Articles
Field notes from building production AI
Agentic AI Architecture: What Breaks in Production
Architecture is a failure-containment problem, not a component list. Which design decision stops which production failure, and how to tell if yours is ready.
Human-in-the-Loop AI: Where the Approval Gate Belongs
Human-in-the-loop and full autonomy are settings on the same system. Where approval belongs, how to tell if a gate is working, and when to remove it.
AI Agent Observability: A Practical Best-Practices Guide
You cannot fix what you cannot see. A vendor-neutral guide to tracing, span-level evals, and what to instrument in production.
AI Agents for Sales: What Actually Ships to Production
What separates a production-grade AI sales agent from a demo: clear ownership, a human approval gate, and hard stop conditions.
The AI Adoption Framework: Why Most Enterprise AI Never Reaches Production
A five-layer framework for moving enterprise AI from successful demos into resilient, production-grade systems.