General Guide6 min readUpdated Fri Aug 21 2026 12:05:37 GMT+0000 (Coordinated Universal Time)

How to Write a Landing Hero in One Sentence

A website skill for a landing hero in one sentence. Name one user and one outcome, add one constraint, ban slogan fog, generate three lines in ChatGPT, Claude, or Gemini, then pick one and read it at billboard speed.

Anyone staring at a blank hero can write a landing hero in one sentence in 15 minutes: one user, one outcome, one constraint. Run it in ChatGPT, Claude, or Gemini before you stack three value props and a gradient and call it a homepage.

A hero is not a manifesto. A manifesto lists virtues. A hero helps one person decide to stay. If the line could sit on a competitor's page with the logo swapped, it is still fog.

AI Overviews often quote the first visible sentence as the site's claim. Write a line a machine can cite and a human can believe. The same line usually becomes the H1.

This loop is a skill, not a brand-voice workshop. If you are still pasting "write hero copy" into a blank chat, read what a skill prompt actually is and save the card at the end.

Write a landing hero in one sentence, five steps

User first. Outcome second. Constraint third. Variants last.

1. Name one user and one job

One sentence you write yourself. "A founder who needs a waitlist page by Friday." "A support lead who needs nav labels people can find." Not "teams of all sizes."

Instruct ChatGPT, Claude, or Gemini: "The user is this role. The job is this verb plus object. Do not draft the hero yet. List what would make the user bounce in the first five seconds."

Bounce reasons are constraints. You will need them in step 3.

If the brief is a pile, shape it like a messy note into a one-page brief before you touch the fold.

2. Ban slogan fog

Paste this list:

reimagine, empower, unlock, next-generation, all-in-one, seamlessly, the future of, made simple used as a period, we believe, finally.

Then add the three words your last campaign overused.

Tell the model: "If a draft uses a banned word, rewrite the whole sentence. Do not swap in a synonym. Change the thought."

Synonym swaps are how "empower" becomes "enable" and you still sound like a keynote.

The same ban-first habit that keeps captions from sounding like ads keeps a hero from sounding like a booth.

3. Force who, outcome, constraint

Every hero, same bones, one sentence:

  • Who — the role, or "you" if the page is already aimed.
  • Outcome — the change after one use, not the feature list.
  • Constraint — time, limit, or honesty. "In one sentence." "Without a demo call." "On the page you already have."

Example: one sentence so a stranger knows the job before the fold.

Instruct: "One sentence, 12 to 20 words. Outcome cannot be a feature list. Constraint must come from the bounce list or the input. No semicolon. No second claim after an em dash."

If a hero image is in play, write its prompt in a separate loop. Do not let image adjectives leak into the sentence.

4. Generate three, pick one, write the subhead as proof

Ask for three hero sentences that share the same user and outcome, but change only the constraint.

Pick one. Do not blend. Then ask for one subhead only: "One line of proof. A number, a time cap, or a limit. 8 to 16 words. Not a second hero."

Refuse a "punchier pair" if the subhead restates the hero. Cut any mission clause. A second slogan is not a subhead.

5. One human pass at billboard speed

Read the sentence once, at the speed of a driver passing a sign. If you need to reread, it failed.

Kill:

  • A we that has no user.
  • An outcome that is actually a category ("better collaboration").
  • A constraint that is a vibe ("effortlessly").

Fix those in your own words. Send the edit back: "Replace only the hero sentence. Do not re-expand the subhead."

One pass. Then stop. Save the shape with the Promptcrates tools for reusable website skills so the next landing page does not open with a manifesto.

Reusable skill card

Trigger: A landing page is about to ship with a slogan, a feature stack, or three buttons in the first screen.

Input: One user role, one job sentence, one bounce-reason constraint, plus your banned slogan list.

Output: Three hero sentences of 12–20 words in the who / outcome / constraint shape, plus one proof subhead of 8–16 words, zero banned words.

Stop: After you pick one pair and read the hero at billboard speed. No blending. No second button of equal weight.

Takeaways

  • One user, one outcome, one constraint. One sentence.
  • Ban slogan fog before you generate. Synonyms still sound like keynotes.
  • The subhead is proof, not a second hero.
  • Answer engines quote the first sentence. Save the loop. The last slogan is disposable.

Frequently Asked Questions

Can the hero mention the product name?

Yes, once, if the name is how people search. The outcome still comes first. A name with no outcome is a logo in sentence form.

What belongs in the subhead?

Proof. One fact, one limit, or one next step the first screen can keep. The subhead is not a second hero.

Should I use ChatGPT, Claude, or Gemini for hero copy?

Any of the three works. Claude is strict with banned slogan words. ChatGPT is fast at variants. Gemini is fine for a first pass. All three need your human pass.

How long is one sentence allowed to be?

Twelve to twenty words. If you need a semicolon, you have two jobs. Split or cut.

Can I add a second call to action under the hero?

One primary action. A text link for the skeptical path is fine. Two buttons with equal weight is a menu, not a hero.

Reviewed by PromptCrates EditorialTags: landing, hero, website, copy, chatgpt, claude, gemini

Related Masterclasses & Guides

View all guides