AI Outreach That Doesn't Sound Like AI: How to Write LinkedIn Messages People Actually Reply To

Most AI-generated LinkedIn messages fail for the same reason: they sound exactly like AI-generated LinkedIn messages. The recipient reads the first line, feels the template-ness of it, and moves on.

You've seen them. "I hope this message finds you well." "I was looking at your profile and was impressed by your experience." "I believe there may be a synergistic opportunity." Nobody talks like this. The AI does it because it's been trained on a lot of bad outreach copy.

The good news: AI outreach can work. The prerequisite is that it doesn't sound like AI outreach.

Why most AI outreach fails

The fundamental problem with AI-generated messages is that they optimize for sounding professional rather than sounding human. Professional sounds safe. Human sounds like a real person sent it.

There are specific patterns that immediately flag a message as AI-generated. Once a reader spots one of these, trust is gone:

• Opening with "I hope this finds you well" or any variant of it
• Starting sentences with "I" (LinkedIn actually flags this)
• Em dashes used as a stylistic choice ("I noticed your company — and was intrigued")
• Exclamation marks used to signal enthusiasm
• Multiple bullet points in a message that should be a quick DM
• The word "synergy" or "leverage" or "impactful"
• Three-paragraph messages when one paragraph would do

These patterns have become reliable tells. People who receive a lot of LinkedIn messages know them instantly.

The anti-slop approach

The best AI-assisted messages are ones where the AI generates a raw draft, then you strip everything that sounds fake and add one specific detail that only a human with actual context could write.

The specific detail is everything. "I noticed you're at Acme Corp" is generic and forgettable. "I noticed you just moved to the enterprise side at Acme — curious how that's going after the startup years" is something only a real person would say, because only a real person would have read enough of their profile to know that detail.

The AI can draft the structure. The human adds the soul.

Before and after: what good AI outreach looks like

Before (typical AI output):

After (AI draft, edited by human):

Same underlying purpose. Completely different result. The second one gets replies. The first gets archived.

The purpose-selection approach

One of the reasons AI outreach fails is that it tries to be everything at once — vaguely sales-y, vaguely collaborative, vaguely friendly. Clear purpose makes messages far more effective.

Before drafting any outreach, define exactly what you want from this specific message. Not from the relationship in general — from this specific message, today.

• Fundraising: you want an intro to their fund or to know if they're writing checks
• Sales: you want to know if they're the right person to talk to about your product
• Hiring: you want to know if they're open to hearing about a specific role
• General: you want to re-establish contact with no agenda attached

The message changes completely depending on which of these it is. A message with a clear purpose is always better than one that hedges. People respond to clarity because it makes it easy for them to say yes or no — and even a clear "no" is more useful than silence.

How wait who? handles this

When you draft a message in wait who?, you pick a purpose first — fundraising, sales, or general reconnection. The AI generates a draft calibrated for that specific intent, not a generic "let's connect" placeholder.

The draft avoids the hallmarks of AI slop: no "hope this finds you well," no em dashes, no exclamation marks, no three-bullet preambles. It's short, specific, and written as if a real person sent it.

You still edit it. You still add the specific detail that proves you actually read their profile. But the blank page problem is gone — and that's usually the thing that stops people from sending the message at all.

AI outreach that sounds like AI gets ignored. AI outreach that sounds like you — specific, purposeful, and human — gets replies.