52 AI Tools Tips - ChatGPT, Claude & Prompt Writing

Prompt Writing Tips

Be specific in your prompts—instead of 'write about dogs' try 'write 200 words about the history of golden retrievers as family pets'.
Give AI a role to play like 'act as an experienced copywriter' to get more focused, professional responses.
Break complex requests into steps—ask AI to first outline, then draft, then refine for better results.
Include examples of the style or format you want in your prompt to guide the AI's output.
Ask AI to think step by step for logical problems to improve accuracy on complex reasoning tasks.
Specify your audience—'explain quantum physics to a ten-year-old' produces very different results than 'explain to a PhD student'.
Use constraints to improve output—'answer in exactly three bullet points' forces concise, focused responses.
Tell AI what not to do as well as what to do—'explain without using jargon' helps avoid unwanted content.
Iterate on prompts—if the first response isn't right, refine your request rather than starting over.
Save your best prompts in a document so you can reuse and adapt them for similar tasks.

Use custom instructions to set permanent preferences like tone, length, and formatting for all conversations.
Ask ChatGPT to critique its own response—'what's weak about this answer?' often improves the next attempt.
Start a new chat for unrelated topics to avoid confusion from previous context bleeding into responses.
Use ChatGPT's code interpreter to analyse spreadsheets, create charts, and process data without leaving the chat.
Ask for alternative versions—'give me three different approaches to this email' provides useful variety.
Paste long documents and ask ChatGPT to summarise, extract key points, or identify action items.
Use voice mode for brainstorming sessions when typing feels too slow or formal.
Ask ChatGPT to explain its reasoning to catch errors and understand how it reached conclusions.

Use AI to overcome writer's block by asking for multiple opening paragraph options to spark ideas.
Have AI rewrite your draft in different tones—formal, casual, persuasive—to find what works best.
Ask AI to identify weak points in your argument or gaps in your logic before publishing.
Use AI to condense long documents into executive summaries or social media-friendly snippets.
Request specific word counts to match platform requirements—'rewrite this in exactly 280 characters'.
Have AI check for clichés and suggest more original phrasings in your writing.
Use AI to localise content—ask it to adapt American English to British English and vice versa.
Get AI to create outlines before you write to ensure logical structure and complete coverage.

Paste error messages directly into AI and ask for debugging help—it often spots issues instantly.
Ask AI to explain unfamiliar code line by line when learning a new codebase or language.
Request code in small, tested chunks rather than asking for entire applications at once.
Ask AI to add comments to your code or improve variable names for better readability.
Use AI to convert code between languages—Python to JavaScript, for example—as a starting point.
Ask AI to suggest test cases for your functions to improve code quality and catch edge cases.
Describe what you want in plain English first, then ask AI to translate it into code.
Request optimisation suggestions—'how could this function be made more efficient?'—to improve performance.

Use descriptive adjectives like 'cinematic lighting' or 'watercolour style' to guide visual output.
Specify aspect ratios upfront—'16:9 landscape' or 'square format for Instagram' saves regeneration time.
Reference specific artists or art movements like 'in the style of Art Deco' for consistent aesthetics.
Add negative prompts to exclude unwanted elements—'no text, no watermarks, no people'.
Generate variations of promising images rather than starting over when you're close to what you want.
Use consistent character descriptions across prompts to maintain recognisable figures in series.

Set up AI to draft routine emails—you edit rather than write from scratch.
Use AI to prepare meeting agendas and talking points from shared documents and notes.
Have AI transcribe voice notes and extract action items to speed up processing.
Create templates with AI for recurring tasks like weekly reports or project updates.
Use AI to research topics quickly by asking for summaries of multiple perspectives.
Automate data entry by describing what you need and having AI format it correctly.

Always fact-check AI responses—models can confidently state incorrect information.
AI doesn't know current events past its training date without web access enabled.
Sensitive or proprietary information entered into AI may be stored or used for training.
AI struggles with precise calculations—use a calculator or spreadsheet for numbers that matter.
Complex multi-step reasoning can lead to errors—break problems down and verify each step.
AI may generate plausible-sounding but non-existent sources—always verify citations independently.