The reason most AI scripts sound like AI is that people start with the AI. They ask it to write, and it writes the average. This method does the opposite. Your brain goes in first, in the messiest possible form, and the AI only ever organises and sharpens what you already said. The result keeps your information gain intact, which is the whole point.
This is the core scripting workflow of the Claude-for-YouTube system, and it takes a rough idea to a filming-ready script in under two hours.
The sequence
Voice note one, the brain dump. Record yourself talking through everything you want to cover. No structure, no plan, just talk. The messier the better. Every story, every point, every opinion. This is where your unique perspective enters the script.
Transcribe it. Any transcription tool will do. Your phone can probably do it on its own.
Organise it. Paste the transcript into Claude and have it find the structure hidden inside the mess. It pulls out your core argument and key points, organises them into a logical flow, flags your strongest moments so they are protected, and marks the thin spots where you need more depth. Crucially, it adds nothing that was not already in your voice note. It only organises what is there.
Voice note two, the performance. Now record a second voice note, this time presenting from that logical flow as if you were talking to camera, paying extra attention to the spots it flagged as thin. This becomes roughly eighty percent of your script.
Transcribe and refine. Transcribe the second note and hand it back. Claude tightens the rambling, removes repetition, strengthens the weak transitions, keeps your voice, and scores the result honestly for information gain so you know if it needs more of your own experience before you film.
Why the second voice note matters
The first note is raw material. The second is you actually presenting, which is closer to how the video will sound. Recording it from an organised flow means you cover everything without the rambling, and the AI has a strong, voice-rich draft to refine rather than a blank page to invent from.
From here the hook, the de-fluffing, and the calls to action finish the job. But the heavy lifting is done, and it sounds like you, because it started as you.
Frequently asked questions
How do I write a YouTube script with AI without it sounding generic?
Start with your own voice notes instead of the AI. Record a rough brain dump, have the AI organise it into a logical flow, record a second voice note presenting from that flow, then have the AI refine the transcript. Your experience goes in first, so the script keeps your voice.
Why record two voice notes instead of one?
The first captures your raw material with no structure. The second is you actually presenting from an organised flow, which is closer to the final video and gives the AI a strong, voice-rich draft to refine rather than a blank page.
How long does this scripting workflow take?
Under two hours from rough idea to filming-ready script, most of which is your two voice notes. The AI steps each take minutes.
One short, honest email a week: what’s real, what’s hype, and what’s worth your money. Subscribe and get my full tool-stack guide, free.
Get the free guide