I work a fairly boring corporate job in logistics. The salary is good, but the work is soul-crushing. Like many people, I always wanted to build something creative on the side — but there was one non-negotiable rule for me:
I didn’t want to be on camera. At all.
My face was not going on the internet. Period.
That’s what led me to research faceless content ideas, and eventually, AI-generated Instagram accounts.
How the Idea Started
The idea didn’t actually come from me — it came from my girlfriend.
She follows a lot of fashion accounts on Instagram and noticed something strange about some of them. The models looked too perfect. The lighting felt slightly off. After digging deeper, we realized many of those accounts were using AI-generated images.
That was the lightbulb moment.
If AI images were already working in fashion, why not try a similar approach in another niche?
Choosing a Niche (Even a Saturated One)
In January of last year, I decided to start a fitness motivation Instagram account.
Yes, I know — fitness is a heavily saturated niche. But that was actually part of the appeal. Fitness content doesn’t require complex storytelling. It’s mostly aesthetic images paired with motivational quotes. Simple, repeatable, and scalable.
At least in theory.
The First 3 Months: Frustration and Almost Quitting
The first three months were rough.
I used Midjourney and a few other AI tools to generate fitness model images. The biggest problem was consistency. Every image looked like a completely different person.
Followers noticed immediately.
Comments like:
- “Who is this?”
- “Thought this was a different girl”
- “Is this even the same person?”
It destroyed any sense of brand identity. Engagement was terrible, and I was spending nearly two hours a day just trying to generate usable images.
By month two, I was close to quitting.
Month 4: The Breakthrough — Consistent AI Characters
Around month four, I started experimenting more seriously.
I tested tools like APOB, Artbreeder, and several obscure AI generators I found through Reddit. The real breakthrough came when I finally figured out how to maintain a consistent AI character across posts.
Once I locked in “her” look — athletic build, dark hair, mid-20s vibe — everything changed.
People started following because they believed they were watching a real person’s fitness journey.
Content Strategy That Actually Worked
The strategy was simple:
- Post once per day
- Always feature the same AI-generated “person”
- Use gym shots, outdoor runs, yoga poses, and athleisure outfits
- Write motivational captions or short first-person stories about “my” fitness journey
Writing captions as a fictional character felt strange at first, but no one questioned it.
Consistency mattered more than originality.
First Money: Brand Deals at 6 Months
By month six, the account hit 8,000 followers, which made it eligible for micro-influencer platforms.
That’s when I landed my first brand deal:
- $75 for a single post promoting a protein powder
The brand provided product shots and messaging. I just integrated it into the feed. Total work time: about 20 minutes.
Income Breakdown at Month 11
Here’s what the account is making now:
Brand Deals
- Average: $280/month
- 4–5 offers per month
- I accept 2–3 that fit the account aesthetic
- Pay ranges from $50 to $120 per post
Affiliate Links
- Around $60/month
- Amazon Associates
- Links to workout gear and supplements
- Conversion rate isn’t great, but it adds up
Total Monthly Income
~$340/month
Not quit-your-job money — but for about 45 minutes a day, the hourly rate is solid.
The Hard Parts Nobody Talks About
1. The Guilt
There are real fitness influencers grinding every day — actually working out, filming content — while I’m running a fictional AI persona.
I’ve made peace with it. When brands ask directly, I’m transparent that it’s an AI-generated character. Still, it feels weird sometimes.
2. The Technical Learning Curve
Consistent faces, realistic lighting, natural poses — this took months to learn.
My first 100 images were unusable. Hands still look weird sometimes. Regeneration is part of the process.
3. Engagement Still Matters
Content creation can be automated — community building cannot.
Responding to DMs, commenting on other accounts, engaging with followers takes more time than image generation itself.
What I Wish I Knew Earlier
- Pick a niche where AI aesthetics make sense
Fitness, fashion, and travel work well. “Day in my life” content would be much harder to fake convincingly. - Don’t cheap out on tools
Free AI generators wasted two months of my time. Paid tools are worth it. - Treat it like a real business from day one
Separate email, separate account, track revenue, costs, engagement, and time spent.
What’s Next?
The next step is expanding to TikTok using short-form AI-generated video content. New image-to-video tools look surprisingly convincing, and early tests show promising engagement.
I’m also considering launching a second account in a different niche — possibly home decor or cooking. Once you understand the system, the model is very repeatable.
Final Thoughts: Is This Really Passive Income?
Right now, I spend about 12–15 hours per month on the account. Sundays are for batch-creating 20–30 images and scheduling posts. Daily check-ins are mostly for engagement.
Compared to the 40+ hours per month I spent early on, it finally feels somewhat passive.
This experience reinforced one key lesson:
Passive income is never passive at the beginning.
The first six months felt like an unpaid part-time job. But once the systems were built and momentum kicked in, the effort-to-income ratio finally started to make sense.