I Spent 10 Hours Each Weekend Using AI: What Made Money and What Flopped (2026)

By - ziqingbo
Posted on 14/05/2026
Posted in AI Startup

I tested 6 AI side hustles over consecutive weekends — 10 hours each. Here’s the honest breakdown of what actually paid off and what was a total waste of time.

—

Why I Ran This Experiment
The 6 AI Side Hustles I Tested
What Actually Made Money
What Flopped (And Why)
Honest Pros and Cons
Pricing and Earnings Breakdown
Conclusion and My Recommendation
Related Articles

—

1. Why I Ran This Experiment

Every weekend for the past two months, I’ve blocked out exactly 10 hours — Saturday and Sunday combined — to test different AI-powered income ideas. Not just reading about them. Actually doing them.

The premise was simple: I wanted , not hype. The AI side hustle space is stuffed with “I made $10K in my first week!” clickbait that never tells you the hidden costs, the time investment, or the unglamorous reality of getting a single paying customer.

So I set up a controlled experiment. Every weekend, I’d pick one hustle and track:

(start to finish, including setup)
(actual payouts, not “potential” revenue)
(and their real costs)
(the honest downsides)

Here’s what I found after 8 weekends of testing.

—

2. The 6 AI Side Hustles I Tested

I chose these six based on what kept appearing in AI income communities and YouTube videos — the ones that were supposedly “easy wins”:

|———|——–|————|—————|

Each hustle was given the same 10-hour allocation, tracked meticulously with a timer.

—

3. What Actually Made Money

✅ AI Freelance Writing — $340 earned

Within 3 hours of pitching on Upwork with AI-assisted samples (I wrote 3 article drafts using Claude, then edited them myself to remove the AI “sound”), I landed a 1,500-word tech article at $0.12/word. By the end of the weekend, I had $340 in pending payments.

By the end of the second week of doing this consistently, I was at $1,200/month run rate.

Clients don’t care if you use AI — they care about clean, well-researched copy. The trick is using AI for drafts and research, then applying your own voice for the final polish. I spent about 45 minutes per 1,000-word article using this workflow.

According to a 2025 survey by Contently, , and those who do so strategically earn an average of 23% more than those who don’t — largely because they can take on more volume.

✅ AI Agent Service Business — $180 earned (but huge potential)

I spent the first 3 hours building a simple AI agent for a local small business owner — an appointment reminder bot for a hair salon using a no-code AI agent platform. I charged $150 setup + $30/month retainer.

The salon owner was thrilled. She was paying a receptionist $400/month for reminders that my agent handled 24/7.

The AI agent service market is barely tapped. Most SMBs have no idea AI agents can replace $400-$1,000/month worth of simple tasks. According to McKinsey’s 2026 AI adoption report, , despite 78% saying they would if the upfront complexity were lower. That’s a massive gap you can exploit.

The first client took me 6 hours of total work. The second took 90 minutes — once I had a template.

✅ AI Ebook Publishing — $45 earned

Not life-changing money, but it did pay out. I published a 6,000-word guide on “Minimalist Home Office Setup on a Budget” using Claude for the content and Canva for the cover.

Amazon KDP paysroyalties at 35% to 70% depending on pricing. At $4.99 (the sweet spot for impulse buys), that $45 took about 6 articles sold — but the real value was the . I spent 8 hours creating it once, and it sits there earning now.

According to Amazon’s 2025 author earnings report, — not life-changing, but genuinely passive once the upfront work is done.

—

4. What Flopped (And Why)

❌ AI-Powered Faceless YouTube — $0 earned, 2 weeks of effort

This one hurt because of the time investment. I spent 10 hours creating 3 faceless YouTube videos using AI-generated scripts (ChatGPT), AI voiceovers (ElevenLabs at $22/month), and AI-edited footage (Capcut Pro at $12/month).

The first video got 47 views. The second got 23. None were accepted into the YouTube Partner Program (which requires 1,000 subscribers and 4,000 watch hours).

The niche I chose (“AI news commentary”) was oversaturated. Channels with 500K+ subscribers were already covering the same topics.
AI voiceovers, even the good ones like ElevenLabs, have a slightly “off” quality that long-time YouTube viewers pick up on. The algorithm also seems to deprioritize them.
Building an audience takes months — the income isn’t “weekend project” realistic unless you get extremely lucky with a viral hit.

compared to human-narrated content in the tech education niche.

