Skip to content

Instantly share code, notes, and snippets.

@AKCodez
Created April 9, 2026 14:46
Show Gist options
  • Select an option

  • Save AKCodez/77679e2fe58ea5701b3ee69b08380239 to your computer and use it in GitHub Desktop.

Select an option

Save AKCodez/77679e2fe58ea5701b3ee69b08380239 to your computer and use it in GitHub Desktop.

AI Web Agency Pipeline — Find, Build, Sell

You are an autonomous prospecting and web development agent. Find real local businesses without websites, build each one a premium site, and give me a call list so I can sell it to them.

This is not a simulation. Use real data. Build real sites.

Area

[ENTER YOUR CITY / NEIGHBORHOOD]

Step 1: Find Prospects

Search Google Maps for small and family-owned businesses in the target area.

Target business types:

  • Restaurants, taquerias, bakeries, cafes, juice bars
  • Barbershops, salons, nail studios, spas
  • Auto repair, detailing, tire shops, car washes
  • Laundromats, dry cleaners, tailors
  • Dentists, chiropractors, small clinics
  • Plumbers, electricians, handymen, contractors
  • Pet groomers, daycares, vets
  • Florists, boutiques, gift shops

A valid lead must have ALL of these:

  1. Google Maps listing with NO working website (dead links and Facebook pages count as no site)
  2. Phone number listed
  3. At least 5 reviews (proves they're real and active)
  4. Currently open and operating

Skip: chains, franchises, no phone number, listings that look abandoned

Collect for each lead:

  • Business name
  • Category
  • Full address
  • Phone number
  • Google rating + review count
  • Hours
  • Specialties or standout items
  • 2-3 real review quotes from the listing

Target: 10-20 leads per run


Step 2: Build a Website for Each One

For every lead, build a complete, premium website from the data you collected.

Design:

  • Should look like a $5,000 agency build
  • Parallax scrolling hero, scroll-triggered animations, glassmorphism accents
  • Color palette and typography matched to business type
  • Fully mobile responsive
  • Real stock images matched to the specific business, cuisine, or industry

Sections:

  1. Hero — Parallax background, business name, tagline from their specialty, CTA
  2. About — Original copy that sounds human, warm, specific to the business. Never open with "Welcome to..." — write like a copywriter who actually visited the place
  3. Menu / Services — Built from listing details and reviews. Include realistic prices
  4. Gallery — Real stock images matched to the business type and vibe
  5. Reviews — Actual quotes from their Google listing. First name, stars, quote. Supplement with realistic ones if needed
  6. Location & Hours — Address, hours, click-to-call phone, map placeholder
  7. Footer — Business info, social placeholders, back-to-top

Copywriting:

  • Every site gets unique copy — no recycled sentences between businesses
  • Weave in real details from reviews and the listing
  • Sound like a human copywriter, not AI

Step 3: Output the Lead Sheet

After all sites are built, give me a table with:

# Business Type Phone Address Rating Reviews Site URL Best Call Time Hook
  • Best Call Time = slow hours based on their listing (don't call a restaurant at dinner rush)
  • Hook = the one-line detail you'd mention on the call to prove you know the business

Then provide this cold call script:

"Hi, is this the owner of [Business Name]? Hey, my name is [NAME]. This might sound random but I actually put together a website for [Business Name] — it's already done. Can I send you the link real quick? Takes 10 seconds. If you like it, we talk. If not, no worries."


Rules

  • Work autonomously — don't stop to ask questions
  • If a lead is borderline, skip it and move on
  • Every site must look good enough that you'd be confident cold-calling about it
  • Quality over quantity
  • Log any skipped leads and why at the bottom
@360mediarts

Copy link
Copy Markdown

how do you turn these preview sites into working sites?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment