How to make the perfect voice clone with Eleven Labs

If you’ve spent any time on TikTok, Instagram Reels, or YouTube Shorts lately, you’ve probably noticed something: AI voiceovers are everywhere. They’re ‘human enoug’” that most people don’t stop to question them.

Tools like ElevenLabs offer big libraries of ready-made voices in different accents and languages, and creators are using them to narrate everything from fact videos to mini-docs.

In this newsletter I want to run through something that isn’t quite so normalised: voice cloning - making a replica of your own voice and using it in production. I’ve been experimenting with this for 2 years now, but only recently had enough demand to justify the time and cost putting in to making a new one. But I can say with certainty that those costs were justified.

Let’s separate two things that often get blurred together:

Both are useful. But cloning has a few specific use cases that make it worth talking about on its own.

Voice cloning sits in a weird cultural and ethical zone. Some people find it creepy - the uncanny valley effect where something feels almost human but not quite. Others see it as a natural extension of digital production.

There are also hard legal and professional realities: a person’s voice is part of their identity and often their intellectual property. In TV, podcasting, and other talent-based industries, contracts frequently forbid cloning a presenter’s voice without explicit permission, and usually compensation. So the rule is simple: never clone someone else without them fully understanding and approving it.

Where cloning gets genuinely effective is when a creator chooses to clone their own voice as a tool to speed up their workflow.

Anyone who edits knows the moment: you’ve finished the cut, then realize you need a missing line, a correction, or a cleaner take. Normally that means re-recording, coordinating schedules, setting up mics again, and hoping the tone matches.

With a voice clone, you just type the new line into ElevenLabs and drop it into your timeline. This can make an annoying fix with many moving parts relatively painless.

This is the one that changed my process.

My videos lean heavily on images, footage, and documentary-style pacing. If I use a clone, I can take my script and generate a full narration track immediately, then start building the edit around it - without filming or recording anything first.

That means:

It massively reduces the time spent recording and also cuts down narration-editing in post. In my own workflow, editing time in Premiere for presenting dropped from about 2 hours to 15 minutes. Recording time fell from ~50 minutes to ~15 minutes. That’s roughly 2.5 hours saved per video - which compounds into weeks over a year.

AI has changed how consumers shop by speeding up research. But one thing hasn’t changed: shoppers still trust people more than AI.

Levanta’s new Affiliate 3.0 Consumer Report reveals a major shift in how shoppers blend AI tools with human influence. Consumers use AI to explore options, but when it comes time to buy, they still turn to creators, communities, and real experiences to validate their decisions.

The data shows:

The most effective brands are combining AI discovery with authentic human influence to drive measurable conversions.

Affiliate marketing isn’t being replaced by AI, it’s being amplified by it.

Download the full report to see what this means for your brand.

ElevenLabs recommends at least one hour of clean audio, ideally closer to three hours, for professional cloning. You can get usable results with less, but quality scales with amount and consistency of input.

Key prep rules:

If you don’t already have an audio archive, the fallback is simple: record 1–2 hours of clean reading with a decent mic in a quiet room. It can be a novel, articles, anything — you just want steady, natural speech.

Once it’s ready, test it with casual lines, then with a real script segment. I treat clone output as a guide track for editing — like a painter’s sketch before the final layer. Then I re-record only what genuinely needs human presence.

AI voice cloning is way beyond a gimmick now. It’s a tool for:

If you produce consistently, the subscription pays for itself quickly in time saved. And if you want to take the next step, pairing voice cloning with a visual avatar can unlock even more automated production workflows.