Miami, 3:08 AM
Ethan is asleep
A Shopify notification lights up his phone:
New order — Berlin — $118.40
Five minutes later: another one
Then another
By the time he wakes up:
- the supplier already received every order
- tracking numbers were sent automatically
- support replied to six customers
- three new ad hooks were ready for testing
- yesterday’s analytics already had a breakdown waiting for him
Ethan didn’t touch any of it
His store did $91,340 in revenue last month
After product costs, fees and ads, he kept $53,180
Tool cost:
Claude Pro — $20/month
Shopify — $39/month
That’s it
- No agency
- No VA
- No copywriter
- No media buyer
Just Shopify and Claude
But here’s the part nobody sees:
This was Ethan’s third store
The first two completely failed
The graveyard
October 2025
First store
Portable desk gadgets niche
Burned $1,700 in five weeks
The product looked cool in videos but solved nothing important
Conversion rate: 0.6%
Almost zero repeat visitors
December 2025
Second store
Pet niche
This one actually started working
Ads converted
Customers bought
Reviews came in positive
Then the supplier destroyed everything
Orders started arriving 18–26 days late
Refund requests exploded
Stripe froze payouts for 45 days
The product wasn’t the problem
Logistics killed the store
Total damage: almost $5,000 gone in under three months
Most people quit here
Ethan didn’t
He changed one thing
The thing that was broken in both stores
Claude was already there
- Open tab
- Random question
- Random answer
- Close tab
That was the entire problem
Most people use Claude like a smarter Google search
Every conversation starts from zero
Nothing compounds
On the third store, Ethan stopped treating Claude like a chatbot and started treating it like employees he hired once
Each workflow got:
- one role
- one permanent system
- one repeatable prompt
After that, Claude stopped being “AI”
It became the backend of the business
That shift changed everything
Six prompts. Two groups
ONE-TIME DECISIONS (run once)
product_find.prompt supplier_audit.prompt listing_builder.prompt
SYSTEMS THAT RUN DAILY
ads.prompt support.prompt weekly_diagnostic.prompt
Each one handles a different part of the store
Together, they replaced most of the operational work
1. product_find.prompt — finds products before they become saturated
Most beginners search TikTok for “winning products”
That guarantees:
- oversaturated markets
- copied ads
- weak margins
- endless competition
This prompt goes the opposite direction
Instead of trends, it searches for recurring frustrations people complain about online
The prompt forces Claude to identify:
- the exact customer
- the emotional trigger
- why competitors are weak
- why the product works specifically right now
- why the problem keeps happening
The important part:
- The product must show its value in under 10 seconds of video
- That alone filters out most bad products immediately
- Four ideas in under a minute
A beginner researcher would spend days doing the same thing manually
2. supplier_audit.prompt — prevents store-killing suppliers
Most Shopify stores don’t die because ads fail
They die because suppliers destroy customer trust
Late shipping
Bad quality
Fake tracking updates
Refund chaos
That’s what killed Ethan’s second store
Now every supplier gets audited before launch
Claude analyzes:
- shipping estimates
- review history
- negative complaints
- response speed
- suspicious rating spikes
It catches patterns humans usually miss
One bad supplier can destroy months of momentum
This prompt exists to stop that before money gets burned
3. listing_builder.prompt — writes pages that sound human
Most Shopify product descriptions feel fake instantly
“Premium quality”
“Innovative design”
“Revolutionary technology”
Nobody speaks like that
The line that changed everything was this:
“Write like someone texting a friend after using the product for two weeks”
That single instruction completely changed the tone
Now the descriptions sound:
- casual
- believable
- human
Not corporate
The highest-converting section ended up being:
“What this product does NOT do”
Most stores hide limitations
This section names them directly
Trust increased almost immediately after adding it
4. ads.prompt — writes TikToks that don’t feel like ads
The best-performing ads in 2026 barely look professional
Phone camera
Messy room
Natural lighting
Real frustration
The more polished the ad feels, the faster people scroll
Claude generates:
- hooks
- emotional pacing
- visual actions
- opening scenes
- alternative intros for testing
Then Ethan films 4–5 versions in one evening
No overthinking
No three-week editing process
Just fast iteration
The winning ad usually becomes obvious within 48 hours
5. support.prompt — handles customer support automatically
Once the store scaled, support volume exploded
40 messages
70 messages
Sometimes over 100 in a day
Same questions repeatedly: Where’s my order? Can I change my address? Why is shipping delayed? Can I get a refund?
Claude now drafts almost every reply automatically
Current approval rate: around 82%
Most replies only need tiny edits before sending
No customer waits half a day anymore
That alone improved trust massively
6. weekly_diagnostic.prompt — fixes the business every Sunday
Every Sunday Ethan exports Shopify analytics and pastes the numbers into Claude
Then Claude breaks down:
- funnel leaks
- weak products
- traffic quality
- conversion bottlenecks
- scaling opportunities
The most important instruction:
“Do not be polite”
Otherwise AI defaults to motivational garbage
The business improved much faster once the feedback became brutally direct
Five minutes every Sunday replaced hours of guessing
The real math
Monthly revenue: $91,340
Expenses:
Cost of goods: −$29,400
Ad spend: −$13,100
Shopify + fees: −$4,660
Net profit: ~$53,180
- Month 1: barely profitable.
- Month 2: $16K revenue.
- Month 3: $41K revenue.
- Month 4: $91K revenue.
The store didn’t suddenly become magical
The systems compounded
That’s the real advantage
The takeaway
Four months ago Ethan had:
- two failed stores
- frozen payouts
- almost $5K lost
- no working system
Last month: $53,180 net profit
Same apartment
Same phone camera
Same person
One thing changed:
Instead of opening Claude from scratch every day,
he built six permanent systems once
Those six systems became the store.
Follow me if you want to learn more about AI systems, Shopify, and modern online business strategies
@insomnia_vip
