Video summary

How I Used Claude AI To Make $102k In 90 Days

Main summary

Key takeaways

Business

Business outcome & experiment results (proof)

  • Took an AI-enabled e-commerce business from zero → $100k+ sales in 90 days
    • Peak reported pace: $2,500/day revenue
  • Additional scale-up: $200k+ sales in first 6 months
  • Framed as a “beginner challenge”:
    • Chosen founder was 17 years old with < $100 starting budget
    • Goal: launch an AI business in 30 days

Core strategy (what to copy)

Use AI to remove bottlenecks in:

  • Store creation (design + content)
  • Product research (market validation + demand signals)
  • Product page generation (copy via JSON/template updates)
  • Ad creative (AI images/videos)

Three-tool execution stack:

  • Shopify (storefront + selling)
  • Claude (research + store/content generation, JSON updates)
  • Higgs Field / “Hicsfield” (AI creative studio for realistic ad visuals)

Playbooks / frameworks embedded in the process

Product selection framework (winning product traits)

A winning product is chosen based on:

  • Solves a painful, uncomfortable insecurity
  • Is not easily available in stores (less competition)
  • Is already selling/validated in the market

AI product research pipeline (Claude Product Finder)

  • Use a structured prompt/project (e.g., Product Finder V5)
  • Feed parameters like:
    • Market seed (e.g., “home and garden”)
    • Budget level (example: beginner $250–$500)
    • Buyer/build preference (example: long-term brand, not quick flip)
    • Model choice (example: Opus)
  • Output is ranked using demand proxies such as similar web traffic / visitors

Concrete operational steps (store + fulfillment)

1) Build the AI Shopify store (fast launch)

  • Use “Build Your Store” (free AI store builder connected to Shopify)
  • Workflow described:
    • Choose an industry (example: home & garden)
    • Select clean/simpler images (website visuals)
    • Claim store → connects and installs the AI store into Shopify
    • Select a Shopify plan
      • Mentioned promo: $1/month for first 3 months
  • Live store validation:
    • Reported speed: < 60 seconds to generate a usable store preview
    • Claim: store is uploaded with products and mostly complete

2) Set up fulfillment with suppliers (drop shipping)

  • Mentions AutoDS as supplier/automation:
    • AutoDS sources items and manages shipping (not from AliExpress directly)
    • AutoDS setup is prompted after store creation; can be skipped initially
  • Drop shipping method:
    • Add product to the store using an AliExpress link
    • Emphasized: orders should not be fulfilled from AliExpress
    • AutoDS handles sourcing/shipping

Product selection example + “why it won”

Case product: fungal nail renewal patch

  • Reported revenue impact: $200k+ revenue
  • Positioning:
    • A night patch for dirty/fungal toenails
    • Solves a socially uncomfortable insecurity
  • Applied the “pain + scarcity + validated demand” rationale

Product research tool example: self-watering planter

  • Claude product finder suggested markets and candidates; example decision:
    • Candidate: self-watering planter
    • Example sourcing cost: ~$5 from AliExpress
    • Suggested selling price: $30–$40
  • Validation tactic:
    • Check Amazon: require ≥ 1,000 sold in the last month (across multiple listings)
  • Final pick rationale:
    • Chosen for a “solve convenience pain” angle (no need to remember watering)

Content + product page execution (automation)

  • How to swap the template product text for the chosen product:
    • In Shopify product template, edit:
      • product.json
  • Use Claude to generate updated JSON:
    • Claude is fed the Shopify template JSON + a reference product
    • Reference chosen via an Amazon listing
      • Fallback: screenshot if access fails
  • Replace the JSON and save

Manual finishing touches noted:

  • Update price
  • Remove/clean variants
  • Rename product to remove warehouse-like naming
  • Mobile-focused edits:
    • Remove unnecessary sections (e.g., bundles)
    • Remove duplicated reviews
    • Change the variant picker to a dropdown

Marketing + sales execution (ads and testing)

Ad creation evolution

  • Previously: complex video shooting
  • Now: AI-generated image ads using Claude + Higgs Field

Testing plan

  • Create at least 5 ads with slight creative variations
  • Launch on Facebook
  • Use $30/day testing budget
  • Expected timeline: start seeing sales as soon as next day

Performance anecdote

  • Using “the same blueprint,” they launched and reached $150 in sales on the first day (for a different business)

Key KPIs / metrics explicitly mentioned

Sales & revenue

  • $100k+ in 90 days
  • $200k+ in first 6 months
  • Peak: $2,500/day
  • Example campaign: $150 first day

Ad testing

  • $30/day Facebook test budget
  • Expected conversion signal: next-day

Product validation thresholds

  • Amazon demand proxy: >1,000 units sold in the last month

Actionable recommendations distilled from the video

  • Daily habit: spend 15–20 minutes using Claude for “fun projects” to build comfort before execution
  • Use strict prompts (don’t just ask “find me a winning product”—add constraints for better output)
  • Select products with social/appearance pain and/or strong emotional discomfort
  • Validate via demand signals (Amazon sold counts; similar web/traffic signals from Claude)
  • Launch with multiple ad variations quickly and run short paid tests before iterating
  • Keep store editing lightweight: rely on AI generation, then do targeted cleanup (variants, mobile layout, pricing)

Presenters / sources mentioned

  • Creator/presenter: Not explicitly named in the subtitles
  • Scott (mentioned as an example user building a Pokémon clone)
  • Mark Builds Brands (source of the AI prompt used for product research)
  • AutoDS (supplier/fulfillment automation platform)
  • Shopify (store platform)
  • Claude (AI model/tool)
  • Higgs Field / “Hicsfield” (AI creative studio)

Original video