n8n

Personalized Video Outreach at Scale with n8n, HeyGen & ElevenLabs

Rodrigo Ferreira

— 5 min read

Personalized video outreach is the fastest way to stand out in a crowded inbox, but recording a unique clip for every lead simply does not scale. This guide shows you exactly how to automate hyper-personalized sales and marketing videos using n8n, HeyGen AI avatars, and ElevenLabs voice synthesis, so each prospect receives a video that says their name, references their company, and speaks to their specific pain point, all generated without you ever touching a camera.

By the end of this article you will understand the full workflow architecture, the real use cases that justify building it, and a step-by-step overview to get your own pipeline running.

The Problem: Manual Video Recording Does Not Scale

Video outreach works. Prospects are far more likely to reply to a personalized video than to a plain text email, because a video feels human, deliberate, and hard to fake. The trouble is the math. If recording, naming, and uploading a single personalized clip takes ten minutes, then 100 prospects costs you almost seventeen hours of focused work. Do that weekly and video becomes a full-time job rather than a growth lever.

Sales reps quickly hit a wall:

  • Time — each clip needs a fresh take, a re-record when you stumble, and manual file management.
  • Consistency — energy and message drift across dozens of takes in a day.
  • Lighting and setup — every recording depends on you being camera-ready.
  • No reusability — a clip for one prospect is useless for the next.

The result is that most teams either send generic videos that defeat the point of personalization, or they record a handful and give up. Automation solves this by separating the asset (your avatar and cloned voice) from the content (the per-lead script), letting you mass-produce truly individualized videos.

The Workflow Architecture

The system is an n8n workflow that orchestrates three external services. n8n is the conductor; HeyGen and ElevenLabs are the production studio. Here is the end-to-end flow:

  1. Trigger — a new row in a Google Sheet, a webhook from your CRM, a new lead in HubSpot or Pipedrive, or a scheduled batch.
  2. Fetch lead data — pull first name, company, role, recent activity, and any custom fields from your CRM or enrichment tool.
  3. Generate the personalized script — pass the lead data to an AI model (via the n8n AI nodes) with a prompt template that writes a short, natural 20–40 second script referencing the prospect by name and their situation.
  4. Generate the voice — send that script to ElevenLabs to produce audio in your cloned or chosen voice, returning an audio file or URL.
  5. Generate the avatar video — send the script (or the ElevenLabs audio) to HeyGen to render a talking avatar video. HeyGen returns a job ID, so the workflow polls until the render is complete and a video URL is returned.
  6. Deliver — send the video link via email (with an animated GIF thumbnail), through your LinkedIn outreach tool, or by writing it back to your sequencing platform.
  7. Log — write the video URL, render status, and timestamp back to your sheet or CRM so nothing is sent twice and every result is tracked.

Skip the build. Get the ready-to-import n8n template → Get it on Gumroad

Why ElevenLabs and HeyGen together

HeyGen can generate audio on its own, but pairing it with ElevenLabs gives you a far more natural, brand-consistent voice, especially if you clone your own. The pattern is simple: ElevenLabs produces the audio, then you feed that audio into HeyGen's avatar so the lip-sync matches your real voice. This combination is what makes the output feel like you rather than a robot, which is the entire reason personalized video converts in the first place.

Benefits of Automating Video Outreach

  • Volume without burnout — produce hundreds of unique videos in the time it took to record five manually.
  • True personalization — every video references the prospect by name and context, not a generic script.
  • Consistency — the same energy, message, and quality on video number 1 and video number 500.
  • Always-on — the workflow runs overnight, so videos are ready when your reps log in.
  • Measurable — because everything is logged, you can A/B test scripts, voices, and avatars against reply rates.
  • Reusable asset — set up your avatar and voice once, then reuse them across every campaign forever.

Real Use Cases

1. Cold outbound sales

A SDR team uploads a list of 300 enriched leads. The workflow generates a 25-second video for each, opening with "Hi {{firstName}}, I noticed {{company}} just expanded into {{market}}..." Reply rates on personalized video routinely outperform plain text by a wide margin, and now the team sends them at scale.

2. Trial onboarding and activation

When a user signs up for a free trial, the trigger fires and they receive a welcome video addressing them by name, pointing to the exact feature relevant to their stated use case. This lifts activation without adding headcount.

3. Event and webinar follow-up

After a webinar, each attendee gets a personalized thank-you video referencing a question they asked or a topic they engaged with, keeping the relationship warm while the moment is fresh.

4. Re-engagement of dormant leads

Pull leads that went quiet 90 days ago and send a fresh, personalized video to revive the conversation, a campaign that would be unthinkable to record by hand.

Step-by-Step Setup Overview

  1. Create your accounts and API keys — sign up for HeyGen and ElevenLabs and generate API keys. In HeyGen, create or select your avatar. In ElevenLabs, choose or clone a voice and note the voice ID.
  2. Set up credentials in n8n — add your HeyGen and ElevenLabs API keys as HTTP header credentials, plus your CRM or Google Sheets credentials for the data source.
  3. Build the trigger — start with a Schedule or Webhook node, or a CRM trigger. For testing, a Google Sheets trigger on new rows is the simplest.
  4. Add a data node — fetch and map the lead fields you will reference: first name, company, role, and one specific detail that proves the video is genuinely personalized.
  5. Add the AI script node — use an AI node with a tight prompt: "Write a friendly 30-second outreach script for {{firstName}} at {{company}}, referencing {{detail}}. Keep it under 80 words, conversational, one clear call to action."
  6. Call ElevenLabs — use an HTTP Request node to POST the script to the ElevenLabs text-to-speech endpoint with your voice ID, then store the returned audio.
  7. Call HeyGen — POST to HeyGen's video generation endpoint with your avatar ID and the audio (or script). Capture the returned job ID.
  8. Poll for completion — add a Wait node and a loop that checks HeyGen's status endpoint until the render is finished and a video URL is returned. This step matters because rendering is asynchronous.
  9. Deliver and log — send the video via your email or outreach node, then write the URL and status back to your source so the lead is marked as processed.
  10. Test, then scale — run a single lead end to end, verify the video quality and personalization, then enable the batch trigger and monitor your logs.

Tips for higher reply rates

  • Keep videos under 40 seconds. Short and specific beats long and polished.
  • Lead with the personalized detail in the first five seconds, before attention drops.
  • Use an animated thumbnail in emails so the message clearly reads as a video.
  • Always log results and iterate on the script prompt based on what gets replies.

Conclusion

Personalized video has always been one of the highest-converting outreach formats, and the only thing holding teams back was the manual labor. By combining n8n's orchestration with HeyGen's avatars and ElevenLabs' voice, you turn a seventeen-hour task into a workflow that runs while you sleep, producing genuinely individualized videos at any volume your pipeline demands. Set up the avatar and voice once, point it at your lead data, and let automation handle the rest.

Ready to automate? Get this template on Gumroad →

Read more