OpenAI has published a prompting fundamentals guide through its Academy platform, aimed at everyday ChatGPT users who want better results without wading through academic literature on prompt engineering. It's basic by design — and that's the point.
What's new
The guide distills prompt engineering into three steps: outline the task clearly using an action verb, add relevant context (files, background, constraints), and specify your desired output format, tone, and length. OpenAI also includes a tiered "Okay / Better / Best" example set using a machine learning explanation prompt — showing concretely how layering in constraints and structure meaningfully changes response quality. Supplementary tips cover breaking complex requests into smaller steps and asking explicitly for multiple options when you want choices.
Why it matters
Most prompting guides either over-engineer the concept into a pseudo-discipline or dumb it down to uselessness. This one threads the needle reasonably well. The tiered example format is the strongest part — watching the same base prompt get sharpened with word limits, structural requirements, and analogy constraints is a cleaner demonstration than any amount of abstract advice. It's pitched at the "first week with ChatGPT" user, but even experienced users will find the output-specification section a useful checklist.
What to watch
OpenAI Academy appears to be building out a structured learning layer around ChatGPT use — this guide links to related resources on working with files and apps. Whether this becomes a full curriculum or stays a loose collection of tips is worth tracking, particularly as competitors like Google and Anthropic have their own documentation ecosystems. For now, this is a solid reference to send someone who keeps complaining that ChatGPT "doesn't understand" them.