0:00
/
0:00

Built an AI Agent That Gets 10x Cold Email Responses

How workflow orchestration turns <1% response rates into 10%

Most cold emails get a <e1% response rate. Send to 1,000 people, get 10 responses.

I built a workflow that gets 100 responses from the same 1,000 emails. Here’s how the system works.

The Problem With Cold Email Automation

You’ve seen them:

“Hey [FIRSTNAME], I noticed [COMPANY] and thought we could help with [GENERIC PROBLEM]...”

They fail because they’re obviously templates. The automation is lazy. Just find-and-replace on a few variables.

Real personalization means using actual information about each person. Not inserting their name. Referencing their specific situation.

What I Built

A workflow that chains together three AI agents:

  1. Research agent - Searches online for context about each lead

  2. Writing agent - Composes unique emails using all available data

  3. Sending agent - Delivers at scale

The workflow runs on our last 500 users. Each person gets an email that references:

  • Their exact signup date

  • Tools they’re using

  • Their role

  • Information from their public profiles

Here’s what one looked like:

“Hey Johann, thanks for signing up to Needle on October 15th. As a co-founder interested in Google Drive workflows, I wanted to reach out personally. I’d love to learn more about how Needle fits into your workflow...”

Not a template. The AI wrote this after researching the specific person.

How It Works

Start with structured data

Google Sheet with basic lead info:

  • Name, email, company

  • Tools they use

  • Signup date

  • Role

This comes from our database and onboarding flow. If you capture any user data, you have this.

AI research layer

For each row, the workflow:

  • Searches for their online presence

  • Pulls relevant context (website, LinkedIn, recent posts)

  • Adds this to what we already know

LLM writes each email

The writing agent gets:

  • Your database info

  • The research results

  • Your messaging guidelines

It generates a unique email. Not a template with variables filled in. An email written from scratch using all that context.

Automated delivery

Emails send automatically. Each one personalized. At scale.

The Results

500 emails sent. Response rate roughly 10%.

People booked calls. Some thanked me for the outreach. Many asked questions about specific workflows.

Why? The emails were relevant. They referenced real information about each person’s situation.

Sure, recipients probably knew some automation was involved. But when the message actually relates to their work, they don’t care about the how.

The Math

1,000 prospects:

  • Generic template: 10 responses

  • AI-personalized workflow: 100 responses

That’s 90 additional conversations. 90 more chances to validate ideas, book demos, close deals.

Build This Yourself

You need:

  1. Lead data in a spreadsheet - Whatever context you have about each person

  2. The workflow template - Available free in Needle’s template library

  3. Your email connected - Gmail, Outlook, whatever you use

  4. Basic customization - Adjust tone and offer to match your brand

The workflow handles research, writing, and sending.

Why This Works

Three things:

Scale - Reach hundreds of people without manual work per email

Personalization - Each message uses actual information about the recipient

Relevance - When people see you understand their situation, they respond

Yes, it’s automated. No, that doesn’t matter if the email is genuinely useful to them.

Try It

Grab the free workflow template and run it on your next campaign.

If you see results, let me know. I’m collecting data on what messaging patterns work best.

Comments

User's avatar