Video summary

$10K In 10 Days With a BRAND NEW Claude AI Dropshipping Store Full Guide (Just Copy Me)

Main summary

Key takeaways

Business

Results / Proof (store performance)

  • Claimed launch achievement: $0 → $10,000 in 10 days using “Claude AI” to automate most steps (product research, store build, creatives, Facebook ad launch).
  • Shopify analytics timeframe shown:
    • Gross sales: $11,000+ (Apr 26–May 11)
    • Average order value (AOV): $47
    • Conversion rate: 3.6% (described as “above industry standards”)
  • Later checkpoint:
    • By June 8, store made $817 that day; described as averaging $1,000/day
  • Scale volume (as stated):
    • 200+ orders in the period from 0 to $10,000 in 10 days

Product selection criteria (manual gating before AI research)

The “system” starts with product criteria to ensure AI research targets something viable:

  • Product solves a real problem
  • Amazon reviews: minimum 4.5 stars
  • Profit constraint: can sell with at least $35 profit (implies per-unit gross profit)
  • Demand trend: rising demand
  • Competition check: 3–5 active competitors selling the product on Facebook/TikTok

Frameworks / processes / playbooks used (bullet extraction)

AI agent “copy/paste system” playbook

  • Use Claude desktop app with Co-work (multi-agent workflow) so multiple steps run in parallel.
  • Use external data via a custom MCP connector:
    • Winning Hunter as the data source (“ads library on steroids”) to retrieve winning competitor creatives and landing pages.

Workflow stages:

  1. Product research (Claude selects from ~10 products)
  2. Extract/compile ad creatives + landing pages into a spreadsheet
  3. Clone competitor landing page into Shopify (HTML → Liquid theme)
  4. Generate product images for the store
  5. Generate video ads by referencing top-performing ads
  6. Build Facebook ad structure + test

“Non-emotional” decisioning framework

  • Humans overcomplicate and second-guess; AI is positioned as data-driven product selection to avoid “analysis paralysis.”

Tooling / operational setup (execution details)

Product research via Winning Hunter MCP

  • Set up a custom connector in Claude (via MCP URL) for Winning Hunter
  • In Co-work:
    • Use a prompt instructing Claude to use Winning Hunter (prompt pasted)
    • Set to “act without asking” (hands-off execution)
    • Use model: Claude Opus 4.8

Output after ~10 minutes:

  • Around 10 products
  • Each includes:
    • product name, what it solves, price, launch date, competitor info
  • Additional option: save results as doc/spreadsheet and pull:
    • actual ad creatives
    • landing pages

Landing page cloning into Shopify

Preconditions:

  • Must be on a paid Shopify plan
  • Free trial blocked because the MCP connector “breaks”

Steps:

  • Add the official Shopify connector
  • Use a skill: “clone link to my Shopify” to clone a competitor landing page
  • Use OpenAI vision once to extract structure from competitor page previews
  • Claude Design generates standalone HTML (~5–8 minutes)
  • Guidance:
    • Ensure ~70–80% of info is correct (images may be missing in preview; acceptable)
  • Export:
    • Download standalone HTML

Convert HTML → Shopify Liquid theme using Claude command (“clawed code” referenced):

  • Install Shopify Claude connector app
  • It generates a new theme (clones an existing theme like “Horizon” or “debut”)

Customization note:

  • Sections, text, colors, titles, icons are stated to be editable (not hardcoded)

Image generation for store assets

  • Skill used: “DTOC infographic generator”
  • Inputs: product image on plain white background
  • Constraint:
    • Claude has no native image model; must route via connectors

Options provided:

  • Higsfield (recommended): can create images + videos
  • Budget alternatives:
    • Google Studio with API key (e.g., “Google Banana Pro”) — requires topping up
    • OpenAI API (GPT image model “image 2” recommended as best among stated options)

Integration:

  • Add connectors in Claude; instruct it to use the chosen API/provider

Concrete product example (case study)

  • Winning product chosen by Claude:
    • “Snake shin protectors” / snake shin pads (protective leg gear)

Claude-provided messaging themes:

  • Hikers
  • Dog walkers
  • People scared of snakes

Seller confidence reasoning (business logic stated):

