Writing Better Prompts
A handful of habits that get sharper results from the AI — one goal per ask, clear constraints, and giving it context.
Last updated: June 24, 2026
The AI is only as clear as the request you give it. None of these are tricks — they're just the habits that reliably get you a usable result on the first try instead of the third.
One ask, one goal
Give each message a single, clear job. "Rewrite this paragraph to be more concise" lands better than "rewrite this, also check the facts, and make it punchier." When a response comes back noisy or off, it's usually because the ask was doing too much — split it into separate messages.
Say who it's for and how long
The AI can't see what's in your head. Spell out the things that change the answer:
- Audience — beginners, experts, a client, your boss.
- Length — a sentence, a paragraph, 300 words.
- Tone — formal, casual, persuasive, plain.
- Format — bullet points, a table, prose, a headline.
A request with these built in needs far less back-and-forth.
Give it context instead of describing it
You don't have to paste your draft into the chat. Use what Muses already has:
- Select text in the editor before you ask — the selection rides along as context automatically.
- Mention sources with
@to point the AI at your own articles and materials. - Let memory carry your standing preferences so you don't restate them every time. See What the AI Remembers.
Break big tasks into rounds
A long, structured piece rarely comes out right in one shot. Outline first, then draft section by section, then revise. Each round gives you something to react to, and the AI builds on what you've already approved. A work mode can keep it focused on one round at a time.
Refine, don't restart
If the result is close but not right, add a constraint and ask again — "shorter," "less formal," "keep the second half." Starting a brand-new conversation throws away the context you've built. Reach for a clean conversation only when you're switching to an unrelated task.