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

By the Income Reality Check Editorial Team · Updated 2026-07-11 · 7 min read

Iman Gadzhi’s 5 AI side hustles: the sales work the video skips

Verdict: Half-true — works only if you do the unspoken work. The five models are legitimate businesses, but four of them live or die on cold outreach and closing clients, and the video gives that part one sentence.

Iman Gadzhi’s “5 Proven Ways To Make Money With AI (No Experience)” has racked up more than 77,000 views by promising a way out of the scroll-and-save paralysis every aspiring hustler knows. The pitch: pick one of five AI-powered side hustles, let the software carry the hard part, and stack $10,000-a-month retainers. Is it real? The models are real. The “no experience” framing is where it quietly falls apart.

What the video actually claims

Gadzhi lays out five options. AI user-generated content (UGC): use a tool like Pictory to spin up fake spokesperson videos for e-commerce brands, charge $50–$200 a video or $2,000/month retainers. AI websites: build sites for local businesses with Lovable or Bolt, charge $999 upfront plus a $99/month maintenance fee. AI short-form clipping: run a creator’s podcast through Opus Clip and sell the clips as a $1,000–$2,000/month retainer. AI voice agents: set up a phone-answering bot with Vappy (Vapi) for a $1,000–$3,000 setup fee and $800–$1,500/month. And AI digital products: build an ebook or template once with his own tools (Synthesize AI, Ghostwriter OS) and sell it forever.

The recurring number is $10,000 a month. Five UGC retainers at $2,000. Or 20 digital-product sales at $500. “Less than one sale per day,” as he puts it. He ranks all five with a “CLEAR” scorecard and crowns digital products the winner at 20 out of 25.

To his credit, Gadzhi is more honest than most in this genre. He lists cons for every model. He says outright that “passive income” is a term he thinks is “BS.” He admits voice agents have the steepest learning curve. That candor is real, and it’s worth acknowledging before the reality check.

What the method actually requires

Here’s the sentence that carries the whole video: “having good sales skills is a key factor in success.” It appears once, under the UGC cons, and then largely disappears. Four of the five models—UGC, websites, clipping, and voice agents—are client-services businesses. In plain terms, you have to find a stranger who runs a real business, convince them to trust you with their marketing or their phone line, and close the deal. Repeatedly.

That’s not an AI skill. It’s a sales job, and it’s the actual job. This is what a social media marketing agency has always been—the “AI” swaps out the production labor, not the client-acquisition labor. To land five $2,000 retainers, industry norms for cold outreach put you somewhere around a 1–3% positive-reply rate. Getting five paying clients can mean thousands of cold emails or DMs, dozens of discovery calls, and a pipeline you refill every month as brands churn (Gadzhi concedes brands “churn through creative and agencies fast”).

The tool subscriptions are the cheap part, and they’re not free. Opus Clip’s watermark-free plans run $15/month (Starter, 150 processing minutes) to $29/month (Pro, 300 minutes)—fine for one client, tight the moment you have five. Voice agents are worse than they look. Vapi advertises about $0.05 a minute, but that’s the orchestration layer only; once you add transcription, the language model, voice generation, and telephony, real per-minute cost lands closer to $0.14–$0.30. A busy dental office doing thousands of call-minutes a month eats into that $1,000 retainer fast, and you’re the one absorbing the overage if you priced it wrong.

Who actually wins this game?

The people crushing these models already had the missing skill. Gadzhi himself built his following on Agency Navigator and Agency Incubator, courses teaching the social-media-marketing-agency (SMMA) model—which is to say, he teaches sales and client acquisition, and has for years. The winners in his comment section tend to be people who can already sell, who have an existing audience to pull from, or who have a budget to run paid traffic while they learn. Early movers matter too: the UGC and clipping lanes were open two years ago and are, by his own scoring, now the most crowded on the board (he gives UGC competition a 1 out of 5).

Digital products, the “winner,” has the same catch dressed differently. Building the ebook is easy now. Getting 20 people a month to buy it is the entire game, and Gadzhi admits it—“the product doesn’t just sell itself because it exists.” He compares it to “a slot machine with unlimited plays.” A slot machine is a good metaphor for something where most pulls lose.

What you’d realistically earn

NerdWallet’s reporting on realistic side hustles is a useful anchor: the median side-hustle income sits around $200 a month, and the average is dragged upward by people who’ve been at it more than a year and built a reputation. Beginners should plan for roughly $100–$400 in the first few months, with $500–$2,000 possible after six-plus months of consistent work—if you stick with it.

Against that, “$10,000/month with no experience” is not a lie so much as a best-case ceiling presented as a starting point. Some people do hit five figures with an AI agency. They are not the median, and they did not get there in month one without selling. Expect months of unpaid pipeline-building before the first retainer clears—and expect to keep selling, because four of these five stop paying the day you stop showing up.

One more thing worth saying plainly for U.S. readers.

The FTC has spent the last two years specifically targeting “make money with AI” pitches. Its 2024 Operation AI Comply sued several AI-e-commerce operations, and the agency later banned Ascend Ecom after alleging it defrauded consumers of at least $25 million with promises of AI stores producing five-figure monthly income. The FTC has also proposed a new Earnings Claim Rule that would force anyone marketing a money-making opportunity to keep written proof behind income figures. None of that names Gadzhi—his video isn’t selling a done-for-you store, and he explicitly disowns “passive income.” But the regulatory temperature around $10k/month AI claims is exactly why the substantiation gap matters. (Readers outside the U.S.: the U.K.’s ASA and Australia’s ACCC police misleading earnings claims similarly.)

Who this is (and isn’t) for

This makes sense if you can dedicate 10–15 hours a week, you’re comfortable doing cold outreach and getting told no a lot, and you treat the AI as a way to deliver a service faster—not as the business itself. A little runway helps: enough savings to work an unpaid pipeline for two or three months. It does not make sense if the appeal was specifically the “no experience, no selling” framing. If cold-emailing a hundred plumbers sounds worse than the problem you’re trying to escape, four of these five will feel like a bait-and-switch, and the fifth is a marketing challenge in disguise.

What to remember

The AI in these models genuinely works—it collapses the production time that used to be the barrier. What it doesn’t do is find the client, close the deal, or drive traffic to your product, and that’s the part that determines whether you earn $0 or $10,000. Gadzhi’s video is a fair map of five real businesses. Just read the one sentence about sales skills as the headline, not the footnote. For more on the same pattern, see our looks at 5 ways to make money in 2026 with AI and how to actually make money using AI.