I’m about to send my first batch of connection requests, and I’m in my head about how to make them not sound like a bot. I know that because I’m using a proxy and potentially managing multiple accounts, LinkedIn’s probably already a little skeptical of my account by default. So my first message to someone—the connection request itself—needs to feel genuinely human.
The problem: I also know that my message needs to be personalized enough to not get ignored, but short enough to not scream “I spent 2 hours researching you.” And I’m supposed to be funneling this through LiSeller’s AI, which is supposed to help, but I’m worried the AI is going to add some canned phrasing that just sounds corporate.
How do people actually write first connection requests in this scenario? Like, are you keeping them super short (like, 3–4 sentences)? Are you leading with a specific detail about their profile? Are you using the AI to draft it or overriding it with your own version? I’m trying to figure out how authentic-sounding first outreach actually works when you’re running accounts through proxies.
What’s the actual formula that works?
Okay, the copywriting secret here is that your first connection request shouldn’t feel personalized—it should feel relevant.
There’s a difference. “Personalized” can feel creepy (“I noticed you graduated in 2014”). “Relevant” feels natural (“I work with engineering teams building X, saw you’re at Y, might be worth a convo”).
My formula for connection requests: (1) Light context about why you’re connecting (one sentence max), (2) One specific detail from their profile or work (NOT their resume history, but something they actually did—shipped something, wrote about something, etc.), (3) A micro-value prop (what you do, really briefly), (4) No ask. Just “let me know if you’d be open to a quick chat.”
Total: 2–3 sentences. That’s it.
In LiSeller, feed the AI your best version of that arc, let it generate variations, and pick the one that sounds most like how you actually talk. Edit it if needed. Hit send.
Technical approach: don’t let the AI write your connection request. That’s template territory.
Here’s what I do:
- Create a master connection request template (2 sentences, max 100 characters after personalization).
- Define your personalization variables:
[First Name], [Company], [Recent Activity] (if LiSeller can pull that).
- Upload your list with those data columns pre-filled.
- Use LiSeller’s variable substitution to generate personalized versions.
- Preview them before sending. Don’t auto-send the first batch.
This way, you get consistency + personalization without AI-generated corporate-speak. Takes maybe 20 minutes to set up, and then it’s automated but still authentic.
As a recruiter, I keep it dead simple: “Hey [Name], I’m recruiting for [Role] at [Company]. Saw your background—think you’d be a fit. Want to chat?”
That’s like 15 words. Takes me 10 seconds to customize. And it works because (1) it’s immediately clear why I’m reaching out, (2) it’s not trying to be clever, and (3) it respects their time.
LiSeller’s AI is great for follow-up messages, but for initial connection requests? Keep it bare bones. Short + direct = more likely to get acceptance.
Don’t overthink it. The connection accept is just the first gate. Your real story comes in the follow-up.
From a safety standpoint: the less you customize connection requests, the safer you are.
Why? Because per-account customization on first contact can flag LinkedIn’s algorithm. If you’re sending 20 handcrafted, unique connection requests, LinkedIn might see that as bot-like behavior (“Why would a human personalize every single first message?”).
Here’s the safer approach: use a standard connection request template (like 80 characters) with minimal personalization. Just first name + company. Something like: “Hey [Name], let’s connect. I work with [Company Size] companies doing . Curious about [Y] in your space.”
Then, after they accept, that’s when you personalize heavily. Your first follow-up message is where the AI magic happens.
Initial request = conservative and safe. Follow-up = where you show personality.
I spent way too long crafting perfect connection requests and then realized I was wasting time. My best performer? Super simple: “Hey [Name], I work with [Industry] companies on [Problem]. Let’s connect.”
I probably get like 30-40% accept rate with that. With more fancy personalization? Maybe 35-40%. Difference is negligible, but the simple version takes me 1 second per person versus 30 seconds.
Use that time you save on writing killer follow-ups instead. That’s where the conversion happens. Connection request is just the door opener.
In LiSeller, you have two paths for connection requests:
Path 1: Manual Template – You write a template, we substitute variables. Simple, safe, human-sounding.
Path 2: AI-Generated – You give the AI a tone (friendly, professional, casual), it generates something on-the-fly.
For day-one first impressions, I recommend Path 1. Just template + variables. It’s faster, you maintain complete control, and it feels natural.
Save the AI for your follow-up sequence—that’s where personalization really matters and where the AI adds value.
If you go AI route, always preview the output and edit before sending. Don’t let auto-send run on AI-generated connection requests without human review.
From a conversion psychology angle: your connection request is setting expectations.
If your connection request is highly personalized and clever, the prospect expects your follow-up to be equally thoughtful. If you then send a generic follow-up, they lose trust.
If your connection request is simple and authentic, they expect a straightforward follow-up, and you can personalize that more heavily without it feeling incongruent.
So strategically: keep the connection request boring and safe. Template + first name + company. Then, 2–3 days after acceptance, send your real personalized message—your hook, your value prop, your story.
That’s when you use the AI. That’s when you show up.