  • Small niche (described as good for beginners)
  • Niche also buys via older audience, so ads “don’t need to be crazy good”
  • Older demographics may be less accustomed to AI-looking content, so AI-style video can feel more “real”
  • Warns against overly large competitive niches (example: beauty niche → high CPMs like $50)

Scaling operations: fulfillment / “private agent”

  • Scaling too fast without the right fulfillment partner can lead to:
    • chargebacks and refunds, eliminating profit

Private fulfillment agent:

  • Fulfill Empire

Reasons given:

  • 10+ years experience
  • sourcing + warehouse reach (China)
  • fast dispatch:
    • warehouse ~1 hour from Hong Kong airport
    • same-day dispatch
  • average shipping time to USA: ~9 days
  • goal/benefit:
    • competitive product cost + shipping pricing

Implied KPI:

  • Reduce refund/chargeback risk by improving fulfillment reliability (not quantified)

Video ad production system (GTM execution)

Quantity / testing plan

  • Recommended video ad count:
    • 3–5 ads baseline
    • up to 10 ads suggested if using AI for speed
  • Scale logic:
    • scaling depends on finding a “creative” that drives performance
    • more quality creatives tested → higher chance of a winner

Ad sourcing → creative reference → AI generation

Find winning videos in Winning Hunter:

  • Search by product name
  • Change:
    • language → English
    • sort by → ad spend (to surface top performers)
  • If results are scarce:
    • generate search phrases/brand names via Claude

Reference example cited:

  • An ad concept using leg pads stopping rocks/snakes; downloaded HD and used as reference.

Generate new creative using Higsfield “Marketing Studio”:

  • Upload reference ad (Add reference)
  • Select product and create avatar:
    • objection handling: generated avatars may look too young
    • prompt to create older outdoor worker (30s/40s/50s, gray hair, etc.)
  • Video generation:
    • reference clips limited to 25 seconds
    • create two generations (two 25s segments)

Voiceover workflow:

  • Create voice with 11Labs (text-to-speech)
  • Settings: English accent (American), male, older voice; model V3 recommended

Script workflow:

  • Transcribe reference with AI
  • Copy script with timestamps
  • Ask Claude to generate a UGC script for the product based on that script
  • Optionally “V3 enhance” for more emotion

Assemble in CapCut:

  • Combine 2–3 clips
  • Add voiceover

Output quality note:

  • Example exported at 720p, but should target 1080p

Facebook ads strategy (structure + budget + testing)

  • Stated structure on a “Myro board”:
    • Campaign budget: $50/day

Ad setup described as:

  • 1 ad set broad with 3 ads
  • 2 ad sets broad with 3 ads each
  • Total stated: 3, 6, 9 ads (they also mention needing 9 ads total; earlier said “10 different ads,” implying a minor inconsistency)

Testing philosophy:

  • Facebook does “most of the hard work” once campaigns/ad sets/creatives are set correctly
  • Creators are the bottleneck; test many creatives to find winners

Actionable recommendations (as stated/implied by the system)

  • Don’t skip product gating:
    • Ensure reviews (≥4.5), rising demand, and ≥$35 profit before automating
  • Use data-backed product selection:
    • rely on Winning Hunter-powered research rather than manual gut feel
  • Clone winning landing pages, then adjust:
    • aim for 70–80% correctness in extracted page info; fix details in Shopify afterward
  • Protect margins with fulfillment reliability:
    • choose an experienced sourcing/warehouse partner to prevent chargebacks/refunds during scale
  • Scale by creative testing, not audience expansion alone:
    • produce 3–10 video ads, use winning references, and test systematically in Facebook broad ad sets

Presenters / sources mentioned

  • Claude AI / Claude Co-work / Claude Opus 4.8 (main AI system)
  • Winning Hunter (ads library via MCP connector)
  • OpenAI (used for vision/data extraction from competitor page previews)
  • Shopify (store building + official Shopify connector)
  • Higsfield (recommended for image/video generation; also “Marketing Studio”)
  • Google Studio / Google Banana Pro (alternative image generation option)
  • OpenAI image model / “image 2” (alternative option via OpenAI API)
  • 11Labs (text-to-speech for voiceovers)
  • CapCut (video editing/assembly)
  • Fulfill Empire (private fulfillment/sourcing “agent”)

Note: 11 Labs and Winning Hunter are the primary operational data providers; Fulfill Empire is the fulfillment execution partner.

Original video