You can use AI to write a cold outreach email that actually gets replies by giving it clear context about who you're emailing and why, then editing its draft to add personalisation and a specific call to action. Tools like ChatGPT (OpenAI's AI chatbot) or Claude (Anthropic's AI assistant) work well for this - they can produce a focused, professional draft in seconds, even if you've never written a cold email before.
TL;DR
- You can use AI tools like ChatGPT (OpenAI's AI chatbot) or Claude (Anthropic's AI assistant) to draft, refine, and personalise cold outreach emails - even if you've never written one before.
- The key is giving the AI enough context: who you're emailing, why it matters to them, and what you want them to do next.
- AI works best as a drafting and editing partner, not a one-click solution - your human judgment still shapes the final email.
- Personalisation and a clear, specific subject line are the two factors most likely to determine whether your email gets opened and answered.
- Check each provider's current plan page for the latest on what's free or paid before you start.
Why Cold Emails Usually Fail (and What AI Can Fix)
Most cold emails fail for the same handful of reasons: they're too long, they open with "I" instead of something relevant to the reader, the subject line is vague, or the ask is unclear. These aren't talent problems - they're structure problems. And structure is exactly what AI is good at.
AI won't do your research for you, and it can't manufacture a genuine reason to reach out. But once you have that reason, it can help you express it clearly, concisely, and in a way that respects the reader's time. That's the core of how to use AI to help you write a cold outreach email that actually gets replies.
If you're new to using AI for everyday tasks, how to use AI in everyday life is a good place to build your baseline before going further.
Step 1: Do the Human Work First
Before you open any AI tool, answer three questions on paper or in a notes app:
- Who am I emailing, and what do I know about them specifically? (Their role, a recent project, something they published, a mutual connection - any real detail.)
- Why should this email matter to them, not just to me? (What problem does your ask solve for them, or what opportunity does it open?)
- What is the single action I want them to take? (A reply, a 20-minute call, a look at a portfolio - pick one, not three.)
These answers are the raw material your AI prompt will be built from. Without them, the AI will produce something generic, because it has nothing specific to work with.
Here's what those three answers might look like in practice - using a freelance designer reaching out to a small e-commerce brand as an example:
- "I'm emailing the founder of a small sustainable clothing brand. I noticed they recently relaunched their website but the product pages feel text-heavy and lack visual hierarchy."
- "Better product-page design could reduce the friction that causes visitors to leave without buying - that's a direct revenue problem for them."
- "I want them to reply and agree to a 15-minute call."
Once you have those three answers written down, you're ready to prompt.
Step 2: Write a Prompt That Gives the AI Real Context
The quality of your cold email depends almost entirely on the quality of your prompt. A vague prompt produces a vague email. A specific prompt produces something usable.
Here's a prompt structure that works, using the freelance-designer example above:
"Write a short cold outreach email from a freelance product designer to the founder of a small sustainable clothing brand. The designer noticed the brand recently relaunched their site but the product pages feel text-heavy and lack visual hierarchy. The email should open by referencing that specific observation, explain briefly how better design could reduce drop-off on product pages, and close with a single ask: a 15-minute call. Keep it under 150 words. Tone: warm, professional, not salesy."
Paste that into ChatGPT (OpenAI's AI chatbot) or Claude (Anthropic's AI assistant) and you'll get a structured draft within seconds. It won't be perfect - but it will be a real starting point, not a blank page.
A few prompt tips:
- Name the recipient's role and context, not just "a business owner."
- Specify the tone - "warm and direct" produces very different results from "formal."
- Set a word limit. Cold emails should be short. Telling the AI to keep it under a certain length forces it to prioritise.
- Tell it what to avoid - "don't use phrases like 'I hope this finds you well'" saves you an editing round.
Step 3: Edit the Draft Like a Human
AI drafts are starting points. Read the output and ask yourself:
- Does it sound like me? If not, adjust the phrasing.
- Is there anything generic I can replace with something specific? Swap "I noticed your brand is doing interesting work" for the actual observation you made.
- Is the subject line doing its job? More on this below.
- Would I reply to this if I received it? Honest answer only.
This editing step is where the real work happens - and it's also where most people skip too quickly. Spending five minutes here is the difference between a cold email that reads as genuine and one that reads as a template.
The Subject Line: Small Detail, Big Impact
The subject line determines whether your email gets opened at all. Anecdotally, shorter subject lines that reference something specific to the recipient tend to outperform clever or mysterious ones - but treat that as a starting hypothesis, not a rule, since what works varies by industry and audience.
Ask your AI to generate several subject line options:
"Give me five subject line options for this cold email. Keep each under eight words. Reference the product-page observation where possible."
Then pick the one that feels most natural, not the most creative. You can always test variations if you're sending at volume.
