---
title: "Comment by Mayank Bansal on \"The Conversation Before the Build: Aligning With AI on Architecture\" | Hashnode"
description: "I had the same realization recently. The moment an agent gets access to real infrastructure, it stops being a prompting problem and starts being an infrastructu"
url: "https://hashnode.com/posts/the-conversation-before-the-build-aligning-with-ai-on-architecture/6a58293f43fef24b32448271/comment/6a5858a2187e6e5b5bb4f73b"
---

[Discussion](https://hashnode.com/posts/the-conversation-before-the-build-aligning-with-ai-on-architecture/6a58293f43fef24b32448271)

Comment [The Conversation Before the Build: Aligning With AI on Architecture](https://hashnode.com/posts/the-conversation-before-the-build-aligning-with-ai-on-architecture/6a58293f43fef24b32448271)

[Mayank Bansal](https://hashnode.com/@monk1)

[Jul 16](https://hashnode.com/posts/the-conversation-before-the-build-aligning-with-ai-on-architecture/6a58293f43fef24b32448271/comment/6a5858a2187e6e5b5bb4f73b)

I had the same realization recently. The moment an agent gets access to real infrastructure, it stops being a prompting problem and starts being an infrastructure problem. We've been moving more and more of those decisions into runtime policies in Failproof AI instead of hoping the model remembers the rules.
