How One Operator Built a $18,800/mo AI Outreach System for $480 in API Costs
A case study of a fully automated workflow that uses AI to find, pitch, and convert local business clients at a fraction of traditional agency costs.
Practical Summary
An individual built a system on a MacBook that uses AI to scan Google Maps for businesses with outdated or no websites. It automatically generates a modern webpage mockup and a short video pitch, then sends a customized message. This automated workflow acquired 47 clients in a month, generating $18,800 in revenue with only $480 in total API costs, demonstrating extreme cost optimization.
Why It Matters
This example provides a concrete blueprint for using AI to automate high-touch sales processes, drastically reducing labor costs (from a team of 8 to a single operator) and optimizing API spend. It showcases how AI tools can create entirely new, lean business models with very low overhead.
System Overview: Automated Local Business Outreach
The core of the workflow is an automated pipeline that identifies potential clients, creates tailored marketing assets, and initiates contact. The system runs on minimal hardware—a MacBook and an iPhone—highlighting its low startup cost and portability.
Step 1: Automated Client Prospecting
The system scans Google Maps across three cities to find small businesses, specifically filtering for those without a website or with an outdated one (described as looking 'like from 2005'). This automated lead generation eliminates manual research.
Step 2: AI-Generated Personalized Pitch Materials
For each identified prospect, the AI automatically generates two key assets: 1) A webpage effect mockup showing what a modern site for their business could look like, and 2) A short (10-second) video rendering. This creates a compelling, visual 'before and after' pitch.
Step 3: Automated Outreach and Scheduling
The system then automatically sends a customized message to the business owner, presenting the mockup and video. When a prospect replies, the AI can handle scheduling directly; in one example, it automatically scheduled a Zoom meeting while the operator was in a taxi.