It feels perfect: Click “Draft with AI.” Give it a skim. Hit send.
You’re not copying and pasting. You’re not using an extension or automated sequence. This is built into LinkedIn. It must be smart, right?
Except, it’s not. And if you’re using it as-is, you’re slowly eroding trust, message by message.
LinkedIn has more real-time professional context than almost any platform: Connections. Roles. Posts. Career shifts. Shared activity.
If any platform could make intelligent, personalized outreach easier, it’s this one.
And yet, the AI messaging feature is flawed. Yes, it is a "beta" and probably will always be a "beta." But it is missing some really important capabilities. It has no awareness or nuance, or any feel to it.
If this was a third party tool it would be seen as a spam engine.
I grabbed a 2nd degree connection. Kimberly, who I do not know, but would be a great example of this feature because of the shared connections, her built out LinkedIn profile, and LinkedIn activity.
1. It Ignores the Obvious
Nine mutual connections? No mention. Shared Groups? No attempt to weave them in. Common abbreviation of the company name? Nope, we went with the formal "The O'Connor Group" instead of TOC.
The AI skipped all of them. Any of those could possible make this stand out as a note Kimberly should read.
2. It Tries to Sound “Professional”
“I was curious about how {Company Name} is currently leveraging”
That’s not professional. That’s stiff, generic, and zero-context. It’s the kind of tone people use when they’re using templates or aren't confident.
Sales leaders coach against this, as I talked about in my last article and here comes AI implementing it at scale.
3. It’s All About Me
The message starts with me. My curiosity about how they are solving for LinkedIn. Not something I noticed. Not sharing have you tried or been experimenting with?
Zero insight or value at from me. Just a vague ask and then we can solve that.
4. It Misses The Little Things
“Hi Kimberly J.” is the fastest way to sound like a bot. Now this won't come up for anyone but this has always been a tell tale sign.
It’s 2025. If your “AI” can’t parse a LinkedIn headline or name string, don’t call it AI.
Just a wild guess, Kimberly J goes by Kim. A quick glance at her LinkedIn URL will tell you that.
Let’s address the obvious: Yes, you need systems to scale outreach. Yes, you or your team are busy. No, no one wants to handcraft messages all day long.
But that’s precisely why quality matters more, not less.
If your first impression looks like automation or what I call "pseudo-personalized", it gets filtered, mentally or literally. The only thing worse than being ignored is losing your credibility.
The best messages feel like someone did their homework. They reflect intent. Curiosity. Context. Timing. They are what I call professionally conversational.
AI should support that posture, not shortcut it.
Right now, LinkedIn’s messaging feature seems like it is working against its tagline of "Unlocking high-quality conversations with the people that matter, at scale." It’s helping you move faster in the wrong direction.
It’s not a clever line. It’s not a 14-step framework.
It’s this:
That’s what good messaging does every day. This can be done with the help of AI.
At Intero Advisory, we’ve spent years helping companies improve their LinkedIn outreach by focusing on what really works. We teach people how to write better messages and pay attention to the small things that make a difference. It’s about knowing where to look, making smart adjustments, and doing it consistently.
It seems small but noticing abbreviations, nicknames, relevant activity mentioning, knowing when to send based on when and how people access LinkedIn. Those are the things that improve your response rates, spark conversations, and, ultimately lead to conversion - be that sales or hiring.
You have the data, the behavior signals, and even the incentive: better messages → better connections → more platform activity.
But until these features are smarter, I'd concerned about letting LinkedIn’s AI messenger do it for you.
The potential small time saved isn’t worth the negative impact on your credibility.
In summary, IMHO - overall LinkedIn's AI is not ready for prime time.