Practical AI tips you can try that same day. Written for people who want to make AI work, not just talk about it.
twice a month on Substack
The structure underneath that decides whether AI works for you, or just makes you sound like everyone else. Context engineering, harness engineering, and the build order that compounds.
Read the full essay →Twenty-four Claude Code terms in plain English. Agent, hook, MCP, skill, shell. Written for anyone who uses Claude Code without writing code themselves.
Read the glossary →A five-phase framework to build an AI assistant that knows your expertise, follows your judgment, and communicates in your voice.
Read the full guide →Five steps to move your team from scattered experiments to shared workflows that stick.
Read the full guide →The tool question, the half-briefing, and the first-draft trap. Three patterns I see in every workshop.
Read the full article →From a scoring sheet to a chatbot used thousands of times by 80+ editors. Built in months, not hours, with the domain expert in the lead.
Read the case study →Social Media Maik creates posts for LinkedIn, Instagram and internal channels. 25 minutes of work, done in 3. The real investment is the knowledge base behind it.
Read the case study →Tone Tara checks and rewrites content in our brand voice. From scoring 20,000 pages to a tool used by 80+ editors daily. Built in months, not minutes.
Read the case study →A complete walkthrough from zero to a working Custom GPT.
Read on Substack →The skill that matters is not which tool you pick. It's how you brief it.
Read on Substack →Three layers, how skills and agents work together, and why it compounds.
Read on Substack →