Before and After: What AI-Assisted Editing Actually Looks Like
Here is a real before-and-after showing how AI-assisted editing transforms a generic cold email into a reply-worthy one.
Before (generic, AI first draft without good prompting):
Subject: Reaching out
Hi there,
I hope you're doing well. I'm a freelance designer and I'd love to work with your brand. I have lots of experience in e-commerce design and I think I could add a lot of value. Let me know if you'd like to chat!
After (AI draft with specific prompt, then lightly edited):
Subject: Your product pages - a quick thought
Hi [Name],
I came across your site after your relaunch last month - the new branding looks great. One thing I noticed: the product pages are quite text-heavy, which can make it harder for shoppers to make a quick decision.
I'm a freelance product designer who's worked on similar challenges for e-commerce brands. I have a few ideas that might be worth a quick look.
Would a 15-minute call this week work for you?
[Your name]
The second version is specific, short, opens with the recipient in mind, and has one clear ask. The AI did the structural heavy lifting; the specific detail ("text-heavy product pages," "relaunch last month") came from the human doing the research first.
Step 4: Personalise at Scale Without Losing Quality
If you're sending more than a handful of cold emails, you can use AI to help you personalise each one without starting from scratch every time.
Create a base prompt template with placeholders:
"Write a cold outreach email from [your role] to [recipient's role] at [company type]. Key observation: [specific thing you noticed]. Value angle: [why this matters to them]. Call to action: [single ask]. Tone: [tone]. Under 150 words."
Fill in the brackets for each recipient before you prompt. This keeps the process efficient while ensuring every email has something real and specific in it - which is what separates replies from silence.
For a broader look at how this kind of workflow fits into your day, how to use AI to save time covers the bigger picture.
A Note on Tone and Authenticity
Cold emails written entirely by AI - without editing - tend to share a particular quality: they're technically correct but feel slightly off. Too smooth, too complete, slightly impersonal. Recipients who receive a lot of email often sense this, even if they can't name it.
The fix is simple: read your email aloud before sending. If you stumble over a sentence or it doesn't sound like something you'd say, rewrite that sentence yourself. One or two human-sounding edits can shift the whole feel of the email.
If you want to build real confidence with prompting and editing AI output - not just for emails, but across everyday tasks - AILE, a gamified AI-learning app, offers short, practical lessons on prompting, editing AI output, and building confidence with AI tools. It's worth a look if you'd rather learn through doing than reading documentation.
What to Do After You Send
Cold email is rarely a one-shot activity. If you don't hear back:
- Follow up once, briefly, a few days later. Reference your original email and keep it to two or three sentences.
- Ask AI to write the follow-up too - give it your original email as context and ask for a short, non-pushy check-in.
- Track what works. If you're sending multiple emails, note which subject lines and openings get replies, and refine your prompts accordingly.
For more on writing emails with AI across different contexts - not just cold outreach - how to use AI to write emails covers the full range.
Frequently Asked Questions
Can I use AI to write a cold email even if I'm not a good writer?
Yes - that's one of the most practical uses of AI for everyday tasks. Tools like ChatGPT (OpenAI's AI chatbot) or Claude (Anthropic's AI assistant) can produce a solid first draft from a short description you provide. Your job is to review it, add personal details, and adjust the tone. You don't need to be a strong writer to do that effectively.
Will the recipient know I used AI to write the email?
Not if you edit it properly. AI-generated drafts often sound slightly generic or overly polished, which is why the editing step matters. Read it aloud, swap in specific details about the recipient, and adjust any phrasing that doesn't sound like you. A well-edited AI draft is indistinguishable from a well-written human one.
What should I never let AI do when writing a cold email?
Don't let AI invent facts about the recipient, fabricate shared connections, or make promises you can't keep. Always verify any claim the AI includes - it can occasionally hallucinate details. The research behind personalisation has to come from you; AI just helps you express it clearly.
How many times should I follow up on a cold email?
There's no universal rule, but a short, polite follow-up a few days after the first email is widely considered reasonable. Beyond two or three follow-ups with no response, most people stop. AI can help you write follow-up messages too - just give it context about what you sent originally.
Are AI writing tools free to use?
Many AI tools offer a free tier with some limitations, and paid plans unlock more features or higher usage. Availability and limits vary by provider and change regularly - check each provider's current plan page before you start to confirm what's included.
What's the biggest mistake people make when using AI for cold emails?
Sending the first draft without editing it. AI gives you a strong starting point, but it doesn't know the recipient, your relationship history, or the subtle nuances of your industry. Treat every AI draft as a rough draft, not a finished product.
Keep going with AILE
Learning AI shouldn't feel like falling behind. AILE, a gamified AI-learning app, turns it into short, friendly, hands-on lessons you can actually finish - no jargon, no gatekeeping. Join the waitlist for early access →