Edit and sharpen your writing with AI

The trick isn't deleting whole ideas — it's deleting the words that surround your ideas. Most drafts are 20% padding you won't miss.

1. Paste the essay and say 'cut this to X words, keep every distinct point, only remove filler and repetition'.
2. Ask it to mark what it cut so you can veto anything you actually wanted.
3. Read the result out loud — if a sentence lost meaning, put it back.
4. Ask 'which paragraph is still the flabbiest?' and cut that one further.
5. Do a final pass yourself; the AI over-trims sometimes and flattens your voice.

Ironically, AI is good at spotting AI-sounding writing. The tells are predictable: too-balanced sentences, empty transition phrases, and words no human says out loud.

1. Paste your text and ask it to flag every phrase that sounds generic or template-like.
2. Ask it to rewrite those flagged bits in a plainer, more conversational voice.
3. Tell it to vary the sentence length — short punchy ones next to longer ones.
4. Ask it to remove 'in conclusion', 'furthermore', 'it's important to note' and similar padding.
5. Read it aloud; anything you'd never say to a friend, change it yourself.

Half the time you can't even see the problem because you know what you meant. The AI reads it cold, the way your recipient will.

1. Paste the email and ask 'how will this land emotionally to the person receiving it?'
2. Ask it to point out the specific lines that read as blunt or accusatory.
3. Ask for a softer version that keeps the same actual message — you still want the point made.
4. Check the softened version didn't go so far it sounds fake or spineless.
5. Send the one that sounds like you on a good day.

If you ask it to 'improve this', it hands you a whole new draft in its voice and you learn nothing. Ask for critique first, edits second.

1. Say 'critique this, don't rewrite it' up front and mean it.
2. Ask three specific questions: is my point clear, is my opening strong, where do I lose the reader.
3. Have it quote your actual sentences when it points out problems.
4. Fix the problems yourself first, then hand it back for a second read.
5. Only ask for a rewrite of one stubborn sentence you genuinely can't crack.

Your first sentence is doing 90% of the work of deciding whether anyone reads sentence two. Most people waste it clearing their throat.

1. Paste your opening and ask 'does this make me want to keep reading, honestly?'
2. Ask for five alternative first lines in different styles — question, bold claim, scene, stat, tension.
3. Tell it to skip any opening that starts with 'In today's world' or a definition.
4. Pick the one that fits your voice, not just the flashiest.
5. Rewrite it once more in your own words so it's actually yours.

Clarity is 'can they understand it'. Style is 'do they enjoy reading it'. You fix clarity first — a beautiful sentence nobody understands is worthless.

• Clarity edits: cut jargon, shorten long sentences, fix confusing pronouns, one idea per line.
• Style edits: rhythm, word choice, voice, where you break paragraphs for effect.
• Do clarity passes before style passes, or you'll polish sentences you should delete.
• AI is genuinely good at clarity; it's weaker at style because style is personal.
• If you're short on time, just do the clarity pass — it fixes 80% of bad writing.
• Never let it 'style' away your actual voice in the name of polish.

Wordiness comes from writing the way you think — in loops and hedges and 'what I mean to say is'. When you talk you fill silence; on the page that filler just sits there.

The habit to build: after every sentence, ask 'did that word earn its place?' The AI can point out your personal filler words, and once you see your patterns, you start catching them before you type them.

It means the sentence you're proudest of — the clever line, the fancy metaphor — is often the one hurting your writing, because you kept it for you, not the reader.

AI is useful here precisely because it isn't attached to your ego. Ask it 'which line am I keeping because I like it, not because it works?' and it'll call out the darling you couldn't bear to cut yourself.

Don't ask it to rewrite the whole thing — you'll end up with something that sounds like every other AI-buffed statement and admissions readers can smell it. Instead, paste it and ask 'where does this go generic, where does the real me disappear?'

It'll show you the paragraphs that could belong to anyone. Those are the ones you rewrite by hand tonight with a specific detail only you have. Use the AI to find the flat spots, not to fill them.

Paste both and ask the AI to read them like a hiring manager who has 30 seconds. Ask specifically: does the cover letter say anything the resume doesn't, or is it just restating bullet points in paragraph form? That's the most common killer.

It'll usually tell you the cover letter is dead weight — most are. Ask it to help you make the letter do one job the resume can't: show why you specifically want this one role, in your actual voice.

Paste it and ask 'what is the one thing this post is actually about, and which paragraphs don't serve it?' Rambling almost always means you tried to make two or three points in one piece.

The AI will spot the tangents you're too close to see. Cut the paragraphs that belong in a different post entirely, and suddenly the real piece is right there, sharper and half the length.

For most everyday writing, AI covers what you'd pay an editor for anyway. You bring in a human when the stakes and the nuance are high enough to justify the cost.

Grammarly catches mechanics. A chat AI catches meaning. They fix different problems, and honestly you can use both — Grammarly for the typos, the AI for whether your point even lands.

You are — as long as the ideas and the choices stay yours. An editor doesn't stop being editing when it's a machine; every published author works with editors and nobody says they didn't write the book.

The line is this: if you're using it to find weak spots and decide which fixes fit your voice, you're still the writer. If you're pasting a prompt and shipping whatever comes out unread, that's not your writing and it'll show. Keep your hand on the wheel.

Trust the diagnosis, question the rewrite. It's genuinely reliable at spotting what's wrong — wordy sentences, weak openings, buried points. It's less reliable at fixing them, because it tends to smooth everything into the same safe, flat voice.

So take its critiques seriously and its rewrites with suspicion. Read every suggested change out loud before you accept it — if it sounds less like you, reject it and fix the problem your own way. The AI finds the leak; you patch it.