HomePrompt Library › ChatGPT Prompt for Cold Emails

ChatGPT Prompt for Cold Emails

A proven cold-email prompt that gets ChatGPT to write short, personalized outreach with strong subject lines and a single clear ask.

Email · Works great in ChatGPT

Copy-ready prompt

Act as a B2B sales copywriter. Write a cold email to [recipient role] at [company] with the goal of [book a call / get a reply].
My offer: [product/service] which helps [audience] achieve [specific outcome].
Personalize the first line using this detail about them: [recent news / role / pain point].
Requirements:
- 3 subject line options under 45 characters.
- Under 120 words, easy to skim on mobile.
- Lead with their problem, not my company.
- One clear call to action (no "let me know if interested").
- Confident, human tone — no jargon or hype.

Want a version tailored to you?

Answer a few quick questions and the Email Prompt Generator builds a custom prompt from your exact details.

📧 Open the Email Prompt Generator

How to write a cold email that actually gets replies

The average professional gets dozens of cold emails a week and deletes most in under three seconds. Winning that three-second decision is almost entirely about relevance and brevity, not clever wording. A large language model like ChatGPT can produce a technically fluent email in seconds, but left to its own defaults it will write something long, self-centered, and generic — the exact profile of a message that gets ignored. The prompt above works because it removes those defaults and replaces them with the constraints that experienced sales writers apply by instinct.

The single most important instruction in the prompt is "lead with their problem, not my company." Most cold emails open with a paragraph about the sender: who they are, what their company does, how long they have been in business. The reader does not care yet. Opening instead with a problem the reader recognizes — a missed quota, a slow onboarding flow, rising ad costs — earns the extra ten seconds you need to make your case. When you fill in the [recent news / role / pain point] field, you are giving ChatGPT the raw material to make that opening feel researched rather than mass-produced.

The 120-word cap matters just as much. Long emails signal that replying will be work, and busy people avoid work. A tight email signals that you respect their time and that the ask is small. If ChatGPT returns something longer, reply with "cut this to 90 words without losing the hook" and it will tighten the copy considerably. The single call to action is the third pillar: giving someone one obvious next step ("Are you open to a 15-minute call Thursday?") converts far better than a vague "let me know if this is interesting," which pushes the mental work of deciding what to do back onto the reader.

Finally, generating three subject lines is not padding. Open rate is the ceiling on every other metric — nobody replies to an email they never opened — so the subject line deserves genuine testing. Keeping each under 45 characters ensures the whole line is visible on a phone, where most email is now first seen.

Why this prompt works

Cold emails fail when they lead with the sender instead of the reader. This prompt forces a problem-first opening, caps the length so it stays skimmable, and demands a single CTA — the three things that most improve reply rates. Asking for multiple subject lines lets you A/B test the part that decides whether the email is opened at all.

How to customize it

  • Swap [recipient role] and [company] for a real prospect, and paste a genuine detail so the first line feels researched.
  • Change the goal to match your funnel — a reply, a demo, or a meeting.
  • If replies are low, ask the AI for 3 shorter variants under 70 words.

Example output

Sample only

Subject line options:

  • Cutting onboarding from 3 weeks to 3 days
  • A faster path for new Acme hires
  • Idea for Acme's onboarding backlog

Email:

Hi Dana,

Saw Acme grew headcount 40% last quarter — congrats. Teams scaling that fast usually hit the same wall: onboarding drags to three-plus weeks and new reps ramp slowly.

We help ops teams like yours cut that to a few days with guided, self-serve onboarding flows. One customer got new hires productive 60% faster in their first month.

Are you open to a 15-minute call Thursday to see if it fits how Acme onboards?

Thanks,
Sam

Prompt variations to try

Warm follow-up after no reply

Act as a B2B sales copywriter. Write a 60-word follow-up to a cold email that got no reply. Reference the original problem [problem] briefly, add one new proof point [stat/result], and end with an easier yes than a call — a single question they can answer in one line. Friendly, not pushy.

Founder-to-founder tone

Write a cold email from me, a founder, to another founder at [company]. Skip corporate language entirely. Open with a specific observation about their product [detail], keep it under 90 words, and make the ask a casual 'worth a quick chat?' Sound like a real person, not a sales team.

Reply-focused (no meeting ask)

Write a cold email whose only goal is to get a one-word reply. Lead with [recipient]'s likely pain point, offer a specific resource or insight for free, and close by asking a yes/no question. Under 100 words, no calendar links, no CTA beyond the question.

Common mistakes to avoid

  • Opening with "My name is X and I work at Y." The reader does not care about you until you have shown you understand their problem — delete the intro and start with their situation.
  • Leaving the personalization placeholder generic. If [recent news / role / pain point] is vague, the email reads as a template. Paste one specific, verifiable detail.
  • Stacking multiple asks. "Reply, or book a call, or check out our site" splits attention. Pick one action and make it the only one.
  • Accepting the first draft as-is. ChatGPT tends toward safe, generic phrasing. Push back once ("make the opening line sharper and more specific") and quality jumps.
  • Ignoring the subject line. A brilliant email with a boring subject never gets opened — test the three options the prompt gives you.

Frequently asked questions

Will ChatGPT write cold emails that sound robotic?

By default it can, because its safe register is formal and generic. The "confident, human tone — no jargon or hype" instruction pushes against that, and a single follow-up like "rewrite this to sound like a real person, not marketing copy" usually fixes any remaining stiffness.

How do I personalize at scale without writing each email by hand?

Keep the prompt fixed and only change the personalization line for each prospect. Spend your research time finding one genuine detail per recipient — a recent hire, a product launch, a public pain point — and paste it into the placeholder. The structure stays constant; the opening stays specific.

What length actually performs best for cold email?

Shorter almost always wins. Under 120 words is a reliable target, and many high-performing cold emails are closer to 75. The goal is a message that can be read in full on a phone screen without scrolling.

Is cold emailing legal?

Legitimate B2B outreach is generally permitted in most regions if you identify yourself honestly, offer a way to opt out, and do not use deceptive subject lines. Rules vary by country (for example, consent requirements are stricter in the EU under GDPR), so check the regulations that apply to your recipients before sending at volume.

Tip: replace the parts in [square brackets] with your own details before you send. The more specific you are — audience, tone, goal, constraints — the better the AI output.