Every new member hears my voice
before they touch their phone.
The moment someone joins my community, a small bot writes a short script, sends it through my cloned voice on ElevenLabs, and delivers a personal voice note. The welcome lands within seconds, sounds like me, and people thank me for the touch. The same trick answers support questions too, in my voice, without me ever picking up the phone.
People love it, and they think I sent it myself.
The welcome email is the cheapest part of a community to neglect, and the part where new members decide whether they belong. Most communities ship a templated email, a generic thank-you screen, and nothing else. There is a different way. Let a small bot write a short personal script the moment a new member joins, run it through your cloned voice, and deliver a real voice note inside their DMs. The note lands in seconds, it sounds like you on a calm day, and members keep telling you it is the warmest welcome they have ever received from an online community. Here is who it is for, what goes wrong without it, how it works, and what you get back.
Who it’s for
Course creators, coaches, paid community owners, membership site owners, and anyone with a Telegram, Discord, or Skool community where new people land every day. Especially useful if you already get the question “is this a real community or just a course” and you want the answer to be obvious within seconds. If the first hour after someone joins decides whether they stay, this is for you.
What goes wrong
Without the bot, new members land in a quiet room. The welcome email is templated, the dashboard is the same one a thousand other people saw, and nobody has said hello yet. Some members never come back after day one because nothing made them feel seen. The cost is silent. You never get to see the members who left in the first hour because the welcome felt cold. A cold welcome is a member you paid to acquire and never met.
How the machine works
The trigger fires the moment a new member joins. The bot reads their first name, the date they joined, and the reason they signed up. It writes a short script in a locked template. It sends the script to ElevenLabs, which renders the audio in my cloned voice. The bot drops the audio file straight into the member’s DM, with a short written intro on top. The member presses play and hears a personal welcome from me, by name, in my real voice, within seconds of joining.
What you get back
Around 2 to 3 hours of admin a week saved on member welcomes and short voice support replies. About 120 hours a year handed back to you for actual creative work. A welcome members keep mentioning in their replies, and a support thread that closes in one round instead of three. The welcome that used to be a templated email is now a personal voice note from me, and people love it.
The welcome email was the quietest part of my community.
For a long time my welcome email was a templated message. New member joins, autoresponder fires, generic text lands in their inbox, the next morning the member opens a quiet dashboard and tries to find their way around alone. I had no way to greet every new member personally, because at ten to twenty new members a day there was no human time to do it. The welcome was the part of the experience I cared about most and the part that felt most automated, in the worst sense of the word.
The cost was not the email itself, the cost was the silence that followed it. Members who joined on a Sunday afternoon never heard from a real person until they posted in a channel three days later. Some never posted at all because nobody had said hello. The hour right after someone joins is the single best chance to make them feel seen, and that hour was running on a template.
So I built a small bot. The moment a new member joins, the bot fires. It reads three things from the signup: their first name, the date they joined, and the short reason they wrote in the signup form for joining. It writes a short script inside a locked template that I tuned once. It sends the script to ElevenLabs, where my voice is already cloned from a clean five-minute sample. The audio comes back in about ten seconds. The bot drops the voice note straight into the member’s DMs with a small written line on top, the member presses play, and hears my real voice greeting them by name.
The cost is small. Around five euro cents per voice note in API and audio render fees. The total time I spend on each welcome is zero. At ten to twenty new members a day, plus around thirty support replies a day that also get the voice note treatment, that is roughly two to three hours a week of admin work returned, around 120 hours a year. Members reply to the welcome more often than they ever did with the templated email, and the same support questions that used to bounce back and forth three times now close in one round, because hearing the answer in my voice settles it faster than reading it.
Proof point: the mechanics behind this bot are the same ones I teach inside Automations Made Easy, where 1,000+ students have built their own version against their own community, their own voice, and their own signup form.
Three moves that turn every new signup into a personal voice note from me
What made this work was treating the welcome as a small live moment, not a templated email. The bot is not a writer, it is a stand-in for the warm hello I would have given each member in person if I had the hours. It hangs on three moves: clone the voice once from a clean sample, write a short personal script from the signup data, and deliver the rendered audio straight into the member’s DMs within seconds. Done right, it gives back around 120 hours of admin a year and lands a welcome that members keep mentioning in their replies.
