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

By the Income Reality Check Editorial Team · Updated 2026-11-01 · 7 min read

Selling Meta ads with Claude AI: what the $10K lazy pitch leaves out

Verdict: Half-true — works only if you do the unspoken work. The AI-assisted Meta ads service model is legitimate and the pricing is realistic, but “lazy” misrepresents what is actually an active freelance sales business.

Patrick Dang’s video — 109,748 views at time of review — has a blunt title: “I Tried the LAZIEST Way to Make Money with Claude AI at Night.” The pitch is to use Claude and an AI image tool called Higgsfield to build Meta ads for brands, cold-pitch those brands on LinkedIn, charge around $2,000 per month per client, and stack five clients to reach $10,000 a month. Two students appear as proof: Brandon, allegedly earning $10,000 a month, and Teresa, closing “thousands of dollars” in clients helping real estate agents with Instagram content. The core model is real. The word “lazy” is where the pitch earns a closer look.

What the video actually claims

Dang — a former Oracle enterprise sales rep who now runs a YouTube channel with 431,000 subscribers — frames this as an “AI-powered service” business. The concept is simple: take a service companies already pay for (in this case, Meta advertising), then have AI do 70–80% of the production work instead of a full agency team. The tools are Claude for competitive research, copywriting, and HTML ad mockups, and Higgsfield for AI-generated product images.

His live demonstration is genuinely compelling. He drops a competitor brand’s URL into Claude, instructs it to analyze the Facebook Ad Library for the brand’s rivals, rank ads by signals like run time and variation count, and then recreate those ad formats in the client’s brand style — all inside Claude’s interface via Higgsfield’s integration. One prompt, a few minutes, a polished-looking set of static ads.

On pricing, he cites market rates of $150 to $1,000 per ad, with monthly retainers ranging from $700 to $50,000. His suggested starting rate for a beginner is $2,000 per month. “If you’re doing this all with AI,” he says, “you don’t mind the price because you’re going to have multiple clients to get to that $10K per month.”

What the method actually requires

First, the tools. Dang mentions early in the video that you can do this for “less than $100.” Technically possible. Claude Pro costs $20 per month; Higgsfield’s plans start at $19 per month for a Starter tier (270 credits), with Plus at $59 and Ultra at $129. For light sample work, $39 to $79 per month is achievable. But once you’re producing ad variations for two or three paying clients under real deadlines, Starter credits run out quickly — and Claude Pro’s usage limits can throttle you at scale. Budget $100 to $200 per month for tools once you have clients, not $39.

Now the larger gap: the sales pipeline. Dang recommends cold-outreaching decision-makers on LinkedIn — director-level and above — with personalized Loom videos that include a live competitor analysis and sample ads already created for the prospect’s brand. This is the right playbook. LinkedIn DMs do outperform cold email, with average response rates around 10% versus roughly 3–4% for cold email, according to 2026 outreach benchmarking data. But a 10% response rate means 90 out of 100 messages go nowhere. Assume a 20–25% close rate on discovery calls after that. To fill five client slots at $2,000 per month, you’re looking at roughly 30 to 50 qualified calls, which requires 150 to 200 personalized outreach messages — each one built around its own competitor research and custom Loom video.

That’s a part-time sales job layered on top of the service delivery work.

How long does each Loom video take? Dang’s demo, done live, runs about four minutes. But the prep work — pulling the Ad Library data, running the Claude analysis, generating and tweaking the sample ads — is closer to two to three hours for a beginner. Do that for 15 prospects a week and you’re burning 30+ hours before a single client pays.

There’s also client retention, which the video skips entirely. Real brands running Meta ads measure performance in click-through rates and return on ad spend. If the AI-generated ads don’t produce results, clients churn — and you’re back building the pipeline from scratch.

Who actually clears $10K a month from this?

Dang is himself the clearest example — and the irony is instructive. His primary income stream isn’t a Meta ads agency. It’s teaching people how to build one, via YouTube sponsorships, his “Sales Legacy” paid course, and coaching programs. He has 20 million organic views and a decade of enterprise sales experience. When he sends a cold LinkedIn message, his public track record makes that first reply far easier to earn than it will be for someone with a fresh account and no case studies.

The people who scale this model fastest tend to be former agency employees or in-house marketers who already understand ad performance benchmarks; people with warm networks of business-owner contacts who skip cold outreach at the start; or operators who already run client services in adjacent fields — copywriting, social media management, graphic design — and are layering AI tools onto an existing client base.

Beginners do break through. Kim’s experience in the video, landing responses from real business owners as a complete newcomer, is plausible. But she’s one of Dang’s coaching clients, a cohort self-selected for motivation and active support — not a random viewer who watched the YouTube video on a Tuesday night. And the FTC’s Endorsement Guides (16 CFR Part 255) require that testimonials featuring exceptional results “make clear to the audience what the generally expected results are.” Brandon’s $10,000-per-month outcome is clearly above average for a beginner. No such disclosure appears in the video.

What you’d realistically earn

The market data validates Dang’s $2,000-per-month starting point. Mid-range Meta ads freelancers and small agencies charge $1,500 to $3,500 per month, with creative services (which the Claude-plus-Higgsfield workflow provides) pushing rates toward the higher end. For context, a full-time social media specialist earns around $21 per hour on Indeed — roughly $44,000 per year. A two-client freelance operation at $2,000 each already matches that on a gross basis before you account for the lower hours involved.

So what does the actual income curve look like? For most beginners: zero revenue in months one and two (building samples, testing outreach, learning to close). A first client by month two or three if you’re sending 10 to 15 personalized pitches per week. Two or three clients by month six if you’re improving your outreach and closing rate. Five clients — the $10,000-per-month threshold — within twelve months if you’re consistent. That timeline is far slower than “this weekend,” and it’s also far faster than most traditional freelance ramp-ups. The model is genuinely accessible. It just isn’t fast.

Who this is (and isn’t) for

This makes sense for you if you have 10 to 15 hours per week available for outreach and delivery, can handle rejection without spiraling, and have even a thin background in digital marketing — running your own Facebook ads, managing a small business’s Instagram, anything. Some prior exposure shortens the feedback loop on what makes an ad actually work, which is the skill the AI tools can’t replace.

It doesn’t make sense if what you want is income that runs while you sleep. NerdWallet’s roundup of realistic side hustles consistently puts service businesses in the active-income column — you’re trading time for money until you either raise rates, productize, or hire. This is no different. High margins, yes. Passive, no.

For a broader look at what Claude-based service work actually pays once you’re established, our coverage of 10 Claude AI side hustles with full-time income potential and Claude AI side hustles that beat a 9-to-5 puts this model in context alongside others that have different time and capital profiles.

What to remember

Dang’s pitch is more honest than most of its genre — he says explicitly that sales skill matters, that 30 days isn’t guaranteed, and that you need to think strategically, not just prompt-and-publish. The gap between what he says and what the title promises is the real story: this is an active freelance business with a real sales pipeline, tool costs that climb with usage, and a three-to-twelve-month ramp before consistent income. Strip the “lazy” framing and you have one of the more workable AI service models out there. Add it back and you’re setting people up to quit after month two when the LinkedIn inbox stays quiet.

View counts and pricing figures may have changed since this article was published.