The average cold email gets a 1-3% reply rate. The best sales teams consistently hit 15-25%. The difference isn't volume — it's the structure of the email and the specificity of the pain.

Here's what works and how to use AI to write it.

Why Most Cold Emails Get Ignored

The four most common mistakes:

1. Opening with "I" — "I'm reaching out because..." signals self-interest immediately. The reader doesn't care about you.

2. Feature-leading — "We have a platform that does X, Y, and Z." Nobody cares about features. They care about outcomes.

3. No research signal — Generic emails that could be sent to anyone get treated like spam.

4. Vague CTA — "Let me know if you're interested" requires a decision. "Are you free for 15 minutes on Thursday?" requires only a yes or no.

The Structure That Gets Replies

Line 1 — Specific observation (1 sentence)

Something specific about them: a recent hire, a product launch, a post they wrote, a problem common to their role. This proves you're not mass-mailing.

Example: "Saw your team just launched a new outbound motion — congrats on the Series B."

Lines 2-3 — The pain they have (1-2 sentences)

State the problem that's common to their role or situation. Don't mention your product yet.

Example: "Most sales leaders I talk to at that stage spend 2+ weeks building their first outreach sequence. By the time it's live, the ICP has changed."

Line 4-5 — The offer (1-2 sentences)

What you do and what it means for them. One concrete outcome.

Example: "We help teams generate multi-touch sequences with A/B variants in under 5 minutes. I can show you what it looks like for your motion."

Line 6 — CTA (1 sentence, specific)

Make it easy to say yes. A specific time or a simple question.

Example: "Do you have 15 minutes on Thursday or Friday?"

Using Elehua AI's Email Composer

In the Email Composer, provide:

• Who you're writing to (role, company context)

• The pain you're addressing

• What you offer

• Desired tone (direct, professional, friendly)

• CTA type (book a call, reply with question, click a link)

For outreach sequences, use the Outreach Sequence Builder — it generates 6-8 touches across email and LinkedIn with timing, A/B variants, and follow-up angles that don't feel like copy-paste.

Personalizing at Scale

AI-generated cold emails need personalization to land. The formula: generate the template with AI, then manually add one specific line per contact (the "specific observation" line). This takes 30 seconds per contact and dramatically increases reply rates.

For 50 contacts: AI writes the template in 2 minutes. You spend 25 minutes adding the observation line to each. Total: 27 minutes for 50 personalized cold emails.

What Not to Do

• Don't use AI to generate fully personalized emails at scale from a contact list without review — quality drops fast and generic-sounding "personalization" backfires

• Don't send 8-touch sequences to cold contacts you've never had a conversation with — this is spam at volume

• Don't ignore reply signals — if someone opens 4 times but doesn't reply, call them instead of sending email #5

The Bottom Line

Cold email is a conversation starter, not a sales pitch. The goal of the first email is to get a reply — not to explain your entire product. AI handles the structure and language. You provide the specific observation that makes each email feel like it was written for that person.

How to Write Cold Emails That Actually Get Replies (With AI)