The risk-tiered execution model for AI agents

A framework for deciding which agent actions can run autonomously and which need human approval. Based on reversibility, blast radius, and confidence thresholds.

Framework

Why we don't recommend LangChain for production agent systems

A controversial but defensible take on framework selection. When abstraction layers help, when they hurt, and what to use instead.

Opinion

Anatomy of a failed AI pilot: what went wrong at a Series C fintech

A teardown of a real AI implementation that stalled. The technical decisions were fine. The organizational ones killed it.

Teardown

How to evaluate an LLM for your use case (without spending $50k)

A practical guide to eval design. What to measure, how to build test sets, and when "good enough" is actually good enough.

Guide

The build/buy/partner decision matrix for AI capabilities

When to build in-house, when to buy a platform, and when to partner with a specialist. A decision framework based on 40+ engagements.

Framework

What "AI-ready" actually means for a mid-market engineering team

Five dimensions of readiness, scored 1–5. Most teams overestimate their data maturity and underestimate their governance gap.

Assessment

Get the AI Ops Brief.

Weekly frameworks and teardowns for IT leaders navigating AI implementation. No hype. No vendor pitches.

Subscribe