💰
Make Money with AI
Making Faceless YouTube Videos With AI: My First Real Month
I spent 30 days building a faceless YouTube channel using only AI tools — here's what actually made money and what was a total waste of time.
Thirty Days, Zero On-Camera Appearances, $312 in Revenue
I never showed my face once. Not in a thumbnail, not in a B-roll clip, not even a hand holding a coffee mug. Thirty days of publishing to a brand-new YouTube channel, using AI for every single production step — and by the end, the channel had cleared $312 in AdSense plus one small affiliate commission. Not retirement money. But real money from scratch, which surprised me more than the number itself.
Here's exactly how I did it, what broke, and what I'd do differently starting today.
The Stack I Used (And What Each Tool Actually Cost)
Before anything else, you need to know the tools. I kept the stack lean on purpose.
| Tool | Purpose | Monthly Cost |
|---|---|---|
| ChatGPT Plus | Scripts, titles, descriptions | $20 |
| ElevenLabs Starter | AI voiceover | $5 |
| Pictory AI Starter | Video assembly from script | $19 |
| Canva Pro | Thumbnails | $13 |
| TubeBuddy Pro | SEO and keyword research | $9 |
Total monthly overhead: $66. That left a net of $246 on the month. Thin, but proving the model works is worth more than the margin at this stage.
My Actual Production Workflow
I published 14 videos in 30 days. Each one took roughly 90 minutes from idea to upload, once I had the process locked in.
- Find a keyword with real search volume. I used TubeBuddy's keyword explorer to find low-competition topics in the personal finance niche — specifically, questions people search but big channels ignore. Think "how to negotiate a medical bill" rather than "how to save money."
- Write the script in ChatGPT. I prompted it with: "Write a 600-word YouTube script on [topic] for a general adult audience. Be direct, conversational, and include three actionable tips." Then I edited. Always edit. The raw output is a first draft, not a finished product.
- Generate the voiceover in ElevenLabs. I tested four voices and landed on one called "Adam" — warm, mid-paced, sounds like a confident podcast host. Paste the script, export the MP3. Done in under three minutes.
- Assemble in Pictory. I pasted the script, uploaded the voiceover, and let Pictory auto-match stock footage to each sentence. It's imperfect — maybe 30% of the footage choices need swapping manually — but it's genuinely fast.
- Make the thumbnail in Canva. Bold text, high contrast, one strong visual. I used a template and modified it. Thumbnails matter more than almost anything else on YouTube, so I spent real time here.
- Write the description with ChatGPT. One paragraph summary, then five bullet points covering what viewers learn, then three keyword-rich sentences at the bottom.
- Upload, add chapters, publish. No scheduling tricks, no magic posting times. I just hit publish.
Pro tip: Don't skip the manual footage review in Pictory. One of my videos auto-selected a clip of a crying child for a segment about "cutting expenses." Catch these before they go live.
What Actually Drove the Revenue
Three of the 14 videos drove 80% of the watch time. All three hit the YouTube search tab, not the homepage. That confirmed what I suspected: search-optimized topics outperform viral bets when you're a new channel with zero subscribers.
AdSense RPM (revenue per thousand views) in the personal finance niche averaged around $8–$12 for me — consistent with what creators report publicly. One affiliate link in a video about budgeting apps earned a $47 commission from a single click. Small, but that's the model compounding over time.
The videos that flopped had one thing in common: I chose topics I found interesting instead of topics with proven search demand. Lesson absorbed.
Pro tip: Use TubeBuddy's "Competition" score alongside search volume. A keyword with 10,000 monthly searches but "Poor" competition is worth more than 50,000 searches in a saturated space.
Honest Limitations of the AI Setup
The voiceover is good. It's not great. ElevenLabs "Adam" sounds natural in short bursts but gets slightly mechanical over eight-minute videos. Some viewers noticed — a few comments said "this sounds AI-generated." It didn't tank retention badly, but it's a real ceiling on quality.
Pictory's stock footage library skews toward generic American business imagery. For my finance niche it mostly worked. If you're in a more visual niche — travel, food, fitness — you'll hit the library's limits fast.
And scripts still need human judgment. ChatGPT confidently wrote a section about a tax law that had changed six months earlier. I caught it because I knew the topic. If you don't know your niche, the AI will mislead you and your audience.
FAQ
Do you need to monetize immediately to make this work?
No. The real play is building a library of search-optimized videos that compound views over months. Revenue follows consistency, not speed.
Which niche works best for faceless AI channels?
High-RPM niches like personal finance, software tutorials, and health information tend to perform well. Avoid niches where visual authenticity matters, like cooking or travel vlogs.
Is this against YouTube's policies?
Using AI tools to assist production is allowed. Purely auto-generated, low-effort content with no editorial judgment can violate YouTube's spam policies. The human editing step matters — legally and for quality.
Can I start with less than $66/month?
Yes. ChatGPT free tier plus ElevenLabs free tier (10,000 characters/month) plus Pictory's free trial gets you a few test videos at near-zero cost before committing.
Bottom Line
A faceless AI channel can generate real revenue in month one — but only if you treat keyword research as seriously as production. The tools handle the heavy lifting. Your judgment determines whether anyone actually finds the videos. $312 isn't a business yet. But the process is proven, the costs are low, and month two is already running.
If you want to follow along with the next 30 days, bookmark AI Profit Automation — I'll post the full month-two breakdown when the numbers are in.