Everyone wants ‘personalization at scale.’ Recipients just see more spam. Here’s what happens when AI cold email crosses the line—and how founders can keep the channel alive.
Loading…
Playbooks on AI and Web3 product marketing, GTM and distribution. No spam, unsubscribe anytime.
Written by Kevan Shah, Web3 and AI product marketing strategist and founder of Hangryfeed.
Head of BD at Zealy and growth operator with 8+ years experience helping AI and Web3 products raise $26M+, scale to $2.2B monthly DeFi volume, and reach 1.2M+ users.
On a Tuesday morning, a VP of product opens her laptop, coffee in hand, and watches her inbox reload.
Thirty-six new “quick questions.”
Every subject line promises a “personalized” idea for her team. Every email opens with the same structure: a compliment scraped from her LinkedIn, a vague nod to her company’s latest launch, and a pitch that could have gone to any SaaS business on the planet.
She deletes the first one.
By email twelve, she stops reading entirely. Her spam filter starts learning what to do next time.
Somewhere else, a founder proudly watches their new AI outreach system hit its daily target: 500 “tailored” cold emails sent before lunch. The dashboard shows progress. The inboxes on the other side show noise.
That’s the tension we’re living in: AI makes it trivial to send more outreach than ever. For most recipients, that just means more spam.
But it doesn’t have to.
A few years ago, “personalization at scale” sounded like a dream.
Instead of blasting generic pitches, you could:
Builders are already doing this. One marketer described wiring up an AI agent that:
On paper, this looks responsible: guardrails, validation, tracking. Another practitioner uses AI to draft and customize emails—but insists on manually reviewing every message to “maintain a human touch” and avoid spam filters.
There’s a clear hope here: AI as a skilled assistant. It handles the tedious work while you stay in control of strategy, targeting, and tone.
Used this way, AI can make cold email better:
So why does the channel feel worse than ever?
The problem isn’t that AI writes emails.
The problem is what people do once the friction disappears.
When you can send hundreds of messages a day without effort, it’s incredibly tempting to lean on a simple formula:
This is where practitioners themselves start calling these tools “fancy spam machines.”
Why?
Because at that point, the core message is interchangeable. If you can swap the company name and nothing breaks, it’s not really personal. It’s a mass message with a few variables.
From the recipient’s side, that looks like:
Multiply this by thousands of senders doing the same thing and you get inbox fatigue. People stop reading. Filters step in. A channel once known for its scrappy effectiveness starts to burn out.
The irony is that many of the senders think they’re being thoughtful. They see “first name” and “company name” in the subject line and believe they’ve moved beyond spam.
Recipients don’t agree.
Under all the debate, one pattern keeps surfacing: spam is less about technology and more about how you use it.
AI doesn’t magically turn your email into spam. High-volume, low-relevance behavior does.
A few simple truths keep coming up:
One automation engineer summed up their frustration bluntly: spam is ineffective, so stop bothering people.
That line matters. It connects ethics with outcomes.
When you annoy people, you don’t just hurt your brand—you also make it harder for everyone else to use the channel at all.
It’s not just marketers complaining about bad open rates. Recipients are pushing back.
You see it in a few ways:
There’s a resigned tone to some of these reactions: “Most of them go straight to spam and we never see them anyway.”
That’s devastating if you rely on cold email as a growth lever.
Because it means:
If you’re a founder or marketer, this is the real risk: not just that your specific campaign underperforms, but that an entire acquisition channel erodes under the weight of automated noise.
Yet inside the same conversations, you’ll also find a quieter, more pragmatic stance:
AI cold email can work—if you use it to improve quality per email rather than to maximize quantity.
People who are seeing some success tend to have a few habits in common:
One thoughtful perspective puts it plainly: if your goal is to engage only in discussions where you can genuinely provide assistance, your outreach feels less like spam and more like an invitation.
That’s a useful lens for any workflow design.
A decade ago, cold email was a scrappy edge. You could write a handful of thoughtful messages a day, reach busy people directly, and stand out simply by being relevant.
Now:
The differentiator isn’t who can automate more. It’s who’s willing to do the work others aren’t:
AI didn’t kill cold email. It just removed friction. Whether the channel survives in a usable form will come down to how many teams choose restraint over brute force.
If you’re building or running AI-powered cold outreach, here’s how to keep your campaigns on the right side of the line.
Instead of asking, “How many emails can we send per day?” ask:
Your daily send volume should follow those answers, not the maximum your infrastructure can handle.
If your message could reasonably go to 80% of your list, your list is too broad.
Use AI to:
The goal: no more “this could be for anyone” emails.
The weakest use of AI is surface-level personalization: “Saw your recent post about X, loved it.”
The strongest uses look more like:
If the AI-generated line could be copy-pasted into another email with just a name change, it’s decoration, not depth.
At minimum:
Think of AI as a junior teammate. You wouldn’t let a new hire email hundreds of prospects unsupervised on their first week. Don’t let your model do it either.
If your metrics stop at opens and click-throughs, you’ll miss the signs that your outreach is slipping into spam territory.
Watch for:
If your numbers say “effective” but the humans say “annoying,” believe the humans.
AI cold email sits on a knife edge.
It can be the invisible assistant that helps you reach exactly the right people with thoughtful, relevant offers. Or it can be another stream of noise that teaches the world to hate your brand and the channel you rely on.
The difference isn’t the model you use. It’s the choices you make about volume, targeting, intent, and respect.