Your sales team spends Tuesday morning on LinkedIn, Wednesday researching company websites, and Thursday crafting emails one by one. By Friday, they have maybe 20 personalised outreach messages ready. A competitor with an automated pipeline sends 500 that same week, with better targeting and response rates. The difference isn't talent; it's workflow design. Building a sales outreach engine that moves leads from research to personalised email sequences requires stitching together several AI tools. The problem is usually that manual handoffs between steps kill momentum and create bottlenecks. You end up with spreadsheets, Slack messages, and someone manually copying data between tools. This Alchemy stack eliminates those handoffs entirely, moving prospects through your funnel with zero human intervention between discovery and send. We'll build a workflow that takes a list of target companies, researches key decision-makers and their problem, generates personalised cold emails tuned to their specific situation, and sends them with tracking pixels embedded. The entire process runs on a schedule you control. For more on this, see SaaS onboarding email sequence with video demos and cost ....
The Automated Workflow
Start with n8n as your orchestration engine. It's more flexible than Zapier for complex logic, and it handles API calls without the per-call pricing overhead that can add up quickly with high-volume outreach. For more on this, see AI Email Outreach for Sales Teams: Finding the Right Tool....
Step 1: Ingest target company data
Your source is a Google Sheet, CSV file, or API endpoint. Each row contains a company name, industry, and the LinkedIn profile URL of your target contact. n8n polls this sheet every morning.
GET /google/sheets/v4/spreadsheets/{spreadsheetId}/values/Sheet1?range=A:E
The payload looks like this:
json
{ "company_name": "TechCorp Ltd", "industry": "SaaS", "target_url": "https://www.linkedin.com/in/alice-johnson", "company_website": "https://techcorp.example.com"
}
Step 2: Research the prospect and company
Feed the company website and LinkedIn URL into ChatGPT Writer, which scrapes publicly available information and generates a brief prospect summary. You're not writing the full email yet; this is raw intelligence gathering. Use ChatGPT Writer's API to send the request: For more on this, see Wispr Flow AI vs ChatGPT Writer vs HyperWrite: AI Writing....
[POST](/tools/post) https://api.chatgpt.com/v1/research
Content-Type: application/json { "company_url": "https://techcorp.example.com", "linkedin_url": "https://www.linkedin.com/in/alice-johnson", "instruction": "Extract: company size, recent funding/news, key problem in their industry, target role responsibilities. Limit to 150 words."
}
Store the response in a variable. n8n will pass this forward to the next step.
Step 3: Generate personalised email copy
This is where ColdConvert AI takes over. Send it the prospect intelligence, your product/service description, and any custom angle (e.g. "we help SaaS companies reduce churn"). ColdConvert generates a personalised subject line and email body that references specific company context.
POST https://api.coldconvert.ai/v1/generate-email
Authorization: Bearer YOUR_API_KEY
Content-Type: application/json { "prospect_name": "Alice Johnson", "company_name": "TechCorp Ltd", "role": "VP Product", "company_summary": "{{nodeOutput.step2.research_summary}}", "your_value_prop": "We reduce customer churn by 23% through predictive analytics", "personalization_angle": "tech industry, post-Series B, high churn risk", "tone": "professional but conversational"
}
ColdConvert returns:
json
{ "subject_line": "Alice, one thing TechCorp's competitors are doing differently", "email_body": "Hi Alice,\n\nI noticed TechCorp recently expanded into EMEA..."
}
Step 4: Add tracking pixels and optimise for deliverability
Before sending, add a unique tracking pixel from your email tracking service (Mailchimp, HubSpot, or similar). You also want to validate the email address format and run a quick domain reputation check to avoid spam folders. Use BurnRate to monitor your API costs at this point. Each research call to ChatGPT Writer (using Claude Opus 4.6 or GPT-4o under the hood) costs roughly £0.02 to £0.05 depending on response length. With 500 prospects weekly, costs add up fast. BurnRate flags when your spending drifts above your optimisation rules.
POST https://api.burncrate.io/v1/analyse-costs
{ "provider": "openai", "model": "gpt-4o", "total_tokens": 5200, "session_id": "outreach_pipeline_weekly"
}
Step 5: Send the email and log the event
Now use Payman AI to handle the actual send. Payman AI is particularly useful here because it can orchestrate human verification steps if your compliance rules require a second pair of eyes on high-value prospects. But for most outreach, you configure it to send automatically.
