Stage 2 · Lesson 2 of 9
Specific instructions and output format
Most weak outputs come from vague prompts, not weak models. Write a prompt the way you would brief a capable new colleague who knows nothing about your product. Say what the task is, who it is for and what the limits are.
Say what to do with missing or odd input. And say exactly what shape the answer must take. "Summarise this" leaves everything open.
"Summarise this support thread in at most three bullet points for an engineer, then a line starting Status: with open or resolved" does not. Keep your instructions apart from the data with markers, such as XML-style tags like <email> and </email>.
Ask for a fixed output format your code can read reliably. Test on ten varied inputs, not just the one you wrote it for.
A question to check yourself, an AI practice session and 1 cheat sheet entry come with the full roadmap. Unlock
Resources for this stage
Start with· follow one of these
- CoursePaidGrokking Prompt Engineering for Software Engineers
Design Gurus · Beginner · Subscription
Prompting patterns for engineers building features, not for chatting
- GitHub repoFreeAnthropic Courses
Anthropic · Beginner
Free notebooks: API fundamentals, then the interactive prompt engineering tutorial
Practise
Keep handy
Also worth your time
Few-shot, chain of thought, JSON enforcement and parsing failures in one place