PROMPT GUIDE
The difference between a mediocre and a great AI answer is almost always the prompt. The same request, restructured — role assigned, context given, one clear ask, format specified — routinely doubles answer quality across ChatGPT, Claude, Gemini, Grok, Meta AI, and Perplexity. Here is the structure that works everywhere, plus a free generator that writes it for you.
Generate your prompt — free, no sign-up
"Act as an experienced financial planner…" focuses the model's knowledge and tone in one line. It is the highest-value sentence in almost any prompt.
Who you are, what you have tried, what "good" looks like. Models cannot over-include context you actually provided — only context you forgot to.
Chat assistants answer whatever comes last. Bury a second question mid-paragraph and you will get one answer — to the wrong thing.
"Give me 5 options as a table with pros and cons, then your top pick." Steps, table, short answer, draft — naming the shape gets you that shape.
Follow-ups like "go deeper on option 2" carry all prior context. That is the chat advantage — use it instead of writing ever-longer first prompts.
Act as [expert role]. Context: [who you are, the situation, what you have tried]. I need: [one clear deliverable]. Constraints: [tone, length, things to avoid]. Format: [steps / table / short answer / draft]. If anything important is unclear, ask me before answering.
Role, context, one clear ask, and a named output format. The free generator on this site assembles all four from a plain description of what you want — tuned for ChatGPT and the other major chatbots.
Yes — role-plus-context-plus-one-ask is universal across ChatGPT, Claude, Gemini, Grok, Meta AI, and Perplexity. The differences are in tone, not structure.
Complete, not long. Every sentence should be role, context, the ask, constraints, or format. If a sentence does none of those, the generator cuts it — and the answers get sharper.
coding agent · lovable · v0