POST https://api.payman.ai/v1/send-email
Authorization: Bearer YOUR_PAYMAN_KEY
Content-Type: application/json { "to": "alice@techcorp.example.com", "subject": "{{nodeOutput.step3.subject_line}}", "body": "{{nodeOutput.step3.email_body}}", "from": "your_sales_email@yourdomain.com", "tracking_pixel": "{{nodeOutput.step4.pixel_url}}", "schedule_send": "immediate", "human_review": false
}
Step 6: Log results and trigger follow-up actions
Store the outcome (sent, bounced, clicked, opened) in a database or back to your Google Sheet. n8n can write to a second sheet tab called "Outreach Log" with timestamps, email addresses, and engagement data.
javascript
{ "prospect_email": "alice@techcorp.example.com", "sent_timestamp": "2026-03-14T09:30:00Z", "tracking_id": "px_12abc34def", "email_opened": null, "link_clicked": null, "status": "sent"
}
Set up a separate n8n workflow that polls your tracking data every 6 hours. When someone opens the email or clicks a link, automatically trigger a follow-up sequence 3 days later if there's no reply. This keeps momentum without manual intervention.
The Manual Alternative
If you prefer more control over email personalisation, skip ColdConvert AI and instead use ChatGPT Writer directly with custom prompts for each prospect. You'll write the emails yourself but still benefit from the research and tracking automation. Alternatively, use Make (Integromat) instead of n8n if you want a visual builder with less coding. Make's pre-built modules handle email sending and spreadsheet updates without custom API calls, though you'll lose flexibility for complex conditional logic.
Pro Tips
Watch your email warm-up period.
If you're sending from a new domain or mailbox, start with 20-30 emails on day one, then gradually increase daily volume over two weeks.
This keeps mailbox providers from flagging you as spam. BurnRate doesn't track email reputation, so monitor your bounce rate separately.
Use variable send times per time zone.
If your prospects are scattered across UK, US, and APAC regions, configure n8n to stagger sends based on recipient time zone. Emails arriving at 09:00 local time get higher open rates. Add a time zone field to your prospect sheet and use n8n's date manipulation nodes.
Cache research results.
If multiple team members are targeting companies in the same industry, store research summaries and reuse them. Don't call ChatGPT Writer five times for five different prospects at the same company. This cuts your API costs significantly.
Set rate limits on API calls.
Both ChatGPT Writer and ColdConvert AI have request limits. Configure n8n to queue requests and space them out; for example, one research call every 30 seconds instead of batching 50 at once. This prevents hitting throttles and avoids temporary blocks.
Monitor reply-to-send ratio.
Track what percentage of sent emails get replies within 7 days. If it drops below 2%, your personalisation angle may have drifted. Pull a sample of recent emails and ask ColdConvert AI to suggest improvements based on low-engagement patterns.
Cost Breakdown
| Tool | Plan Needed | Monthly Cost | Notes |
|---|---|---|---|
| n8n | Pro (self-hosted) or Cloud | £40–£100 | Self-hosted is cheaper at scale. Cloud plan sufficient for <10k workflows/month. |
| ChatGPT Writer | Premium API access | £15–£50 | Depends on research token usage. ~500 prospects weekly costs ~£30/month. |
| ColdConvert AI | Starter | £29 | Flat rate for up to 500 emails/month. Scaling to 2000+/month requires Growth plan (£79). |
| Payman AI | Standard | £0–£50 | Free tier handles 100 sends/month. Pay-per-send after that (~£0.10/email). |
| BurnRate | Pro | £19/month | Tracks spending across all your AI tool subscriptions. Saves money through optimisation alerts. |
| Email tracking (HubSpot free tier) | Free | £0 | Or Mailchimp free for basic open/click tracking. |
Total estimated cost for 500 emails/week: £150–£200/month.
At that volume, your cost per outreach email lands around £0.07 to £0.10, including research, personalisation, sending, and tracking. A sales hire for outreach costs £3000–£4000/month. You're paying for the equivalent of 3–5 weeks of a junior salesperson's time.