Automated sales email sequence with personalised follow-ups

Most sales teams know the feeling: you've crafted what you think is a compelling cold email, hit send to fifty prospects, and watch the reply rate flatline. The problem isn't your product or your pitch. It's that each prospect receives the exact same generic message, and in 2026, generic is invisible. Personalisation works, but it requires hours of manual research, pulling company details, finding hiring patterns, reviewing recent news, then manually tweaking each email. At scale, this becomes impossible. What if you could generate truly personalised cold email sequences without touching a single email manually? Not mad libs substitution where you just swap in names, but actual intelligence about each prospect built into the opening line, the value proposition, and the call to action. The good news: with the right workflow, you can. You feed in a list of prospects with their LinkedIn profiles or company URLs, and out comes a fully personalised email sequence ready to deploy. This post shows you how to wire ChatGPT Writer, ColdConvert AI, and HyperWrite together using an orchestration platform, so every prospect gets a genuinely tailored message without a human touching it after setup. For more on this, see Wispr Flow AI vs ChatGPT Writer vs HyperWrite: AI Writing.... For more on this, see AI Email Outreach for Sales Teams: Finding the Right Tool....

The Automated Workflow

The workflow has five stages: prospect intake, research enrichment, email generation, sequence creation, and delivery. Here's how the pieces fit together.

Why orchestration matters.

You could theoretically do this manually: research a prospect, paste their details into ColdConvert AI, refine the output in HyperWrite, then move to the next one. But the moment you have fifty prospects, this method collapses. An orchestration platform (Zapier, n8n, or Make) sits in the middle and automates the glue between tools so data flows automatically from one step to the next. For this workflow, we recommend n8n because it handles conditional logic and API calls better than Zapier for more complex sequences, and it's cheaper at scale. If you prefer a visual, low-code interface, Make works just as well.

Step 1: Trigger from your CRM or spreadsheet

Start with a list of prospects. This could live in a Google Sheet, Airtable, or your CRM (HubSpot, Pipedrive, etc.). The trigger watches for new rows. Example: A Google Sheet with columns for Name, Company, Industry, LinkedIn URL, Recent News. In n8n, use the Google Sheets trigger node set to watch for new rows.

Step 2: Enrich prospect data

Before generating an email, gather context. You can use a simple HTTP request to grab public company data or manually include enrichment in your spreadsheet. For lightweight enrichment, use the built-in Gemini 2.5 Flash via API to analyse the LinkedIn URL or company domain and extract key information (company size, recent funding, job postings).

CopyPOST https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash:generateContent?key=YOUR_API_KEY { "contents": [ { "parts": [ { "text": "Extract company info from this LinkedIn page: [URL]. Return: company size, recent news, industry, hiring activity." } ] } ]
}

Store the response in a variable so you can pass it to the email generator.

Step 3: Generate initial email with ColdConvert AI

ColdConvert AI is specifically built for this: you give it prospect details and it generates a personalised cold email. Use the API endpoint to automate this.

CopyPOST https://api.coldconvert.ai/v1/generate-email { "prospect_name": "Sarah Chen", "company": "TechVenture Inc", "industry": "B2B SaaS", "context": "Recently hired 5 engineers; Series A funding in Q4 2025", "your_product": "Sales automation platform", "tone": "professional but conversational", "max_words": 150
}

The response comes back with a personalised email draft. Don't stop here; pass it to the next step.

Step 4: Refine with HyperWrite (optional but recommended)

HyperWrite's API allows you to apply tone adjustments, grammar checks, and writing style refinements. If your email from ColdConvert AI feels slightly off, send it through HyperWrite to tighten the language.

CopyPOST https://api.hyperwrite.com/v1/refine { "text": "[email from ColdConvert AI]", "tone": "confident_and_warm", "style_guide": "concise B2B", "improvements": ["grammar", "clarity", "impact"]
}

This is optional; many teams skip it if ColdConvert AI output is already strong.

Step 5: Build the sequence

A single email rarely converts. Create a follow-up sequence: initial email, then a second email 3 days later, then a third at day 7. In your orchestration platform, create three separate email sends scheduled at different intervals. Use the same personalisation data for all three, but vary the message. Example sequence structure: - Email 1 (Day 0): "I noticed you just [recent company event]. Quick thought on how [your product] could help with [specific problem]."

• Email 2 (Day 3): "Following up on my previous note. I've helped [similar company type] reduce [metric] by [amount]."
• Email 3 (Day 7): "Last touch: if you're not interested, no problem. But if [specific scenario] is on your roadmap, I think we should talk." Use Make or n8n's delay node to space these out.

Step 6: Send via your email platform

Finally, send the emails through your connected email tool (Gmail, Outlook, or a dedicated email API like SendGrid or Mailgun).

CopyPOST https://api.sendgrid.com/v3/mail/send { "personalizations": [ { "to": [{"email": "prospect@company.com"}], "subject": "Quick thought on [company name]'s hiring growth" } ], "from": {"email": "you@yourcompany.com"}, "content": [ { "type": "text/html", "value": "[personalised email body from step 4]" } ]
}

Tying it together in n8n

Here's the node sequence: 1. Google Sheets trigger (watches for new rows) 2. Gemini 2.5 Flash node (enriches prospect data) 3. ColdConvert AI HTTP request node (generates email) 4. HyperWrite HTTP request node (refines email, optional) 5. Delay node (wait 0 days for initial send) 6. SendGrid node (sends email 1) 7. Delay node (wait 3 days) 8. Edit fields node (swap in email 2 copy) 9. SendGrid node (sends email 2) 10. Delay node (wait 7 days total) 11. Edit fields node (swap in email 3 copy) 12. SendGrid node (sends email 3) Connect each node's output to the next node's input. The data flows automatically, and each prospect gets their own sequence without manual intervention.

The Manual Alternative

If you prefer more control over each email, use this approach: export your prospect list, run each row through ColdConvert AI manually, review and refine the output in HyperWrite, then schedule sends in your email platform. This gives you a chance to tweak messaging and catch any odd phrasing. The trade-off is time: this scales to maybe 20 prospects per week before burnout. For most teams, the automation is worth it. The emails are genuinely personalised, the system is consistent, and you can review a sample of outputs before running the full workflow. For more on this, see Personal brand building with consistent social media content.

Pro Tips

Rate limiting and API costs.

ColdConvert AI and Gemini both have rate limits.

If you're sending to 100+ prospects, space out your requests across an hour or use Zapier's built-in rate limiting to avoid hitting the ceiling. This slows down the workflow but prevents API errors.

Test with a small batch first.

Run the workflow on 10 prospects, then review the emails before scaling to 100. You'll catch any data formatting issues or odd phrasing that could hurt reply rates.

Track replies automatically.

Connect your email platform's webhook to your CRM so replies are logged automatically. This closes the loop and lets you measure which emails and sequences actually work for your audience.

Customise the prompt for your product.

The ColdConvert AI prompt should reflect your actual value prop. Change "Sales automation platform" to whatever you sell, and include specific metrics your customers care about.

Monitor deliverability.

Personalised emails sent from a shared domain can trigger spam filters if volume is high. If you're sending more than 50 per day, use authenticated sender domains and warm up your sending reputation gradually.

Cost Breakdown