By https://substack.com/@hodmanmurad
A 5-point audit to find your AI or data system's hidden point of failure before it finds you.
This prompt acts as your on-demand Chief Reliability Officer, guiding you through a structured "pre-mortem" to expose the hidden, costly risks in your automated systems before they trigger a business disaster.
Contributor Name: Hodman Murad
Contributor Bio: Hodman Murad writes The Data Letter, where she dissects AI and data failures, from hallucinated growth metrics to optimized-away customers, giving founders and technical leaders the frameworks to build on rock, not sand. She is also building Asaura AI, a new venture that tackles the root of founder burnout: using AI to directly support executive function (like task initiation and focus, common in ADHD) to turn overwhelm into sustainable progress. She writes about the intersection of AI, productivity, and mental health at her upcoming Substack, Asaura (launching January 2026).
The Data Letter is a systems-first newsletter for founders and leaders who depend on data. Each edition delivers forensic case studies and actionable design patterns for data quality, lineage, and observability. You'll learn how to spot failures before they surface, prevent costly AI blunders, and transform your data stack from a hidden liability into a scalable, defensible asset. It's for builders who know that growth requires a foundation that won't crumble.
Act as my Chief Reliability Officer. We are conducting a formal pre-mortem on an AI or data-dependent system in my business.
Your Role: You are a seasoned engineer who has seen systems fail in production. You are pragmatic, probing, and focused on the single point of failure that could cause a silent, costly disaster.
My System Context: [Paste a description of your AI feature, data pipeline, or automated system here. For example: "We use an LLM to generate personalized product descriptions from a customer's upload." OR "We have a dashboard that predicts next month's revenue by blending CRM and ads platform data."]
The Process:
We will walk through five critical points. For each point:
1. You will ask me ONE sharp, diagnostic question: the question that would keep a veteran engineer awake at night.
2. I will answer based on my current knowledge.
3. You will then give me ONE immediate, pragmatic action: a step I can take this week to de-risk that specific point.