Clone the voice once from a clean, calm sample
The clone is the whole foundation. Record three to five minutes of clean audio in a quiet room, in your normal speaking voice, the same energy you would use to greet a real new member in person. Trim the sample to the calmest takes only and upload it to ElevenLabs once. The clone you get back sounds like you on a calm day, not a stiff narrator. The sample is the single piece you have to get right. Spend an hour on it once, and every voice note for the next year inherits that warmth.
Write a short personal script from the signup data
The second move is the locked script template. The bot fills it in with three variables drawn from the signup: the first name, the date they joined, and a one-line reason they wrote for joining. The template stays short, under 80 words, and warm, the way you would greet a friend who just walked in. It never tries to sell, it never lists the perks of the community, it only says hello by name and points to the one place to start. Lock the template once, every welcome that follows feels personal because the variables make it personal.
Deliver the audio inside the DM within seconds
The last move is the delivery. The bot does not send an email, it sends a DM inside the platform the member is already on, with a short written line on top and the voice note attached. The member presses play and hears my voice within ten seconds of joining. That speed matters. A welcome that lands in the inbox tomorrow is part of the funnel. A welcome that lands as a voice note inside the same chat they signed up in is part of the moment. The first one feels marketed at, the second one feels welcomed.
Once those three moves are in place, the welcome stops being the quiet part of your community. It becomes the part new members talk about with the rest of the room.
Before the bot
- Templated welcome email, same words for every new member, every time
- The first hour after signup was silent, no real human had said hello yet
- Some members never came back after day one because nothing felt warm
- Support questions bounced back and forth three times before they closed
- I felt guilty about the welcome but had no human time to do it manually
After the bot
- A personal voice note from me lands inside the DM within seconds of joining
- Members hear their name in my real voice, with a real reason for being warm
- Support replies go out in my voice on top of the written answer, in one round
- About 2 to 3 hours a week of admin time returned, around 120 hours a year
- Members keep telling me it is the warmest welcome they have had online
Prompt 1: write the short personal welcome script from the signup data
The clone gives you the voice. This prompt gives you the words. Use it to write a short personal welcome script from the three variables your signup form already captures, so every new member hears something warm and specific the moment they join.
Personal welcome script writer
Act as a warm, slightly informal community host writing a short welcome script for a new member. The script will be read aloud by a voice clone of me and delivered as a voice note inside the member's DM within seconds of them joining, so it must sound spoken, not written, and it must stay under 80 words. The new member's first name: [paste] The date they joined (today): [paste] The short reason they wrote for joining in the signup form: [paste] The one place inside the community I want them to start (channel name, thread, or video): [paste] My usual tone when I greet a real new member in person (one sentence): [paste] Produce the script as a single short paragraph, no line breaks, no list, no bullet points, no salesy phrases. Start with their first name. Mention what they wrote in the signup form, in one short sentence, so the welcome feels heard. End by pointing to the one place to start, in a single calm sentence. Keep the script under 80 words. No "in this community", no "we are so excited", no "welcome aboard". Sound like a person greeting a friend who just walked in.
The output is the script that goes straight to ElevenLabs for rendering. Lock the template once, the bot fills the variables on every new signup.
Prompt 2: turn a written support answer into a voice note reply
The welcome side of the bot is half of the value. The support side is the other half. Use this prompt to turn any written support answer your team has already drafted into a short spoken script, so the same answer can also land as a voice note from me on top of the written reply.
Written-to-spoken support reply rewriter
Act as a warm, calm support voice for a paid community. A member has asked a support question. My support bot has already written a clean written answer in plain English. You need to turn that written answer into a short spoken script that I can render through my voice clone and send as a short voice note on top of the written reply, in case the member prefers to listen. The member's first name: [paste] The member's original question, in their own words: [paste] The written answer my support bot produced, in plain English: [paste] The one line at the end I always close with (my sign-off): [paste] Produce the spoken script as a single paragraph, no line breaks, no list, under 90 words. Start with the member's first name. Restate the question in one short sentence so the member knows you actually read it. Give the answer in spoken language, not written language, breaking long sentences down to the way I would actually say them. Close with my sign-off line. No "as mentioned above", no "per my last message", no formal phrasing. Sound like a person explaining something to a friend on the phone.
The script goes through the same voice clone as the welcome. The member receives the written answer and a short voice note in the same reply, and the support thread closes in one round.
The 3-minute overview of how this works
Before the build steps, watch this short overview. It is the exact video from the Automations Made Easy page, and it walks through the mechanics behind machines like this one. 1,000+ students have used these mechanics to save two hours a day, with zero coding.
