Video summary

I Tried Selling AI Websites to International Clients (Real Results)

Main summary

Key takeaways

Business

Business Idea & Value Proposition

  • “AI Websites” for international clients: Identify high-value businesses in a target country that don’t have a good website, auto-generate a tailored site using AI, then sell + customize.
  • Positioning: Use AI to quickly build conversion-focused websites (e.g., booking flows, credibility sections, service pages, pricing), then negotiate on deliverables and customization.

Targeting & Lead Generation Process (Playbook)

Step 1: Choose a target country

  • Uses a proxy for accessibility: “top countries where Indian freelancers provide services.”
  • Example chosen: Canada

Step 2: Choose niche types to test (3 categories)

  1. Home contractors
  2. Cleaning companies
  3. Doctors / health professionals

Step 3: Find leads

  • Uses Instagram instead of Google Maps calls (based on constraints mentioned: card/payment and similar issues).
  • Approach:
    • Search for local accounts (e.g., Toronto).
    • Prioritize profiles with low views + low follower counts (assumed newer businesses that may lack proper websites).
    • Check whether they already have a website; if yes, assess if it’s weak (e.g., booking-only pages or low-quality sites).

Step 4: Outreach experiments (DM scripts & strategies)

  • Contractors
    • “New clients?” temptation angle.
    • Flirty/funny messaging test.
  • Cleaning
    • Send a sample website link first (light effort: “minimal and premium-looking” template).
  • Doctors
    • Send a personalized website per doctor (higher effort, but leverage credibility and a “warm” personalized experience).

AI + Website Production Workflow (Operations Playbook)

Generate content and page structure

  • Use Claude to draft prompts/content (e.g., before/after sections, testimonials, pricing, and booking intent).

Build the website quickly

  • Use Horizons (and “Hosting Horns”) to generate the site from the prompt.
  • Claim: ~5 minutes to create a site.
  • Includes one-click customization/personalization.

Personalization & iteration (when client engages)

Adjust:

  • Logo and brand colors
  • Services page content
  • Pricing structure
  • Contact CTA (e.g., a button linking to Instagram)

Go-live & integrations

  • Mentioned integrations: Stripe, PayPal, Google Adsense, and others.
  • Included: free domain for 1 year, business email, storage.

Quant Results & KPIs (From the Outreach Trial)

Overall DM volume & outcomes

  • Total DMs sent: 50
    • 25 contractors → 0 responses
    • 20 cleaning companies → 1 interested (asked for customization + prices)
    • 5 doctors → 1 notification (liked a comment) but no DM reply; follow-up described as a false flag

Interested-to-close conversion

  • Effective close rate (based on interest): 1 / 50 = 2%
  • The one cleaning company interest was followed through to a closed/negotiating client (voice notes + WhatsApp discussion + final site changes).
  • Delivery time: sample/personalized sites created quickly (claimed ~5 minutes for the initial build).

Pricing targets / negotiation

  • Reference (client’s previous website):
    • Previous website made for $400
  • Stated current market rate:
    • $600–$700
  • Seller’s offer (for a first international client):
    • $400–$500
  • Closing status:
    • Negotiating stage with partner review on May 29 mentioned.

Concrete Examples (What Was Actually Done)

Contractors experiment

  • Found contractors on Instagram with low engagement.
  • DM angles included:
    • “Are you taking on new clients?”
    • Conversation shift after response (described as slightly manipulative / not fully ethical).
  • Flirty/funny DM test:
    • Example message: “Helm you cute brother… are you taking new clients?”
  • Result: no replies

Cleaning companies experiment

  • Built a sample site for “Canada Cleaning” including:
    • before/after slider
    • testimonial section
    • pricing section
  • Sent DMs with the sample website link + offered customization.
  • Result:
    • One lead asked for customization + pricing
    • Then progressed to full build edits

Doctor experiment

  • Located doctors (Instagram content creators) in Toronto with no website.
  • Created personalized website per doctor (via Claude prompt + Horizons).
  • Sent DMs praising their content + offered a “free gift” (custom site).
  • Result:
    • Only a like/comment behavior
    • No real conversion after follow-up

Actionable Recommendations Implied by the Results

  • Niche-market fit matters more than “fun” messaging
    • Flirty/direct outreach to contractors produced 0 responses.
    • Cleaning companies responded when given a sample website and clear booking/credibility framing.
  • Send the website first to compress the sales cycle
    • Sample/personalized site links in DMs improved engagement enough to reach customization + pricing conversations.
  • Personalization should match expected ROI
    • Doctors received attention signals but conversion was weak—suggesting personalization alone wasn’t enough without a stronger conversion path or offer framing.
  • Win with fast iteration after interest
    • The winning client advanced negotiation after rapid adjustments:
      • brand identity + services + pricing

Frameworks / Playbooks Used (Explicit or Implicit)

  • Experimentation loop (A/B testing outreach by niche)
    • Different DM scripts per business type
    • Track responses → interest → conversion
  • Lead qualification heuristic
    • “Low views/low followers” on Instagram ≈ newer/smaller businesses likely lacking websites
  • AI production pipeline
    • Claude (content/prompt) → Horizons (site generation) → quick edits from client-provided assets/preferences

Presenters / Sources

  • Presenter: “Dhaval” (mentioned as Future Dhaval and Past Dhaval)
  • AI / tools mentioned:
    • Claude
    • Horizons / Hosting Horns (platforms referenced for website generation, hosting, and integrations)

Original video