--- title: Your AI Is “In Production.” That Doesn’t Mean It’s... description: Most teams ship LLM features without proper production safeguards, leading to failures like prompt injection, secret leaks, hallucinated policies, and runaway... canonical: https://daily.dev/posts/your-ai-is-in-production-that-doesn-t-mean-it-s-production-ready--focyn39b8 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Your AI Is “In Production.” That Doesn’t Mean It’s Production-Ready. | daily.dev og:description: Most teams ship LLM features without proper production safeguards, leading to failures like prompt injection, secret leaks, hallucinated policies, and runaway... og:url: https://daily.dev/posts/your-ai-is-in-production-that-doesn-t-mean-it-s-production-ready--focyn39b8 og:image: https://api.daily.dev/og/posts/FOCYn39B8.png og:image:alt: Your AI Is “In Production.” That Doesn’t Mean It’s Production-Ready. 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. # Your AI Is “In Production.” That Doesn’t Mean It’s Production-Ready. **[Medium](https://daily.dev/sources/medium_js)** · 3 min read · 0 upvotes · 0 comments ## Summary Most teams ship LLM features without proper production safeguards, leading to failures like prompt injection, secret leaks, hallucinated policies, and runaway API costs. The AI Production Readiness Framework (APRF), published by StackRail, is a vendor-neutral, gated methodology designed to answer one core question: can this AI application safely operate in production? Unlike maturity score models that average away critical failures, APRF uses mandatory pass/fail gates across 8 domains and 27 pillars — a single gate failure blocks deployment. It covers Core (40 gates) and Regulated (61 gates) profiles, with lenses for RAG, agents, voice, and coding agents, plus crosswalks to NIST AI RMF, ISO 42001, OWASP LLM Top 10, and SOC 2. A self-assessment tool is available for teams to evaluate their readiness in 15–30 minutes. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What does the AI Production Readiness Framework (APRF) actually check for an AI agent with tools? It checks for a charter and step budget, allowlisted and schema-validated tools, non-bypassable human approval on high-impact actions, the ability to kill a looping agent, and spend ceilings to prevent runaway costs. Teams must demonstrate these with artifacts like configs, logs, and drills; failing to do so results in a gate fail rather than a partial score. _daily.dev surfaces frameworks like this for teams hardening agentic AI before shipping to production._ ### Why does APRF avoid giving AI systems an overall readiness percentage score? Averaging scores lets a disastrous area, such as secrets hygiene, get masked by a good-looking metric elsewhere, producing a misleadingly high overall percentage. APRF instead treats mandatory checks as pass or fail blockers and takes the minimum capability attainment across pillars rather than a mean, so failures always surface instead of being averaged away. _engineers weighing AI maturity scoring approaches can track frameworks like this on daily.dev._ ### What profiles and gate counts does the APRF v0.10 catalog define for AI systems? Version 0.10 defines a Core Profile with 40 gates for Tier-2 customer-facing AI and a Regulated Profile with 61 gates for Tier-3 or regulated systems, spread across 8 domains and 27 pillars covering security, safety, data, model lifecycle, agents, reliability, cost, and governance, with additional lenses for RAG, agents, voice, and coding agents. _daily.dev helps teams evaluating AI governance checklists stay current on emerging standards like this._ ## Similar posts on daily.dev - [AI Guardrails: Implementing Safety for Production LLM Apps](https://daily.dev/posts/ai-guardrails-implementing-safety-for-production-llm-apps-z1y9eixhb) · BigData Boutique blog · 1 upvotes · 0 comments - [From AI Prototype to Production: The Engineering Gaps Developers Often Miss](https://daily.dev/posts/from-ai-prototype-to-production-the-engineering-gaps-developers-often-miss-msmbjxcci) · SitePoint · 1 upvotes · 0 comments --- Tags: [#security](https://daily.dev/tags/security), [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#ai-safety](https://daily.dev/tags/ai-safety) [View this post on daily.dev](https://daily.dev/posts/your-ai-is-in-production-that-doesn-t-mean-it-s-production-ready--focyn39b8) ```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":"Your AI Is “In Production.” That Doesn’t Mean It’s Production-Ready.","url":"https://daily.dev/posts/your-ai-is-in-production-that-doesn-t-mean-it-s-production-ready--focyn39b8","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/your-ai-is-in-production-that-doesn-t-mean-it-s-production-ready--focyn39b8"},"datePublished":"2026-07-28T07:32:00.312Z","dateModified":"2026-09-14T06:03:02.702Z","description":"Most teams ship LLM features without proper production safeguards, leading to failures like prompt injection, secret leaks, hallucinated policies, and runaway...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/6d8a913bf30b8441c2537b0235f7d69d?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/6d8a913bf30b8441c2537b0235f7d69d?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Medium","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":"Medium","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/medium","url":"https://daily.dev/sources/medium_js"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/your-ai-is-in-production-that-doesn-t-mean-it-s-production-ready--focyn39b8","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"security,llm,ai-agents,ai-safety","timeRequired":"PT3M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Medium","item":"https://daily.dev/sources/medium_js"},{"@type":"ListItem","position":3,"name":"Your AI Is “In Production.” That Doesn’t Mean It’s Production-Ready."}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/your-ai-is-in-production-that-doesn-t-mean-it-s-production-ready--focyn39b8#faq","mainEntity":[{"@type":"Question","name":"What does the AI Production Readiness Framework (APRF) actually check for an AI agent with tools?","acceptedAnswer":{"@type":"Answer","text":"It checks for a charter and step budget, allowlisted and schema-validated tools, non-bypassable human approval on high-impact actions, the ability to kill a looping agent, and spend ceilings to prevent runaway costs. Teams must demonstrate these with artifacts like configs, logs, and drills; failing to do so results in a gate fail rather than a partial score. daily.dev surfaces frameworks like this for teams hardening agentic AI before shipping to production."}},{"@type":"Question","name":"Why does APRF avoid giving AI systems an overall readiness percentage score?","acceptedAnswer":{"@type":"Answer","text":"Averaging scores lets a disastrous area, such as secrets hygiene, get masked by a good-looking metric elsewhere, producing a misleadingly high overall percentage. APRF instead treats mandatory checks as pass or fail blockers and takes the minimum capability attainment across pillars rather than a mean, so failures always surface instead of being averaged away. engineers weighing AI maturity scoring approaches can track frameworks like this on daily.dev."}},{"@type":"Question","name":"What profiles and gate counts does the APRF v0.10 catalog define for AI systems?","acceptedAnswer":{"@type":"Answer","text":"Version 0.10 defines a Core Profile with 40 gates for Tier-2 customer-facing AI and a Regulated Profile with 61 gates for Tier-3 or regulated systems, spread across 8 domains and 27 pillars covering security, safety, data, model lifecycle, agents, reliability, cost, and governance, with additional lenses for RAG, agents, voice, and coding agents. daily.dev helps teams evaluating AI governance checklists stay current on emerging standards like this."}}]} ```