Want to learn the mechanics behind voice note bots like this one?
I teach the same mechanics that make this cloned voice welcome run inside Automations Made Easy. 1,000+ students have used these mechanics to save two hours a day and turn quiet onboarding into the warmest part of their community. No coding required.
Week one: seventy welcomes sent, every one in my real voice.
Most people quit a system like this in the first week because the first welcome feels too small. There is no big spike, no retention chart bending up, no inbox flood. The first week of a cloned voice welcome is meant to feel quiet. The point is to confirm the bot fires on every new signup, the script reads warm, the audio renders in my voice without drift, and the voice note lands inside the DM within seconds, every single time.
Week one looks like this. Around 70 new members join across the week, plus around 200 support replies that get the voice note treatment on top of the written answer. Each welcome and each support reply costs about five euro cents in audio render fees, so the week costs around 13 euro in API fees and zero hours of my time. That is around 2 to 3 hours of admin work returned to me in the first week, the kind of small chunks that used to eat my evenings. Members reply to the welcome inside the first hour around three times more often than they did to the templated email, which is the leading signal that the change is landing.
The point of the first week is not the retention chart. The point is to prove the loop closes: someone joins, the bot fires, the script writes itself in the right tone, my voice renders the audio, and a personal voice note lands inside the DM within seconds, without me touching the phone.
From there the maths is simple. Around 2 to 3 hours of admin returned every week is about 120 hours a year handed back to me, roughly a quarter of work returned, on a task I was already neglecting because there was no human time for it. Members keep mentioning the welcome inside their replies, which compounds quietly into the kind of goodwill that brings new members in through word of mouth and keeps existing ones around longer than they otherwise would have stayed.
All of this runs while I record, write, sleep, or take a week off. The bot does not care. The trigger watches every signup, the script writer keeps the tone warm, ElevenLabs renders the audio, the DM lands inside the platform. The community grows on a schedule that matches the calendar, and the welcome finally feels like the warmest part of the experience instead of the quietest.
Prompt 3: tune the voice clone settings so the audio sounds like me on a calm day
The clone is only as good as the settings you render against. Use this prompt to lock the ElevenLabs settings for the welcome and support voice notes once, so every audio render sounds consistent, warm, and unmistakably me, not a stiff narrator or a sped-up version of myself.
Voice clone settings tuner
Act as a voice production engineer working with ElevenLabs voice cloning. I have already trained a clone from a clean 5-minute sample of my voice. I now need to lock the rendering settings so every voice note the bot produces sounds like me on a calm day, both for warm welcomes and for short support replies. A short description of how my voice naturally sounds (pace, warmth, accent if any): [paste] A 30-second sample script of a typical welcome I would say out loud: [paste] A 30-second sample script of a typical support reply I would say out loud: [paste] The platform the voice note will be played on (Telegram, Discord, Skool, or other): [paste] Produce four things. One: a target stability value (0 to 1) and the reasoning behind it for warm welcomes. Two: a target similarity boost value (0 to 1) and the reasoning behind it. Three: a style exaggeration value (0 to 1) tuned for a calm, slightly informal speaker, not a narrator. Four: a small rendering checklist (three to five short lines) for me to listen for in the first ten renders, so I can spot any drift early before locking the settings for the year.
The output is the settings profile the bot reuses on every render. Lock it once, every future voice note sounds like the same calm version of me.
Prompt 4: pick the right delivery line that sits on top of the voice note in the DM
The voice note carries the welcome, but a one-line written intro sits on top of it inside the DM and decides whether the member presses play. Use this prompt to pick the right delivery line for your platform, so the press-play rate stays high and the voice note actually gets heard.
DM delivery line picker
Act as a community manager writing the short written line that sits on top of a voice note inside a member's DM. The voice note carries the personal welcome from me, but most members will only press play if the written line in the DM signals it is short and personal, not a sales message. The platform the DM lands in (Telegram, Discord, Skool, or other): [paste] The member's first name: [paste] The length of the voice note in seconds: [paste] The reason they wrote for joining the community (so the line can echo it lightly): [paste] The tone I usually use when I write a short DM to a friend (one sentence): [paste] Produce three options for the one-line written intro that sits on top of the voice note in the DM, each one under 14 words, each one ending in a way that makes pressing play feel natural. Each option should mention the member's first name, signal that the voice note is short, and avoid every salesy or formal phrase. No "I hope this finds you well", no "looking forward", no "feel free", no emoji clusters. Sound like a real human dropping a 20-second voice note into a chat.
