AI Side Hustles Half-true — works only if you do the unspoken work

By the Income Reality Check Editorial Team · Updated 2026-10-26 · 7 min read

Mark Tilbury’s 7 ways to make money with AI: what the pitch skips

Verdict: Half-true — works only if you do the unspoken work. The AI does the production; you still have to find the clients and stand behind the output.

Mark Tilbury and his guest “Eman” — the entrepreneur Iman Gadzhi — spend this video ranking seven ways to make money with AI on a tier list, with income claims that climb from “about 100 bucks” per job to “around $10,000 a month.” The video has over 425,000 views and a clean premise: you don’t build the AI, you sell it to businesses that can’t use it themselves. Is that real? Mostly, yes — the tools work and the service businesses behind them are genuine. But almost every dollar figure in the video depends on one skill the two hosts mention only in passing.

What the video actually claims

The format is a tier list. Tilbury and Gadzhi trade off presenting six AI-powered service businesses, score each one, and then tease a seventh they refuse to reveal.

The six they do explain: AI real-estate photo and walk-through video enhancement (pitched at roughly $100 a job rising to “5 to 10,000 dollars a month” on retainers); repurposing a creator’s long-form content into newsletters and social posts using Claude (“a few hundred to a couple of thousand dollars a month per client”); “lead reactivation,” where an AI texting tool like HighLevel re-contacts a business’s old leads for “$50 to a couple hundred dollars for each appointment” and supposedly “$10,000 a month”; AI website makeovers for local businesses (“a few hundred to a couple of thousand dollars” plus a monthly fee); AI-generated UGC ads for brands (which Gadzhi calls the biggest earner, “a few thousand dollars a month per client”); and AI customer-support chatbots built on tools like Tidio or Chatbase (a setup fee plus monthly).

Then the pitch pivots. The seventh idea — the one both men call “the biggest opportunity right now” — is withheld. Instead of explaining it, they direct viewers to a free five-day event, the “AI Income Challenge,” starting November 7. That’s the real call to action here, and it’s worth naming plainly: the video is partly a lead magnet for an event. To the hosts’ credit, they repeatedly tell viewers to check the output, read it through, and verify the AI’s answers before sending anything to a client. That’s more honesty than most videos in this genre offer.

What the method actually requires

Here’s the pattern across all six ideas. AI collapses the production cost — the photo edit, the ad script, the website draft — to a few dollars in credits. It does nothing for the distribution cost: finding a business that will pay, convincing them, and keeping them happy. The U.S. Small Business Administration’s own guidance on getting new customers is a useful reality check — referrals, networking, introductory offers, follow-up systems. None of that is automated by a Higgsfield subscription. The video’s answer to “how do I get my first client?” is always the same: make a free sample and send it cold. That works. It’s also unpaid sales labor, and it’s the actual job.

Then there’s quality control, which the hosts correctly flag as the thing clients pay for — and which, by definition, is not passive. Someone has to read every repurposed newsletter, test every chatbot answer, and check every enhanced listing photo against the real property. Tool costs are real too, if modest: a HighLevel, QuillBot, Chatbase, or Higgsfield subscription typically runs from the low tens to low hundreds of dollars a month before you earn a cent.

A few of these ideas also carry legal exposure the video waves at but doesn’t detail. The lead-reactivation idea — texting a dentist’s or gym’s old leads — sits squarely inside U.S. telemarketing and consent law (the TCPA), which requires prior express written consent for automated marketing texts and an easy opt-out; the U.K., EU, Australia, and Canada each have their own regimes. Gadzhi says to only text a business’s own opted-in leads and to check local rules, which is the right instinct, but “it’s your client’s list, not yours” is not a complete compliance answer.

The UGC-ad idea is the one to watch. Generating a fake “creator” who holds a lip plumper and says “I’ve had so many compliments” is a testimonial for a product the AI avatar never used. In the U.S., the FTC’s Consumer Reviews and Testimonials Rule, in force since October 2024, doesn’t ban AI avatars outright — but it does prohibit testimonials that misrepresent a real experience, and it authorizes civil penalties of tens of thousands of dollars per violation. If the ad implies a genuine user when none exists, that’s the brand’s problem and, potentially, yours.

Does “no experience” survive contact with a paying client?

Not quite — and the hosts half-admit it. Gadzhi says the UGC-ad idea is “only A tier if you’re willing to learn how ads actually work,” because “making an ad that sells is a different skill.” That’s the whole game. Anyone can generate a clip. Knowing which hook stops the scroll is experience by another name.

The money also isn’t as effortless as the headline suggests. Independent 2026 rate roundups put the median UGC video around $175–$200, with roughly 80% of per-video rates under $500; beginners cluster at $150–$400, and only established creators command four figures. Freelance web designers average about $75 an hour, with small-business sites commonly selling for $2,000–$5,000 — numbers that match the video, but that you reach only after you can reliably win the work.

You are being sold the production step. You are buying the sales job.

Who actually wins this game

The people who clear real money here already have the one asset the video treats as free: a way to get clients. Gadzhi mentions, almost in passing, that he “ran a marketing agency” before this — so he knows how to price a retainer, handle objections, and keep a brand happy. That’s not a footnote; it’s the reason his version works. The winners are operators with an existing audience, a cold-outreach system, agency experience, or a budget to test offers. They’re also early movers: Gadzhi repeatedly says the real-estate photo idea is “low competition for now” and that “window will close.” Being early is doing real work before the crowd arrives — which is the opposite of passive.

What you’d realistically earn

Set expectations against a calendar, not a highlight reel. A motivated beginner sending free samples cold should plan for roughly $0 for the first one to three months while building a portfolio and a pitch that lands. Converting one or two local clients at $300–$1,000 each over the following months is a realistic win. The $5,000–$10,000-a-month figures the hosts quote are achievable, but they describe a small agency with several retained clients and systems — a year or more of consistent work away, not a weekend.

Worth remembering why the FTC cares about “$10,000 a month with AI” framing at all. In its 2024 Operation AI Comply sweep, the agency sued ventures that promised consumers they could make “$10,000 monthly” and “7-figure businesses” with AI tools, alleging tens of millions in consumer losses. Tilbury and Gadzhi aren’t selling a $35,000 program — their free event is a different animal. But the number itself is one regulators now treat as a red flag, and you should hear it that way too.

Who this is (and isn’t) for

This fits someone with maybe 10–15 hours a week, a tolerance for rejection, a small tool budget, and either some sales instinct or a genuine willingness to build it. If you like talking to local business owners and following up, the AI genuinely removes the part that used to need a design team. It does not fit someone expecting hands-off income, allergic to outreach, or hoping to skip the learning curve on what makes an ad or a website actually convert. For that person, “no experience needed” quietly becomes “no clients earned.”

What to remember

The tools are real, the service businesses are legitimate, and the hosts are more candid than most about quality control. The gap is distribution: AI makes the thing cheap to produce and does nothing to sell it. Treat the income numbers as the ceiling for a working agency, not the floor for a beginner — and treat the November event as the product it is. For more in this vein, see our looks at five AI side hustles featuring Iman Gadzhi and the earlier five “proven” ways to make money with AI.