First, the real numbers, because you deserve them before the pitch: a creator we studied ran this system end to end, contacted 100 local businesses with automated personalized emails, and closed exactly one deal for $149. That is a 1% conversion rate and $149 in revenue from his first automation run. He said so himself, on camera. We are opening with that because most articles in this space would have titled this "How I Make $149 Every Day With Claude" and hoped you would not do the math. One run, one sale. What makes it interesting is not the first $149. It is that the whole thing, finding the business, building the deliverables, the portfolio, and the outreach, was done by Claude with the right connections and skills, which means the second run costs almost nothing but electricity.
Here is his exact system, translated into steps you can copy.
The idea in one paragraph
Local businesses on Google Maps have photos, reviews, and money, and most have marketing that has not changed since 2019. The offer is a premium package: a cinematic store walkthrough video, a modern website, a cinematic product ad, and five ad creatives. Individually priced, the package anchors at $745. Sold together, it goes for $149, an 80% discount that makes the owner feel like saying no is the expensive option. That pricing move is his, not ours, and he was blunt about the psychology: make the offer feel premium, then discount it so hard it feels stupid to refuse.
Step 1: find one business, not one hundred
The automation comes later. First you need a portfolio, and for a portfolio you need one real example. Open Google Maps, pick a city (he used Austin), and look for a business that photographs well: good reviews, plenty of images, ideally something with character. He picked a coffee shop bookstore combo, and that choice matters, because a creative business makes a creative portfolio.
You can ask Claude to do the picking. Give it the city and country and ask for well-reviewed local businesses with enough public images to work from. His reasoning for targeting good businesses rather than struggling ones: when a business that is already winning sees a package like this, it does not want to lose its edge, and a business with bad photos gives you nothing to build with.
Step 2: build the package
This is where the tools stack up. Claude on its own cannot generate cinematic video, so he connected it to Higgsfield, an AI media platform, through its MCP connector, which gives Claude access to current image and video models like Seedance and Nano Banana. Then he added downloadable skills, which are reusable command recipes, so the whole pipeline runs from a single prompt instead of twenty.
The package has four parts:
The walkthrough video. Screenshot every photo of the store from Google Maps, upload them, and ask for a cinematic 9:16 store walkthrough. Claude generates frames, you approve them, it generates clips. He slows the middle of each clip in his editor for that cinematic feel. His honest caveat: with Maps screenshots the result is decent, and once a real client sends proper photos from every angle it gets dramatically better.
The website. He pulled a UI style he liked from Pinterest, attached it as inspiration, and had Claude build a full site with animations from the business's Google Maps link. Deployed free on Vercel in a few clicks, live link ready to share.
The product ad. Using product images (his came from the website build itself), a skill generates keyframes, then a full cinematic ad. Ask for "hypermotion" if you want the high-energy style that is everywhere on social right now.
The five ad creatives. One prompt, five premium-looking static ads. His note on this part is the real business insight: creatives are not a one-time sale. Every product launch is a reason for the client to come back.
Step 3: the portfolio that does the selling
He asked Claude for brand names, picked "Main Street Motion", generated logo options, then uploaded the entire package and a Pinterest portfolio design he liked. Fifteen minutes later Claude produced a portfolio site with every deliverable presented interactively, including the scrollable website demo. Deployed to Vercel as a folder upload because of the video files. Total cost so far: subscriptions and a domain if you want one.
Step 4: the outreach run
Now the automation. A skill takes any city and produces a list of 100 qualified local businesses. He connected his Gmail to Claude, which drafted a personalized email for every business: mentions the business by name, describes what it does, points out what could be better about its online presence, and links the portfolio. The drafts sit in Gmail until you approve the send, which is the right amount of human in the loop.
Replies came in the same day. When a business asked to see the work before paying, he verified the business was real, sent the package watermarked, got paid, and sent it clean. That is the whole transaction.
The honest math, again
One hundred emails, one sale, $149. If you run it in a new city every week, his numbers say roughly $600 a month before you improve anything, and there are obvious improvements: better targeting, social proof pages (his interested replies kept asking for social media presence), follow-up sequences, and the recurring creative work from existing clients. But we are not going to project those numbers for you, because he has not published them yet. What exists on the record is one run, one sale, and a system that costs very little to run again.
The skill gap in all of this is not video editing or web design. Claude and the connected tools do that. The gap is knowing how to direct AI tools, chain them together, and talk to a client without underpricing yourself. That is learnable in weeks. We teach exactly that stack at Mindwand, and in the interest of the disclosure this article opened with: Mindwand is our product. Skool communities and Coursera courses cover similar ground, and the format matters more than the brand. Pick one and finish it.
What we would change about his system
Two things. First, the cold email volume: 100 personalized emails from a personal Gmail in one day is how you end up in spam folders, so warm the account up and send in smaller batches over a week. Second, the pricing: $149 for that package is arguably too cheap once you have three portfolio pieces instead of one, and his own anchor math ($745 of individual value) suggests the discount can shrink as the proof grows. If you want the guided version of this playbook, the client outreach templates, and the pricing frameworks, that is the kind of thing the Mindwand roadmap walks you through step by step, ours again, same disclosure.
His run proves the system works at exactly one data point. Your job is to make it two.