The bot uses the winning delivery line on every DM. Press-play rates stay high, the voice note actually gets heard, and the welcome lands the way it was meant to.
The exact build, step by step
Record a clean five-minute voice sample and clone it once on ElevenLabs
The clone is the whole foundation, so this step is the one that matters most. Sit in a quiet room, on a half-decent microphone, and record three to five minutes of your normal speaking voice, the same energy you would use to greet a real new member in person. No script reading, no narrator voice, just you talking the way you talk on a calm day. Trim the recording to the calmest takes only, upload it to ElevenLabs once, and let the clone train. You will not touch this step again for the year.
Wire the signup trigger that fires on every new member
The trigger is the source of truth for the whole bot. Point a small workflow (n8n, Make, or Zapier all work) at the signup event of your platform. Every time a new member joins, the workflow fires once with their first name, the date they joined, and the one-line reason they wrote in the signup form for joining. Filter the trigger so it only fires on confirmed signups, not on abandoned ones. The trigger does not write anything yet, it only carries the three variables forward to the next step.
Fill the locked welcome script template with the signup variables
This is where the personal touch is built. The bot opens the locked welcome template you tuned once (under 80 words, warm, single paragraph) and fills the three placeholders: the first name, the date, and the reason they wrote for joining. The template never gets replaced, it only gets filled. Every welcome that follows has the same warmth and rhythm because the bones are the same, but the surface feels personal because the variables are real. This is the move most people skip, and it is the one that decides whether the voice note feels generic or personal.
Render the audio through ElevenLabs in my cloned voice
The filled script gets posted to the ElevenLabs API with the clone ID, the locked stability, similarity boost, and style settings from the tuning prompt. The audio comes back in about ten seconds as an MP3 file. The clone reads the script in the same calm tone as the original sample, every time, with no drift across welcomes. The bot saves the file with a name that ties it to the member, so you can replay any single welcome later if you want to check the tone is still on rule.
Drop the voice note into the member’s DM with the short delivery line on top
The bot opens a DM inside the platform the member is already on (Telegram, Discord, Skool, depending on the community), writes the short delivery line you locked from prompt 4, and attaches the MP3 voice note right under it. The member presses play and hears me by name within ten seconds of joining. Email is never used, on purpose, because the DM lands inside the room they came to join, not the inbox they came to escape. The voice note is short enough that they listen to the whole thing, which is the only metric that matters in the first hour.
Reuse the same clone for support replies and log the press-play rate
Once the welcome side works, the same clone takes on the support side without any extra work. When the support bot writes a written answer, the same workflow turns it into a short spoken script (prompt 2), renders it through the same clone, and attaches the voice note under the written reply. The bot logs every voice note that was sent and whether the member pressed play within an hour, in a small Airtable row. You watch the press-play rate stay above 80 percent and you know the system is healthy. No login, no Studio, no checking.
Trigger fired
A new member joins the community, the workflow fires once with their first name, the date they joined, and the reason they wrote for joining.
Script written
The locked welcome template gets filled with the three signup variables. The script stays warm, stays under 80 words, and reads like a friend greeting a friend.
Voice rendered
ElevenLabs renders the audio through my cloned voice in about ten seconds, with the locked stability and similarity settings. The clone reads with the same calm tone every time.
Delivered and logged
The DM lands inside the platform with the short written line on top and the voice note right under it. The press-play rate gets logged in a small Airtable row so the loop stays watched.
Build this cloned voice welcome inside the same playbook 1,000+ students use
Automations Made Easy teaches the mechanics behind bots like this one. Step by step, no code, plain English. Turn quiet onboarding into the warmest part of your community and let every new member hear from you personally within seconds.
The six months after I switched it on
Here is the shape of the first six months after I turned the cloned voice welcome on for the community. The line is intentionally modest in the early months and steady from there, because that is how a welcome actually behaves: every new member adds a small layer of goodwill, and the layers stack quietly.
Admin hours returned per month from welcomes and support voice notes
Caption: Monthly admin hours returned to me by the bot, across welcomes and short support voice notes. Three things matter on this chart. The line settles around 10 hours a month, about 120 hours a year, roughly a quarter of working time returned. The hours come from a task I was already neglecting because there was no human time for it. And every single one of those hours happens without me opening a single chat tab.
