary Roblox turns historical code review feedback into 'exemplar' rules that guide AI agents, pushing AI-generated review comment acceptance to nearly 70%, compared with 50 to 55% for human comments. About 5% of pull requests now auto-merge after passing AI review, automated safety checks, and risk assessments. Caching and shared context cut the cost of individual rule evaluations by 90%. VP of Engineering Andrew Swerdlow says Roblox tracks agent sessions and token usage, avoids locking into one AI vendor, and expects engineers to shift toward specs, intent, and judgment rather than writing code directly. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How does Roblox get AI-generated code review comments accepted more often than human ones? Roblox mines historical code review feedback to build organization-specific rules called 'exemplars' that guide AI-generated reviews. This lifted AI review comment acceptance to nearly 70%, compared with 50 to 55% for human-written comments, because the AI applies institutional engineering knowledge that generic models lack. _daily.dev surfaces engineering practices like this for teams building their own AI review pipelines._ ### What percentage of pull requests get merged automatically without human review at companies using AI code review? At Roblox, about 5% of pull requests are automatically merged after passing AI code review, programmatic safety checks, and risk assessments. Human engineers still approve the underlying rules and guardrails rather than reviewing every individual change, shifting their role toward defining policy instead of line-by-line review. _Track how far teams are pushing autonomous merging on daily.dev before trusting it in your own pipeline._ ### How can engineering teams reduce the cost of running AI code review at scale? Caching and shared context can cut the cost of individual code review rule evaluations significantly. Roblox reduced this cost by 90% using session tracing, token usage tracking, and shared context across evaluations, treating observability into agent sessions as essential for finding inefficient workflows. _Compare cost-cutting tactics for AI review pipelines with other teams on daily.dev._ ## Similar posts on daily.dev - [Agentic Code Review](https://daily.dev/posts/agentic-code-review-0ue8fcsa1) · Addy Osmani · 147 upvotes · 14 comments - [AI Code Review at Scale: LinkedIn's Multi-Agent Approach](https://daily.dev/posts/ai-code-review-at-scale-linkedin-s-multi-agent-approach-7lpkuhkzi) · InfoQ · 0 upvotes · 0 comments - [Why AI Code Reviews Are Slowing Down PRs](https://daily.dev/posts/why-ai-code-reviews-are-slowing-down-prs-yubjqxjqr) · Aviator Blog · 0 upvotes · 1 comments --- Tags: [#devops](https://daily.dev/tags/devops), [#ai-agents](https://daily.dev/tags/ai-agents), [#ai-coding](https://daily.dev/tags/ai-coding) [View this post on daily.dev](https://daily.dev/posts/roblox-s-approach-to-tripling-ai-code-review-acceptance-rates-ew09cjypd) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Roblox’s approach to tripling AI code review acceptance rates","url":"https://daily.dev/posts/roblox-s-approach-to-tripling-ai-code-review-acceptance-rates-ew09cjypd","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/roblox-s-approach-to-tripling-ai-code-review-acceptance-rates-ew09cjypd"},"datePublished":"2026-10-30T14:38:32.656Z","dateModified":"2026-10-30T14:38:59.454Z","description":"Roblox turns historical code review feedback into 'exemplar' rules that guide AI agents, pushing AI-generated review comment acceptance to nearly 70%, compared...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/aa1affe19326af5ffb7fda7da855b002?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/aa1affe19326af5ffb7fda7da855b002?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Engineering Enablement","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Engineering Enablement","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/3c8216b036d34c1cbe1678385355f35d","url":"https://daily.dev/sources/engineering_enablement"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/roblox-s-approach-to-tripling-ai-code-review-acceptance-rates-ew09cjypd","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"devops,ai-agents,ai-coding","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Engineering Enablement","item":"https://daily.dev/sources/engineering_enablement"},{"@type":"ListItem","position":3,"name":"Roblox’s approach to tripling AI code review acceptance rates"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/roblox-s-approach-to-tripling-ai-code-review-acceptance-rates-ew09cjypd#faq","mainEntity":[{"@type":"Question","name":"How does Roblox get AI-generated code review comments accepted more often than human ones?","acceptedAnswer":{"@type":"Answer","text":"Roblox mines historical code review feedback to build organization-specific rules called 'exemplars' that guide AI-generated reviews. This lifted AI review comment acceptance to nearly 70%, compared with 50 to 55% for human-written comments, because the AI applies institutional engineering knowledge that generic models lack. daily.dev surfaces engineering practices like this for teams building their own AI review pipelines."}},{"@type":"Question","name":"What percentage of pull requests get merged automatically without human review at companies using AI code review?","acceptedAnswer":{"@type":"Answer","text":"At Roblox, about 5% of pull requests are automatically merged after passing AI code review, programmatic safety checks, and risk assessments. Human engineers still approve the underlying rules and guardrails rather than reviewing every individual change, shifting their role toward defining policy instead of line-by-line review. Track how far teams are pushing autonomous merging on daily.dev before trusting it in your own pipeline."}},{"@type":"Question","name":"How can engineering teams reduce the cost of running AI code review at scale?","acceptedAnswer":{"@type":"Answer","text":"Caching and shared context can cut the cost of individual code review rule evaluations significantly. Roblox reduced this cost by 90% using session tracing, token usage tracking, and shared context across evaluations, treating observability into agent sessions as essential for finding inefficient workflows. Compare cost-cutting tactics for AI review pipelines with other teams on daily.dev."}}]} ```