Marketing

How to Write Cold Emails with AI

Craft personalized cold outreach emails with strong subject lines that get opened, read, and replied to.

Cold email success depends on relevance and brevity. AI can generate highly personalized first lines based on prospect research, write concise value propositions that respect the reader's time, and craft subject lines optimized for open rates — giving sales teams scalable outreach without sacrificing personalization.

What makes cold emails get deleted versus replied to

Cold emails that get deleted share a profile: they open with the sender's name and company, spend two paragraphs describing the product, and close with 'I'd love to set up some time to learn more about your needs.' They are about the sender, not the recipient. Cold emails that get replies are different in one fundamental way: they demonstrate that the sender already knows something specific and relevant about the prospect. The most effective personalization is not using someone's first name — it is referencing something real: their recent funding round, a specific pain point mentioned in their LinkedIn post, or a result their company recently published. AI can generate this kind of personalized opening when you feed it prospect research.

The anatomy of a cold email that converts

The highest-converting cold email structure is four elements: a personalized first line that references something specific about the prospect or their company, a one-sentence statement of the problem you solve that is relevant to their situation, a one-sentence credibility proof (a metric, a named client, a specific result), and a single low-friction CTA — not 'let me know if you're interested' but a specific small ask like 'Would a 15-minute call on Thursday work?' The total word count should be under 100 words. Shorter emails have higher reply rates because they respect the reader's time and signal confidence — a salesperson who writes 300 words to explain their value proposition doesn't understand their own value proposition.

How AI scales personalization without sacrificing quality

The bottleneck in cold outreach is writing personalized first lines at scale. AI solves this by generating personalized openers from prospect research data. Provide a list of prospects with company name, recent news, job title, and one specific detail (funding round, product launch, LinkedIn post topic) and ask AI to write a personalized first line for each. Review and edit the best ones — AI's first lines are often 80% of the way there and need only light editing to sound natural. This approach cuts the time-per-email from 15 minutes to 2 minutes while maintaining the relevance that drives replies.

Step-by-step guide

Research the prospect

Provide the prospect's role, company, recent news, or a specific pain point for the AI to reference.

Define the offer and CTA

Specify what you are offering and the exact single action you want the prospect to take.

Write the email

Ask for under 100 words: personalized first line, problem, brief solution, and a low-friction CTA.

Test subject line variants

Generate 3 subject line options: benefit-driven, curiosity, and name-drop for A/B testing.

Ready-to-use prompts

Single personalized cold email

Write a cold email from [YOUR NAME], a [YOUR ROLE] at [YOUR COMPANY], to [PROSPECT NAME], [PROSPECT TITLE] at [PROSPECT COMPANY]. Context about the prospect: [SPECIFIC DETAIL — recent funding/news/LinkedIn post/pain point you know about]. Your offer: [WHAT YOU SELL AND THE SPECIFIC RESULT IT DELIVERS]. Proof point: [METRIC OR NAMED CLIENT]. CTA: a request for a [X]-minute call. Constraints: under 100 words total, no preamble ('I hope this finds you well'), open with the personalized detail, do not describe the product until sentence 3, end with a specific day and time suggestion not 'let me know if you're free.'

Why it works

The constraints are the most important part — each one removes a specific failure mode: preamble wastes the opening, product description before relevance signals self-absorption, open-ended CTAs get ignored. The specific day suggestion converts better than a vague time request.

3-email sequence

Write a 3-email cold outreach sequence for [YOUR COMPANY] selling [PRODUCT/SERVICE] to [PROSPECT ROLE] at [COMPANY TYPE]. Sequence structure: Email 1 (Day 1) — problem-led, no mention of product until last sentence, under 80 words. Email 2 (Day 5) — case study or social proof, one specific metric from a named client (use '[CLIENT]' as placeholder), under 100 words. Email 3 (Day 10) — honest breakup email with a direct question they can answer in one word. For each email: subject line, preview text (under 60 characters), body, and P.S. line. All emails must be readable in under 30 seconds.

Why it works

The three email types (problem-led / proof-led / breakup) address different psychological states of a prospect. The breakup email specifically outperforms follow-ups because it removes pressure and asks a direct yes/no question — reply rates on breakup emails routinely exceed earlier emails in the sequence.

Practical tips

✦Feed AI real prospect research (company news, LinkedIn post, funding round) to generate personalized first lines — AI personalization without real data produces generic output.
✦Keep cold emails under 100 words — every word over 100 reduces reply rate; shorter emails signal confidence and respect for the reader's time.
✦Write your subject line last, after the email body — the best subject line often comes from the strongest phrase inside the email itself.
✦Generate 5 subject line variants per email and A/B test them in your sending tool — subject line testing has the highest ROI of any cold email optimization.
✦For sequences, make email 3 a breakup email with a direct yes/no question — breakup emails consistently outperform standard follow-ups because they remove pressure and signal authenticity.