How to Write Better Prompts for Claude and ChatGPT
Most disappointing AI responses come from vague prompts, not weak models. Here’s a repeatable method for writing prompts that get useful answers on the first try.
📌 Related reading: Digital Nomad Lab · Home Tech Lab
In This Guide
Steps
Step 1: State the Goal, Not Just the Task
Instead of “write an email,” say “write a short email asking my landlord to fix a leaking faucet, polite but firm.” The goal and tone context change the output dramatically. A model given only the task guesses at tone and audience, and it usually guesses generic — the more specific detail you give it up front, the less back-and-forth editing you’ll need afterward.
Step 2: Give the Model Relevant Context Up Front
Paste in the actual document, code, or background details you’re working with rather than describing them from memory — models perform much better with source material directly in the prompt. When I tried summarizing a long report from memory versus pasting the actual text, the pasted-text version consistently caught details I’d have otherwise left out or misremembered.
Step 3: Specify the Format You Want Back
If you want a table, bullet list, or specific word count, say so explicitly. Models default to prose unless told otherwise, which often isn’t what you actually need. A simple line like “return this as a markdown table with three columns” removes almost all ambiguity about output shape.
Step 4: Ask for One Task at a Time on Complex Requests
For multi-step work, break it into a sequence of prompts rather than one giant request — this makes it easier to catch and correct errors early instead of discovering them at the end. A single sprawling prompt that asks for research, an outline, and a final draft all at once tends to produce a weaker version of each step than three focused prompts run in sequence.
Step 5: Iterate Instead of Starting Over
If the first response is close but not quite right, tell the model specifically what to change rather than rewriting your original prompt from scratch — this preserves useful context from the first attempt. Something as short as “keep the structure, but make the tone more casual” is usually enough to get a workable second draft.
Step 6: Build a Small Library of Prompts That Already Work
Once you land on a prompt structure that reliably works for a recurring task — a weekly summary, a code review request, a travel-planning brief — save it as a template rather than reconstructing it from memory each time. This matters for any recurring workflow, whether you’re drafting emails at a desk job or researching visa and cost-of-living details as a digital nomad comparing cities on a budget.
Tips
- ✅ Save prompt templates that work well for recurring tasks (emails, summaries, code reviews).
- ✅ Ask the model to explain its reasoning briefly if you need to verify a factual claim.
- ✅ Use explicit constraints (“under 150 words,” “no bullet points”) when format matters.
- ✅ For home automation prompts (naming routines, writing scene descriptions for a smart speaker), the same specificity rule applies — a vague request like “make my evening routine” gets a far weaker result than naming the exact devices and times involved, which matters if you’re also trying to reduce your smart home’s standby power usage through smarter scheduling.
Warnings
⚠️ Always verify factual claims, statistics, or citations an AI model provides — models can generate plausible-sounding but incorrect information. ⚠️ Be cautious pasting sensitive personal, financial, or client data into a prompt unless you’re confident about the platform’s data handling policy.
Q&A
Does prompt length matter?
Not directly — a longer prompt with relevant context usually outperforms a short vague one, but padding with irrelevant detail doesn’t help.
Is there a difference in prompting Claude vs. ChatGPT?
The core principles (clear goal, context, format) apply to both, though each model has slightly different strengths worth testing for your specific use case.
Should I use the same prompt style for coding tasks and writing tasks?
The core structure carries over, but coding prompts generally benefit even more from pasting exact error messages or code rather than paraphrasing them.
Sources: Anthropic — Claude Prompt Engineering Overview, OpenAI — Prompt Engineering Guide. Fact-checked based on public sources as of July 20, 2026.
