Ivona MikulčićNovember 17, 20266 min read
LinkedIn voice messages for prospecting, when and how to send them
Yes, you can send voice messages on LinkedIn, and most reps never use the feature. That gap is exactly why it works: a 20 to 30 second audio note cuts through a feed of copy-pasted text pitches and sounds like an actual person instead of a template.
This article covers when to send linkedin voice messages, what to say in them, and how to send them at scale without recording each one by hand. You’ll get a script structure, real examples, timing rules tied to LinkedIn connection data, and the mistakes that turn a good idea into another ignored message. By the end, you’ll know exactly where a voice note fits in your sequence and how to automate it without losing the personal feel that makes it work.
Why voice messages stand out on LinkedIn
Most LinkedIn inboxes look the same: connection request, text pitch, follow-up text, repeat. A voice note breaks that pattern because audio is still uncommon enough to trigger curiosity, and the format stands out in a feed of identical text messages.
When a prospect hears your voice, they get information that text cannot communicate: energy, authenticity, and confidence. That makes it harder to ignore than another paragraph of copy-pasted outreach. And because so few reps use voice, the channel carries a novelty advantage that plain text lost years ago.
Voice notes also force brevity. LinkedIn limits voice messages to 60 seconds and restricts them to 1st-degree connections, so there is no room for a wall of text. That constraint works in your favor: a short, relevant audio message respects the listener’s time and earns attention instead of demanding it.
When to send one
Send a LinkedIn voice message after your connection request is accepted, as a second or third LinkedIn touch. Not before.
Sending an audio message to someone you haven’t connected with feels intrusive. It’s an intimate format that works best with at least minimal prior permission. Once the invite is accepted, you have a 1st-degree connection and a natural opening to follow up.
The timing matters because of what happens after the accept. According to lemlist internal data, the median LinkedIn invite acceptance rate across campaigns is 27.3%, with the top 10% reaching 53.9% or higher. That accepted invite is a warm signal. Use it.
Before you record, pull context from the prospect’s LinkedIn profile: a recent post, a job change, a company announcement. That context is what separates a voice note that gets a reply from one that gets skipped. If you want to build a full LinkedIn outreach strategy around these signals, structure your sequence so the voice note lands after an initial text message, not in place of it.
According to lemlist internal data, three LinkedIn touches in a row reply at 10.4% in France and 14.5% in the rest of the world (per lead contacted), compared to 1.6% and 1.9% for a single email. Adding a voice note as one of those touches gives the sequence a format shift that text alone cannot deliver.
What to say in 30 seconds
Use a four-part structure: open with their name and your name plus context (5 seconds), make a specific observation about them or their company (15 seconds), state one relevant reason you’re reaching out (10 seconds), and close with a single low-friction ask (5 seconds).
Here are three examples you can adapt:
“Hey Sarah, it’s Alex from lemlist. I saw your team opened four AE roles in EMEA, so it looks like outbound is a priority this quarter. We help sales teams run multichannel sequences without stitching together five tools. Would a 15-minute call next week make sense?”
“Hi Marcus, Tom here. I noticed your post about reply rates dropping on cold email. We’ve been seeing the same thing, and the teams getting around it are adding LinkedIn and voice touches to their sequences. Happy to share what’s working if you have 10 minutes.”
“Hey Priya, it’s Jordan. Congrats on the Series B announcement. A lot of teams at your stage start scaling outbound and hit deliverability issues fast. That’s what we solve. Would it be worth a quick chat?”
Each example stays under 30 seconds, leads with something specific to the prospect, and ends with a single ask. No product dump, no feature list.
How to send voice notes at scale
Recording one voice note at a time works when you have five prospects. It breaks at fifty.
In lemlist, you write one voice script with AI variables (like {{Icebreaker}}), and lemlist turns it into as many personalized LinkedIn voice notes as leads you have, cloned in your own voice. The integration with ElevenLabs handles the voice cloning. You can clone your voice or choose one of the proposed natural-sounding AI voices.
Here is how it works inside a lemlist sequence:
Add a LinkedIn voice message step to your multichannel campaign
Write your script using variables for personalization (prospect name, company, icebreaker)
Clone your voice or select a pre-built voice
Preview the output for a few leads before launching
lemlist generates and sends a unique voice note for each lead in the sequence
The result is a voice message that sounds like you recorded it one by one, but runs at campaign scale. For a deeper walkthrough of the setup, see the guide on AI voice notes in lemlist.
This fits naturally into a broader LinkedIn message sequence where voice is one touch among several, not the only channel.
Mistakes to avoid
Going past 30 seconds. Using the full 60-second limit tends to make the message sound like a scripted pitch rather than a genuine human observation, which kills reply rates. Aim for 20 to 30 seconds. If you cannot say it in that window, the message needs editing, not more time.
Leading with the pitch. The first five seconds decide whether someone keeps listening. If those seconds are “We’re a platform that helps companies…” the prospect is already gone. Start with their name and something specific to them.
Skipping context. A voice note without a reference to the prospect’s situation sounds like a robocall. Mention a post, a hire, a company milestone. That is the difference between personal and generic.
Making every touch a voice note. Voice works because it is rare. If every step in your sequence is an audio message, you lose the contrast that makes the format effective. Use it once or twice in a sequence, paired with text and email. A good LinkedIn follow-up message in text after a voice note reinforces the conversation without overusing the format.
Recording in a noisy environment. Background noise signals low effort. Find a quiet spot, or use AI voice cloning in lemlist to generate a clean recording every time.
Turn LinkedIn leads into pipeline
Find prospects and automate personalized LinkedIn outreach.