--- title: How We Gave AI Agents the Keys to Production and Still... description: monday.com built a shared platform that lets LLM agents read and propose changes to production infrastructure like Terraform and Helm. Agents run in a loop... canonical: https://daily.dev/posts/how-we-gave-ai-agents-the-keys-to-production-and-still-sleep-at-night-msgbrdmo3 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: How We Gave AI Agents the Keys to Production and Still Sleep at Night | daily.dev og:description: monday.com built a shared platform that lets LLM agents read and propose changes to production infrastructure like Terraform and Helm. Agents run in a loop... og:url: https://daily.dev/posts/how-we-gave-ai-agents-the-keys-to-production-and-still-sleep-at-night-msgbrdmo3 og:image: https://api.daily.dev/og/posts/MsGbRdmo3.png og:image:alt: How We Gave AI Agents the Keys to Production and Still Sleep at Night 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. # How We Gave AI Agents the Keys to Production and Still Sleep at Night **[monday Engineering](https://daily.dev/sources/monday)** · 10 min read · 0 upvotes · 0 comments ## Summary monday.com built a shared platform that lets LLM agents read and propose changes to production infrastructure like Terraform and Helm. Agents run in a loop calling wrapped tools, with every call checked against a policy dispatcher. An independent risk classifier, a separate Bedrock call with no access to the agent's own reasoning, scores each change from the raw diff and plan output. After one quarter, about 50% of an average 1,688 monthly Infra tickets were resolved without human escalation, across 7 agents serving 6 Infra teams. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How can I let an LLM agent open or approve production infrastructure PRs safely? Combine per-run scoped credentials, wrapped tools with a policy dispatcher, and an independent risk classifier that scores changes from raw diffs without seeing the agent's own reasoning. monday.com's platform also wraps external content in untrusted tags to resist prompt injection, enforces branch protection at the Git host level, and keeps a full audit trail per run in LangSmith so every action can be reconstructed. _Engineers wiring agents into infra pipelines track guardrail patterns like these on daily.dev._ ### How do you prevent an AI agent from trusting its own risk assessment of a code change it wrote? Use a separate model call with fresh context that has no access to the agent's conclusions, scoring the run purely from the diff, the plan output, and the files touched. monday.com calls this an independent risk classifier; a critical verdict blocks approval in the dispatcher before it reaches GitHub, and humans review the decision instead. _Teams weighing how much autonomy to give coding agents can follow approaches like this on daily.dev._ ### What results has monday.com seen from using AI agents to handle infrastructure tickets? About 50% of an average 1,688 Infra tickets per month were resolved by agents without human escalation after one quarter. The platform grew to 7 agents serving 6 Infra teams, after manual scripting and self-service automation had only covered about 30% of the ticket load. _Anyone sizing up AI agent ROI for ops work can follow real deployment numbers on daily.dev._ ## Similar posts on daily.dev - [Agents in Production, Part 2: Build an AI-Native Coding Agent That Turns Tickets into Reviewed Pull Requests with Managed Agents](https://daily.dev/posts/agents-in-production-part-2-build-an-ai-native-coding-agent-that-turns-tickets-into-reviewed-pull--j8febodig) · DigitalOcean Community · 2 upvotes · 0 comments - [Why better tickets help agents write better code](https://daily.dev/posts/why-better-tickets-help-agents-write-better-code-suwct5vww) · Atlassian · 1 upvotes · 0 comments - [AI Agents in DevOps: Hype vs. Reality in Production Pipelines](https://daily.dev/posts/ai-agents-in-devops-hype-vs-reality-in-production-pipelines-zp70wtm0z) · DevOps.com · 1 upvotes · 0 comments - [AI Agents at work: real-time platform insights in Slack](https://daily.dev/posts/ai-agents-at-work-real-time-platform-insights-in-slack-as53pzdrz) · monday Engineering · 17 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#terraform](https://daily.dev/tags/terraform), [#iac](https://daily.dev/tags/iac), [#prompt-injection](https://daily.dev/tags/prompt-injection), [#amazon-bedrock](https://daily.dev/tags/amazon-bedrock) [View this post on daily.dev](https://daily.dev/posts/how-we-gave-ai-agents-the-keys-to-production-and-still-sleep-at-night-msgbrdmo3) ```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":"How We Gave AI Agents the Keys to Production and Still Sleep at Night","url":"https://daily.dev/posts/how-we-gave-ai-agents-the-keys-to-production-and-still-sleep-at-night-msgbrdmo3","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/how-we-gave-ai-agents-the-keys-to-production-and-still-sleep-at-night-msgbrdmo3"},"datePublished":"2026-10-26T09:19:40.780Z","dateModified":"2026-10-26T09:30:50.177Z","description":"monday.com built a shared platform that lets LLM agents read and propose changes to production infrastructure like Terraform and Helm. Agents run in a loop...","image":"https://media.daily.dev/image/upload/s--CxzD6vbw--/f_auto/v1722860399/public/Placeholder%2005","thumbnailUrl":"https://media.daily.dev/image/upload/s--CxzD6vbw--/f_auto/v1722860399/public/Placeholder%2005","isAccessibleForFree":true,"articleSection":"monday Engineering","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":"monday Engineering","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/3bce2c4601a8468097dd077a4a9e1417","url":"https://daily.dev/sources/monday"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/how-we-gave-ai-agents-the-keys-to-production-and-still-sleep-at-night-msgbrdmo3","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,terraform,iac,prompt-injection,amazon-bedrock","timeRequired":"PT10M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"monday Engineering","item":"https://daily.dev/sources/monday"},{"@type":"ListItem","position":3,"name":"How We Gave AI Agents the Keys to Production and Still Sleep at Night"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/how-we-gave-ai-agents-the-keys-to-production-and-still-sleep-at-night-msgbrdmo3#faq","mainEntity":[{"@type":"Question","name":"How can I let an LLM agent open or approve production infrastructure PRs safely?","acceptedAnswer":{"@type":"Answer","text":"Combine per-run scoped credentials, wrapped tools with a policy dispatcher, and an independent risk classifier that scores changes from raw diffs without seeing the agent's own reasoning. monday.com's platform also wraps external content in untrusted tags to resist prompt injection, enforces branch protection at the Git host level, and keeps a full audit trail per run in LangSmith so every action can be reconstructed. Engineers wiring agents into infra pipelines track guardrail patterns like these on daily.dev."}},{"@type":"Question","name":"How do you prevent an AI agent from trusting its own risk assessment of a code change it wrote?","acceptedAnswer":{"@type":"Answer","text":"Use a separate model call with fresh context that has no access to the agent's conclusions, scoring the run purely from the diff, the plan output, and the files touched. monday.com calls this an independent risk classifier; a critical verdict blocks approval in the dispatcher before it reaches GitHub, and humans review the decision instead. Teams weighing how much autonomy to give coding agents can follow approaches like this on daily.dev."}},{"@type":"Question","name":"What results has monday.com seen from using AI agents to handle infrastructure tickets?","acceptedAnswer":{"@type":"Answer","text":"About 50% of an average 1,688 Infra tickets per month were resolved by agents without human escalation after one quarter. The platform grew to 7 agents serving 6 Infra teams, after manual scripting and self-service automation had only covered about 30% of the ticket load. Anyone sizing up AI agent ROI for ops work can follow real deployment numbers on daily.dev."}}]} ```