How to Use AI Writing Tools Like ChatGPT, Claude, Gemini and Copilot for Email Outreach, Follow-Up Sequences and Personalized Campaigns
On this page
AI as a Starting Point, Not a Send Button
AI writing tools have changed how sales teams, marketers, and business development professionals approach email outreach. A task that used to take 20 minutes per email -- researching a prospect, drafting a personalized message, writing a compelling subject line, and creating a follow-up sequence -- can now be reduced to 5 minutes with AI assistance.
But the key word is assistance. The best use of AI in email outreach is not to write emails and send them. It is to generate drafts, variations, and ideas that a human then edits, personalizes, and approves. An AI-generated email sent without editing is obvious to recipients: it reads as polished but generic, technically correct but emotionally flat. The recipient may not consciously identify it as AI-written, but they feel that something is off -- the email does not sound like a real person talking to them.
The workflow that produces the best results: use AI to generate a first draft quickly, then spend the time you saved on genuine personalization that AI cannot do. The combination of AI speed and human authenticity outperforms either one alone.
Writing Effective Prompts for Cold Emails
The quality of AI-generated email depends almost entirely on the quality of the prompt. A vague prompt produces vague output. A specific, detailed prompt produces output that needs minimal editing.
A weak prompt: "Write a cold email to a marketing manager about our email extraction tool."
The output will be generic because the prompt provides no context about the specific prospect, no information about their likely pain points, no guidance on tone, and no constraints on length.
A strong prompt: "Write a cold email to [Name], the marketing manager at [Company]. [Company] is a B2B SaaS company with approximately 50 employees that sells project management software. They appear to be growing quickly based on recent job postings for SDRs and account executives. The email is from a sales rep at a company that sells email extraction software. The pain point to address: as [Company] scales their outbound sales team, they need to efficiently build prospect lists from multiple data sources. The email should be 4-5 sentences, casual and direct (not salesy), and end with a specific question rather than a generic call to action. Do not use the phrases 'I hope this email finds you well,' 'touching base,' 'synergies,' or 'circle back.' Do not use exclamation marks."
This prompt produces dramatically better output because it gives the AI specific context about the prospect, identifies a specific pain point tied to observable evidence, provides tone and length constraints, and explicitly excludes the cliche phrases that mark an email as template-generated.
Personalizing at Scale
True personalization -- the kind that makes a recipient feel like the email was written specifically for them -- requires information that AI does not have unless you provide it. The most effective personalization framework gives the AI a "personalization fact" for each prospect.
Gather one specific fact about each prospect: a recent LinkedIn post, a company announcement, a conference talk they gave, a product launch, a hiring milestone, or a specific challenge visible from their public presence. Then prompt the AI with this fact: "Write an opening line for a cold email to [Name] at [Company] that references the fact that they recently [specific fact]. Make the opening line feel natural, not forced -- it should connect the personalization to the reason for the email, not just mention it as a disconnected compliment."
This approach scales because the opening line is the only part of the email that needs to be unique per prospect. The body of the email (the value proposition, the social proof, the call to action) can be templated. The personalized opening hooks the recipient's attention, and the templated body delivers the message.
For batch personalization, structure your prospect data in a spreadsheet with columns for name, company, title, personalization fact, and industry. Then prompt the AI to generate personalized opening lines for each row: "For each of the following prospects, write a personalized opening line that references their specific fact and naturally transitions to the topic of [your product/service]. Keep each opening to 1-2 sentences."
Creating Follow-Up Sequences
Follow-up emails are where AI provides the most leverage. Writing 4-5 follow-up variations that each offer a different angle, a different value proposition, or a different call to action is tedious and repetitive work that AI handles well.
Prompt for a complete follow-up sequence: "Write a 5-email follow-up sequence for a cold outreach campaign selling [product/service] to [target audience]. Email 1 is the initial outreach (already written). Write emails 2-5 as follow-ups. Each follow-up should: be sent [X] days after the previous email, be shorter than the previous email (email 5 should be 2-3 sentences), offer a different angle or piece of value (do not just say 'following up on my last email'), and include a clear call to action. Angles to use across the sequence: [list 3-4 different value propositions, case studies, or relevant insights]. Tone: conversational and direct. Never guilt the recipient for not responding."
The AI will generate a sequence where each email builds on the previous one while offering fresh content. Edit each email to match your voice, check that the facts and claims are accurate, and adjust the timing based on your sales cycle.
Subject Line Generation
Subject lines are the highest-leverage element of any email campaign, and they are also the element where AI's ability to generate large numbers of variations is most useful.
Prompt: "Generate 20 subject lines for a cold email selling [product/service] to [target audience]. The email's main value proposition is [specific benefit]. Requirements: each subject line should be under 50 characters, no clickbait or misleading promises, no ALL CAPS, no excessive punctuation, no emojis. Vary the approaches: some should be direct, some should ask a question, some should reference a specific challenge, some should use a number or statistic. Do not use the word 'revolutionary,' 'game-changing,' 'unlock,' or 'supercharge.'"
From 20 subject line options, you can select 3-4 to A/B test, giving you data on which approach resonates with your specific audience.
Avoiding the AI Sound
AI-generated emails share recognizable patterns that savvy recipients notice. Avoiding these patterns is essential for maintaining authenticity.
Common AI tells to edit out: unnecessarily formal language ("I wanted to reach out to explore potential synergies"), hedging phrases ("I believe," "it seems like," "if I am not mistaken"), excessive enthusiasm (three exclamation marks in five sentences), perfect grammar in contexts where casual language is more natural, and formulaic structure (compliment, then pivot, then pitch, then CTA, every time, in exactly that order).
The simplest editing rule: read the AI draft out loud. If it does not sound like something you would actually say in a conversation, rewrite those parts. AI is excellent at structure and content; humans are better at voice and nuance.
Building the Outreach List
Before you can write emails -- with or without AI -- you need a list of people to email. Export prospect data from LinkedIn, company websites, industry directories, conference attendee lists, and professional association member pages. Save relevant pages as HTML and upload to Email Extractor to extract email addresses. Download as CSV with sources, then add personalization columns (personalization fact, industry, company size) to the export for use in AI-assisted batch personalization.
