80% Reply Rate... How I use AI to see into the future before I hit send

What If You Could See Into the Future Before You Hit Send?

As sellers, we’re typically hitting send on an email and crossing our fingers.
But what if you could hit pause and see into the future first?

What if, before you send, you already knew how your prospect would feel reading your message? Where they'd pause. Where they'd hesitate. Where they'd delete.

That's what the approach I share in this article unlocked. 

Imagine AI not as your writer, but as your test subject. Your emotional decoder. Your way to reduce uncertainty and craft for response, not just delivery.

Let me show you how I used AI not to write an email but to predict how it would land.

What you’re about to see is the exact process I followed to turn a decent message into one that got an 80% reply rate and why AI made that possible in a way no human peer review ever could.

This Approach Got a Response Like This

and this…

and this…

This wasn’t a fluke, it was the result of a creative way to approach email messaging.

Not just writing better.
Predicting better.

These responses are part of a campaign that led to an 80% reply rate.

Sounds fake… it’s not.

And it mainly happened because in addition to using AI as a writer… I started using it as the recipient.

Let’s rewind and break down exactly how this approach works and how you can use it right now.

The Experiment

Task: Reach out to Scratchpad users at prospect accounts that fit our ICP. Learn if they would be open to sharing an exec summary with their operations leader.

Goal: Leverage end user champs to drive interest in Scratchpad at the leadership level. As well drive intro’s and higher response rates when we reach out directly to leaders.

The easy part: Identifying active users at our target accounts and the leaders we want to connect with. Thank you Koala!

The hard part: Getting a busy sales rep to help us sell to their leadership. No comp. No incentive. Just a favor!

How I Used AI to Get There

Here’s the process that led to that 80% reply rate:

1. Wrote a Draft Manually
   I started with my own message, without AI, to capture what I wanted to say vs asking AI to draft a message for me. The approach I typically would take is to provide AI with context and examples of what I want and ask it to draft the message.

2. Let AI Play the Role of the Recipient
   Instead of asking AI to edit this message, I asked AI to respond as if it were the seller receiving it. This helped me see into the future and understand what might land—or fall flat.

3. Got Feedback, Then Iterated
   Based on AI’s reaction, I updated the message with AI support. Then I asked AI again, "How would the rep respond now?"

4. Refined for Clarity & Motivation
   With each loop, I worked to reduce friction, clarify the ask, and add subtle motivations for the recipient to take action.

5. Tested Multiple Variants
   I created both long and short versions, more formal, more casual and got feedback on which versions what get to the desired result.

6. Selected the Final Version Based on Context
   I picked the one that felt most likely to get a “yes” from a user who already liked Scratchpad but might be cautious about internal politics.

What Seeing Into the Future Unlocked

Seeing how the message would actually land before sending it unlocked more than better copy it reshaped the strategy entirely:

• Shifted the tone from "I'm asking for a favor" to "We're in this together"

• Exposed vague wording like "chat" and helped reframe it as a meaningful, org-wide opportunity

• Removed friction by softening the CTA and giving the recipient agency

• Clarified intent so the message felt more like a collaboration than a pitch

This is the kind of nuance that’s nearly impossible to self-diagnose but becomes obvious when AI simulates how the recipient might react. It’s not just editing. It’s emotional insight, at scale.

Try This Yourself in 5 Minutes or Less

1. Take an email you plan to send — like a prospecting message.

2. Paste it into ChatGPT, and ask:
   “Imagine you’re the person receiving this—how would you respond and what’s missing?”

3. Give context about the recipient:

   • Their role

   • Goals

   • Challenges

   • Company

   • Even their tone/style preferences

4. Tell ChatGPT what you want the email to achieve — e.g., book a meeting, start a conversation, etc.

5. Use the feedback to improve the message.

6. Repeat until ChatGPT says the message would likely achieve your goal.

If this sparked an idea, forward it to a teammate or better yet, steal it and run your own version.

Want to see exactly how seeing into the future improved my initial message from “meh” to 80% reply rate?

The full ChatGPT conversation including the line that unlocked the 'yes' is in the premium thread below 👇

What You’ll Learn (and Why It’s Worth It)

1. How to craft a low-pressure referral ask that actually gets a response
→ Learn how to structure an ask that feels collaborative, not transactional — one that earned an 80% reply rate.

2. The psychology behind why this message works
→ Breakdown of how tone, clarity, and social dynamics (like championing, recognition, and risk reduction) impact outcomes.

3. Step-by-step message iterations with commentary
→ See how each word choice and sentence structure is adjusted to improve the likelihood of a “yes.”

4. How to personalize without overstepping
→ Learn how to reference someone's usage and past experience in a way that builds trust and rapport, not defensiveness.

5. A/B comparison of two versions with guidance on when to use each
→ Get a ready-to-use framework for deciding which message style fits different personas (e.g., cautious vs. confident champions).

Subscribe to Premium Guides to read the rest.