Why Leads Ignore AI Emails (& How to Fix Your Messaging Approach)

Most AI-created outreach is instantly recognized as spam.

Not because AI can’t write well. It can, often better than most people (if we’re being brutally honest). 

The real problem is that most people feed AI bad instructions, poor data, weak personalization samples, and no strategic framework.

After reviewing thousands of messages and onboarding dozens to AiSDR, one thing has become very clear to me:

Personalization only works when it’s structured.

Good AI messaging isn’t about “sounding human”. It’s about establishing clear rules on what to do and what not to do.

4 fixable reasons AI outreach fails

Many AI messages fail for simple and fixable reasons as people look for patterns that make your message easy to write off as AI-generated noise.

At AiSDR, we handle these under the hood so you don’t have to worry about them sinking your campaign.

Em-dash overuse

AI loves em-dashes. I love em-dashes. Many humans and authors like Stephen King love em-dashes.

I don’t hear anyone accusing Stephen King of using AI to create his work.

But fairness doesn’t matter here. Perception does.

The hard truth is that some people instantly see messages as “AI spam” the moment they spot an em-dash. They’ll delete the message purely on that basis.

If you’re selling something and want to max your chances of a deal, you should probably avoid em-dashes.

If you don’t want to abandon them entirely, a simple compromise is:

• Swapping em-dashes for spaced en-dashes (“ – “)
• Save em-dashes for one-time intentional uses (our content manager likes to call this your “mic drop” moment)

A tiny detail. Yet a big trust signal.

AI’s (limited?) vocabulary

“Delve”. “In today’s ever-changing world”. “Crucial”.

Pre-2022, these phrases made your writing sound polished and dynamic. Post-2022, they make your emails sound like they’re written by ChatGPT.

You don’t lose anything by cutting them. In fact, you gain:

• Better readability
• More natural tone
• Fewer spam filters (mental and actual)

If your AI assistant is still defaulting to this type of language, you either need better prompting or a better tool.

Vague flattery

“Impressed by your last post”

Impressed… how? Which post? Why mention it to me via email when you could’ve just left a comment?

Expressions like “Great post!” or “Loved your insight” add no value to outreach.

If your compliment can be copy-pasted under any LinkedIn post or in any email and still make sense, it’s not personalization. It’s noise on a mass scale.

Instead of generic praise, aim for a:

• Specific idea you agreed with
• Point you disagreed with (and why)
• Concrete takeaway you found useful

Specific praise builds credibility. Vague praise just signals “I skimmed your post feed and hit send.”

Personalization that isn’t relevant

I still get emails that mention my university, location, or previous employer. 

This is 2022/2023-level personalization. It’s easy to generate, and just as easy to ignore. And if your AI is still writing emails like this, you should probably switch.

Personalization isn’t your goal. Relevance is.

If the reference or fun fact doesn’t connect directly to your reason for reaching out, it’s not strengthening your message.

It’s distracting from it.

Relevance beats relatability every time. Personalization should help justify your reaching out. Not decorate it.

5 messaging tactics that yield consistent results

Here’s the good news: There are also patterns that work extremely well. The best part is that AI is uniquely positioned to execute them at a scale most human teams will struggle to match.

Across customers and campaigns, we’ve seen these five tactics consistently improve reply and conversion.

Stack 2–3 intent signals

Single-signal outreach feels lazy.

It’s better than nothing, but it’s not compelling.

That’s why you need to stack multiple intent signals, e.g. “saw you’re hiring” + “saw you just raised a Series A” + “saw you’re expanding into Europe”.

That’s 3 signals:

• Pain point: hiring needs
• Context: new funding
• Growth goal: Europe expansion

It tells the recipient: “I’m paying attention to what you’re dealing with”. And it sets you up to outline how you can help.

Write like you speak

If your email doesn’t sound like something you’d say out loud, it’s probably overwritten. 

Real humans use:

• Emphasis
• Exclamation points
• Humor
• Slang
• Contractions (you’re, they’re)
• Textspeak (btw, tbh)

If your message reads like a policy document, people will treat it like one: Something they’ll skim. Not something they’ll answer.

Just don’t overdo it. Clarity still matters. 

Prioritize honesty

The most refreshing messages are often the most direct, i.e. “I’m reaching out to sell you something but only because I think it solves X problem you mentioned.”

This works for two reasons:

1. It acknowledges the obvious. You’ve sent a sales email.
2. It grounds your pitch in something real and specific.

Nowadays, honesty is a pattern interrupt that makes people more likely to respond. 

Most outreach still pretends it’s just a “friendly check-in”. People see through that stuff.

You’ll be surprised how many replies you get that start with: “Finally, someone honest.”

Inject personality

Memes. Jokes. Casual intros. Slightly unhinged subject lines.

All of this can and does work, but only so long as it fits your brand and audience.

There is a caveat though. If your outreach is casual and fun, your website, deck, and product have to be premium and polished. Otherwise, you lose pricing power and trust if your emails are relaxed yet your website looks terrible.

Make your references ultra-specific

“Loved your podcast.”

“Great newsletter.”

“Insightful post.”

Too vague.

Better: 

Now you’re:

• Showing you listened
• Demonstrating your own thinking
• Starting a conversation rather than handing out praise

Don’t just reference their content. Engage with it.