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

By the Income Reality Check Editorial Team · Updated 2026-11-08 · 7 min read

Nate Herk ranks 13 AI businesses — which numbers actually survive a fact-check

Verdict: Half-true — works only if you do the unspoken work. The rankings and the dollar figures are mostly accurate, but they describe what the top of each market charges, not what a beginner earns while learning the one skill the video treats as an afterthought: getting clients.

Nate Herk’s video “Every AI Business Ranked from Worst to Best” counts down 13 ways to make money with AI, from faceless YouTube channels at the bottom to becoming a chief AI officer at the top. It’s one of his channel’s bigger hits, at 81,254 views when we found it. To his credit, Herk opens with a disclaimer most of these videos skip: the prices he quotes are “what businesses charge, not a promise of what you’re going to make.” That honesty is real. So is the catch — almost every figure he cites is a ceiling, and the work that gets you to that ceiling barely appears on screen.

What the video actually claims

The format is a scored ranking. Each of the 13 AI businesses gets three marks out of five — technical difficulty, market demand, and durability — and Herk works from worst starting point (AI SaaS, because distribution is brutal) up to best (chief AI officer, because it “can pay more than most businesses on this list”).

Along the way he drops specific price tags. AI website builds that dev agencies still charge 5K–10K for. AI receptionists at “1.5k to 5k plus a monthly retainer.” Single automation workflows at “500 to 2k.” AI workshops from $1,000 for a half day to 5K for a full day. Consulting audits at “1K to 15K.” And at the summit, a chief AI officer base salary he puts at “around 280K to 650K a year.”

He backs the ranking with a wall of statistics: AI integration work on Upwork up 187% in a year, AI video editing the fastest-growing Upwork skill at up 329%, a 62% wage premium for workers with AI skills, and — the headline for his number-one pick — an IBM survey of 2,000 CEOs finding 76% of their organizations now have a chief AI officer, up from 26% a year earlier. The thesis tying it together is clean: businesses at the bottom sell “the thing you build,” and the ones at the top sell “your judgment.” Sell judgment, earn more.

Do the headline numbers check out?

Mostly, yes — which is what makes this video more useful than the average AI-hustle pitch, and also where the subtle slippage hides.

The chief AI officer data is solid. CNBC, citing that same IBM study, reports 76% of more than 2,000 organizations now have a chief AI officer, up from 26% in 2025. The pay band is real too: Glassdoor puts the average Chief AI Officer salary at about $354,637, with the 90th percentile near $649,000 — so Herk’s “280K to 650K” is a fair reading of the top end. The wage-premium stat is where he rounds up. PwC’s 2025 Global AI Jobs Barometer found a 56% average premium for AI-skilled workers worldwide; 62% is the figure for the professional-services sector specifically, not the global average. A small thing, but it’s the pattern of the whole video: the more flattering number gets the screen time.

Here’s the part the CNBC coverage includes and the video leaves out. The same reporting quotes a Gartner analyst who doesn’t “expect that to go mainstream,” and an industry advisor asking whether the role is “transitional” — likely to be folded back into other executive portfolios once the AI hype settles. A separate CNBC piece is blunter in its headline: hiring a chief AI officer “may be a mistake.” The number-one “business” on Herk’s list is not a business at all. It’s a C-suite job you get hired into, one that requires years of credibility inside an organization, and one that a chunk of the people studying it think won’t exist in its current form in five years.

What the method actually requires

Strip the scores away and 11 of the 13 picks are the same business: you sell a service to other companies. Websites, receptionists, automations, workshops, lead gen, creative, consulting, audits — all of it is agency or freelance work with “AI” in front of it. And service work lives or dies on one thing the video mentions but never quantifies: finding people willing to pay you.

