A practical guide series on writing prompts that actually work - the techniques, structures, and evaluation habits that turn a vague ask into a reliable result, whether you're using Claude, GPT, Gemini, or building an agent on top of any of them.
Same model, same underlying question - the difference between a vague prompt and an engineered one is almost always role, context, and format.
New to prompt engineering, or not sure it's a "real" skill? This guide makes the case and sets up the vocabulary the rest of the series uses.
A small set of moves - shown by example, not just named - that cover the majority of real prompting problems.
The three techniques that solve most prompting problems on their own - what each one is for, and when adding examples or reasoning steps actually helps.
→The anatomy of a well-built prompt - system role, background context, and an explicit output format - and why structure beats cleverness.
APA Mastery runs live, practical sessions on working with modern AI tools - not just theory.
See What's On →Patterns for harder problems, and how to tell if your prompt is actually working instead of just guessing.
What to reach for once single-shot prompting stops being enough - reasoning-and-acting loops, voting across multiple attempts, and breaking one prompt into several.
→How to reliably get JSON, not prose, out of a model - JSON mode, schema-constrained output, and what to do when a field comes back malformed anyway.
→What actually changes when you move a working prompt from one model family to another, and what to keep in mind if you're supporting more than one.
→What prompt injection actually is, why it's hard to fully prevent, and the practical mitigations that meaningfully reduce the risk in real applications.
→How to tell if a prompt change actually helped - test sets, rubrics, and the discipline of changing one thing at a time.
→Treating prompts like the production code they are - version control, system vs. user prompt separation, and change management as your prompt library grows.
Every technique in this series - zero-shot, few-shot, chain-of-thought, and the prompt structure checklist - as one printable PDF. Drop your email and it's yours, plus a heads-up whenever new guides go live.
No spam, unsubscribe any time.