What other students built
I teach the simple skills behind machines like this in Automations Made Easy. Students who built their own version sent back what changed in their first month.
“My welcome email used to be three lines and a calendar link. The cloned voice welcome turned signup into the moment people remember. New members reply within the first hour now, which they never did before.”
“I run a paid coaching group. The voice note welcome made new members feel like they joined a real room, not a course. I get fewer refund requests in the first week and more questions in the chat, which is exactly the shape I wanted.”
“Wired it for support replies too. The same question used to bounce back and forth three times. Now the written answer goes out with a short voice note from me on top, and the thread closes in one round.”
“Saved me about two and a half hours a week across welcomes and support. Members keep mentioning the voice note in their replies, which is something nobody ever did about my old welcome email.”
What’s inside Automations Made Easy
AME isn’t a library of pre-built automations. Every business is slightly different. What’s reusable across all of them is the underlying mechanics: how to set up little machines that listen, write, render, and follow up while you sleep, and how to wire the pieces together without writing code.
The program walks you through six modules: The Right Tools (the cost-effective, no-code stack I actually use), Task Selection Mastery (which automations are worth building first), Design Secrets (mapping an automation before you build it), Zero to Hero (complete beginner to confident automator), Real-World Application (we build a full automation together, end to end), and Monetization Mastery (turn the skill into a side-business).
It also includes done-for-you templates you import in two clicks, over-the-shoulder training videos, and the same playbook 1,000+ students have used to save two hours a day. No coding required. If you can copy and paste, you can build this.
The voice note bot: common questions
Pulled from what readers and Automations Made Easy students ask most.
How do I clone my voice without sounding robotic?
You record around three to five minutes of clean audio in a quiet room, in your normal speaking voice, then upload that sample to a voice cloning tool like ElevenLabs. The clone you get back sounds like you on a calm day, not a stiff narrator. The trick is the source sample. If you record on a phone in a noisy room, the clone inherits the noise. If you record in a quiet room with the same energy you would use to greet a new member in person, the clone inherits that warmth. I trimmed my sample down to the calmest takes only, and the clone has never been called out as AI by a single member.
Won’t people feel cheated when they realize it is AI?
Most people never realize, and the ones who do tell me they still love the touch. The reason is simple. The script is real, the warmth is real, the attention to their name and their reason for joining is real. The only piece that is automated is the audio render. People are not upset that the audio was rendered, they are upset when nobody bothers to greet them at all. A personal voice note within seconds of joining beats a generic welcome email by a wide margin, even when the listener knows the voice was cloned. The personal touch is in the message, the cloning is just delivery.
Can I use this for support replies too, or just welcomes?
Both, and the support side is where it earns its keep. My support bot reads the question, writes the answer in plain English, then hands the answer to the same voice clone. The member gets the written answer in chat and a short voice note from me on top, in case they prefer to listen. Members reply faster, the same question rarely comes back twice, and the support thread closes in one round instead of three. Around thirty support replies a day get the voice note treatment in my community, on top of every welcome, with zero extra time from me.
What does each voice note cost to generate?
Around five euro cents per voice note on a ElevenLabs plan sized for a community my size. A welcome note runs about twenty to thirty seconds, a support reply runs about thirty to forty-five seconds, both stay well under the cost of one minute of voice rendering. At ten to twenty new members a day plus around thirty support replies, that is roughly two to three euro a day, around seventy euro a month, for a personal touch that members keep mentioning in their replies. The cost is small enough that I never think about it, the goodwill is the part I notice.
How long before members notice the difference in retention?
You see the difference inside the first week, but it shows up in replies before it shows up in retention numbers. New members start the first hour by replying to the voice note with a short thank you, which they almost never do for a welcome email. That early reply is the leading signal. The retention number itself moves slowly over the first one to two months as the early-reply members stay engaged longer than the silent ones used to. The lift is modest but consistent, and the goodwill compounds because members tell other members about the welcome before they ever see a sales message from me.
More Money Makers like this one
Built from the same handful of mechanics. Each one captures, converts, or earns money that would have walked out the door.
Run this cloned voice welcome yourself, or learn the mechanics inside Automations Made Easy.
If you want to learn the mechanics behind voice note automations like this and build your own at home, Automations Made Easy is the playbook. Step by step, no code, plain English. If you want to talk through which signup trigger, voice clone, and script template would fit your specific community first, I take a small number of consulting clients each month.
1,000+ students · Save 2 hours a day · No coding required