Consider the math Herk himself sets up. He says a single automation workflow goes for “500 to 2k,” then notes in the same breath that the work is “getting extremely commoditized” and that businesses are hiring for it internally. Both are true. So what’s the real bottleneck? It’s not building the automation — AI did make that part a weekend task. It’s landing the client, scoping the job, proving you delivered, and getting the next one before the price collapses. That’s sales, positioning, and reputation. None of it is passive, and none of it happens in an afternoon.

The platform data makes the gap concrete. On Upwork — one of the main places this work gets sold — the platform’s own resources show a wide earnings spread, with AI-skilled freelancers commanding a roughly 40% rate premium but plenty of sellers clustered well below a full-time income. Rate premiums are not the same as paychecks; a high hourly rate on zero booked hours is zero. And the consulting tier Herk ranks near the top? The U.S. Bureau of Labor Statistics puts the median wage for management analysts — the closest established category — at $101,860 a year, with the top 10% above $171,640. Respectable, durable, real. Also a long way from the five-figure-per-audit fantasy, and earned by people with established books of business.

One more thing worth knowing if you’re in the U.S. and you ever start quoting income figures to sell a course or coaching program of your own. The FTC’s Business Opportunity Rule requires that earnings claims be truthful, substantiated, and backed by written evidence, and the agency has proposed stronger penalties for sellers who inflate them. Herk’s “charges, not earnings” disclaimer isn’t just good manners. It’s the line that keeps a claim legal.

Who actually wins this game

The people who win Herk’s ranking are rarely the beginners he’s addressing. They’re operators who already have one of four things: an audience, a paid-traffic budget, prior domain expertise, or early-mover timing in a niche before it saturated.

Herk is candid that this describes him. His number-three pick, AI education and personal brand, is his own main business — “you guys are watching this video right now,” he says — and he admits he “did agency work and consulting and workshops long before I had a personal brand.” That’s the real sequence, and it’s the opposite of the funnel most viewers imagine. The audience came after years of unglamorous client work, not before. The chief AI officers clearing $650K aren’t career-switchers who watched a tutorial; they’re people with a decade of technical and organizational credibility. Lead-gen agencies “adding AI on top,” as Herk puts it, already knew how to generate leads. The AI is the easy layer. The hard layer was always there.

What you’d realistically earn

Let’s be concrete, because the video isn’t. Herk’s quoted prices are legitimate rates — but they’re per-engagement ceilings for established operators, not monthly take-home for someone starting out.

A realistic first-year arc for a motivated beginner doing AI service work looks less like “$1,000 half-day workshops” and more like this: several months of $0 while you build a demo, a portfolio, and a pitch; your first paid gig priced low (Herk’s own free guide reportedly argues your first client “should cost you money”); then, if you stay focused on one offer and get good at outreach, a climb toward a few thousand dollars a month over the course of a year. Some people reach the rates in the video. Most who quit do so before the sales skill compounds — which is exactly why Herk’s closing advice (“pick one and make it work”) is the most valuable sentence in the whole ranking. Against the implied dream of six-figure AI income, the honest range for year one is closer to the Upwork and BLS reality than the headline numbers.

Who this is (and isn’t) for

This path fits you if you have 10–15 hours a week to spend on outreach and delivery, some tolerance for months of low or no income, and either an existing skill (marketing, SEO, sales, a trade’s worth of domain knowledge) or the patience to build one in public. The website and workshop tiers are a genuinely sensible on-ramp — low cost, visual, fast to demo. It’s a poor fit if you’re expecting the “AI does it all” promise to cover the selling, if you need income this month, or if you’re drawn specifically to the chief-AI-officer summit without the corporate track record that role actually requires.

What to remember

Herk made a more honest video than the genre usually produces, and his framework — sell judgment, not just output — is sound. But the scores and the dollar figures describe the destination, not the road. Every business on his list runs on client acquisition, and that’s the one line item he priced at zero.

If you’re weighing these against one another, see our breakdowns of the 6 best AI businesses to start before 2027 for beginners and the 9 one-person businesses with more demand than competition.

View counts and figures cited above were accurate at the time of review and may have changed since publication.