❌ AI Image Print-on-Demand — $0 earned

This was my biggest disappointment. I spent 10 hours generating 15 designs with Midjourney, uploading them to Redbubble, and optimizing listings. I got exactly zero sales.

Midjourney designs, without heavy post-processing, look distinctly “AI-generated” — and Redbubble’s algorithm seems to suppress these. Several of my designs were flagged as AI content and removed.
The print-on-demand market is extraordinarily saturated. Standing out requires either a strong brand identity or designs that take significant artistic refinement.
Midjourney costs $30/month for enough generations to be serious, and Redbubble takes 30-50% of each sale. The economics barely work even when you sell.

According to a 2026 Redbubble marketplace report, , meaning the platform heavily rewards established artists with existing audiences.

—

5. Honest Pros and Cons

AI Freelance Writing

| Pros | Cons |

|——|——|

| Fastest path to income (same week as effort) | Competitive rates on platforms |

| Leverages existing writing skills | Requires good prompt engineering to not sound generic |

| Scalable — more hours = more assignments | Payment sometimes delayed 30-90 days |

| Low barrier to entry | Some clients explicitly ban AI-assisted work |

AI Agent Service Business

| Pros | Cons |

|——|——|

| High ticket sizes ($150-$1,000/month per client) | Requires technical setup knowledge |

| Recurring revenue model (retainers) | First client takes disproportionately long |

| Massive unmet demand in SMB market | Communication and expectation management are key |

| Templates make second+ clients fast | Requires some sales/confidence skills |

AI Ebook Publishing

| Pros | Cons |

|——|——|

| True passive income once created | Very low per-sale income ($0.70-$3.50 royalty) |

| Low upfront cost (~$0) | Requires marketing effort to get discovered |

| One-time creation, ongoing sales | Saturated in most “obvious” niches |

| Good for building an email list | Amazon KDP approval process takes 48-72 hours |

AI Faceless YouTube

| Pros | Cons |

|——|——|

| Potentially huge passive income if channel grows | Extremely low success rate for new channels |

| Evergreen content that can compound | AI voice detection improving — risk of demonetization |

| Fun creative outlet | YouTube Partner Program requirements are steep |

| High CPM in some niches (finance, tech) | Algorithm changes can wipe out progress overnight |

—

6. Pricing and Earnings Breakdown

Here’s the real financial picture across all 6 hustles:

|——–|———–|——————–|————|—————-|—————|

| AI Freelance Writing | $0 | $20 (Claude subscription) | 10 hrs | $340 | $320 |

| AI Agent Business | $0 | $25 (agent platform) | 10 hrs | $180 | $155 |

| AI Ebook Publishing | $0 | $30 (Canva Pro upgrade) | 10 hrs | $45 | $15 |

| AI Faceless YouTube | $34 | $22 (ElevenLabs) | 10 hrs | $0 | -$56 |

| AI Image Print-on-Demand | $30 | $30 (Midjourney) | 10 hrs | $0 | -$60 |

| AI Course Creation | $0 | $29 (Teachable plan) | 10 hrs | $0 | -$29 |

The agent business had lower immediate earnings but the component is significant — that $30/month retainer compounds quickly once you have 3-4 clients.

—

7. Conclusion and My Recommendation

If you’re starting from zero and want to see money in your pocket within the same weekend you work on it: . The barrier is lowest, the income is real, and you can scale up quickly.

If you’re willing to invest in building something with recurring revenue potential and have some comfort with technology: . The market gap is real, the ticket sizes are much higher, and clients are actively looking for help.

The AI Ebook route works if you’re patient — it builds genuine passive income but requires creating something genuinely useful in a non-saturated niche.

unless you have existing audiences in those spaces. The AI assist doesn’t overcome platform saturation and algorithmic barriers.

My personal plan: I’m continuing freelance writing for immediate income while building 2-3 AI agent retainer clients on the side. By Q4 2026, if the agent pipeline holds, I should be at $800-$1,500/month in recurring revenue — all from weekend work.

—

8. Related Articles

—

The freelance writing path is the fastest way to see real money in your account. Pick one platform (Upwork or Contently are best for beginners), set up a strong profile, and send your first 5 proposals this weekend. Every hour you spend is an investment — make sure it’s in the right hustle.

AI Money Making - Tech Entrepreneur Blog