I’ll admit it—I hate writing cold emails.
But as a formal salesperson, it has become an important addition to my prospecting arsenal. Over the years, I’ve developed my own system: research the prospect, find a recent trigger (a press release, podcast quote, or even a hiring post), and connect it back to a sharp insight about their business. It’s time-consuming, but it’s worked. My reply rates have been decent. My meetings booked? Respectable.
But then came AI. And suddenly, everyone was saying, “You can automate outreach with just a few clicks.”
I was skeptical—but curious. So, I ran a one-week experiment earlier in April this year.
The Experiment Setup
The subject detail was to showcase my freelance lead generation capabilities. I split my prospect list into two equal groups:
- Group A: Emails crafted by me, fully manual.
- Group B: Emails generated by AI tools (ChatGPT and a specialized cold email AI assistant).
Same target audience (mid-level to senior decision-makers in B2B SaaS).
Same product. Same call-to-action. Just two different writers: me vs. the machine.
The Email Flow Setup
Each email followed a structure I’ve always believed in:
Personalization ➝ Pain Point ➝ Value Proposition ➝ CTA
I gave the AI a detailed brief:
- Buyer persona
- Common industry pain points
- Product features and differentiators
- Tone preferences: conversational, not too salesy
The tool spat out draft after draft—some usable, some clunky. But it was fast.
What the AI Got It Right
- Blistering speed
Writing 40 custom emails typically takes me 3–4 hours. AI did it in 7 minutes. That alone was impressive. - Structural consistency
The emails were cleanly formatted, grammatically sound, and followed logical flows. No “wall of text” issues. - A/B testing potential
I could generate multiple versions of the same email instantly—great for testing subject lines or CTAs. - Buzzword filtering
Interestingly, the AI avoided fluffy sales jargon when I prompted it properly. No “synergies” or “cutting-edge solutions.”
Where AI Fell Short
- Context blindness
One prospect recently merged with another company. AI missed that nuance and wrote as if the org structure hadn’t changed. That would’ve been a huge miss if sent. - Overgeneralization
Phrases like “I noticed your company is in growth mode” sound personal at first—until you realize it’s vague and could apply to anyone. - Weak CTAs
AI defaulted to soft closes like “Would love to hear your thoughts”. I found myself rewriting most CTAs to be more specific and actionable. - Tone mismatches
Some emails felt… robotic. Even with prompts like “make it sound human,” there was still something slightly off—like a rep who just memorized the playbook but doesn’t get the buyer.
The Performance Breakdown
| Metric | Human-Written | AI-Written |
| Open Rate | 42% | 39% |
| Reply Rate | 11% | 7% |
| Meetings Booked | 9 | 6 |
| Time Spent Writing | ~4 hours | ~10 mins |
Not bad for AI. But not magic either.
Human Insights: The Intangibles
Some things just can’t be automated (yet):
- Gut feel for timing: I bumped one email a few days later because I saw the company was hosting a webinar that week. AI wouldn’t have known that.
- Personal references: I once opened an email with “I saw your CEO on a panel about AI bias and completely agreed with his point on explainability.” That got a reply in 15 minutes.
- Natural curiosity: When I write an email, I’m genuinely curious about the person I’m reaching out to. AI can fake interest—but buyers can feel when it’s real.
The Hybrid Model That’s Actually Working
By mid-week, I stopped doing a strict AI vs human split. Instead, I tried something new:
AI-first draft ➝ Human second pass
Let AI lay the foundation. Then I add:
- A stronger hook
- A personalized reference
- A sharper CTA
- A touch of me
These “hybrid emails” performed just as well as my originals—and took half the time.
Final Thoughts: Will I Keep Using AI?
Absolutely—but not on autopilot.
AI now sits in my outbound workflow as a drafting assistant, not a copywriting replacement. It helps me scale ideas faster, test more variations, and stay consistent even on busy weeks.
But real B2B selling still needs a human touch—especially when you’re trying to start conversations, not just clicks.
You Might Also Like…
Personalization Engines: AI in B2B Website and Email Experiences
ChatGPT-5 Hype Check: What B2B Marketers Should Really Take Away
The Definitive Guide to B2B Marketing Tools for Growth in 2025
Disclaimer note:
The opinions expressed in this post are those of the author. They do not purport to reflect the opinions or views of any company or their associates.
Follow me on my digital channels:
Website: https://asparkofb2b.com/
Facebook (A spark of B2B) https://www.facebook.com/profile.php?id=100089042254709
Twitter (aSparkofB2B) https://twitter.com/aSparkofB2B
LinkedIn (a-spark-of-b2b) https://www.linkedin.com/company/a-spark-of-b2b
Medium https://medium.com/@dexterwrites2022
#AIinSales #ColdEmail #B2BMarketing #SalesProspecting #LeadGeneration #SalesTips #AIForBusiness #EmailOutreach #B2BSales #SalesStrategy
