Make $$$ Selling AI to Boring Businesses
Greg Isenberg · 2026-10-26
💡 Quick Take
1. The speaker uses custom MCP (Model Context Protocol, a system that connects AI models to external tools) connectors inside Claude Code to scrape local directories and find broken landing pages in minutes.
2. Claude Code connects directly to Meta advertising accounts through an official MCP tool to draft and publish localized ad campaigns from text prompts.
3. According to the speaker, Opus 5.5 automatically turns raw phone video into polished short-form video ads for local service funnels.
4. The speaker tested AI voice receptionists in his mobile diesel repair business and recommends avoiding them because prospective clients hang up on robotic voices.
5. According to the speaker, agencies should skip optimizing for personal AI search bots like Muse or Instinct right now because real homeowners still look for contractors on Google Maps.
📊 Detailed Explanation
James Dickerson targets service companies generating between two million and twenty million dollars in annual revenue. He focuses on high-ticket trades like plumbing, heating and air conditioning, septic pumping, and commercial kitchen hood cleaning. A typical five-million-dollar plumbing shop spends one hundred fifty thousand to two hundred fifty thousand dollars every year on advertising. Their average cost per lead sits near one hundred twenty-nine dollars. The speaker advises avoiding low-ticket, recurring services like lawn care or restaurants. Emergency repairs cost homeowners thousands of dollars, which makes fixing even a small leak in a website funnel worth massive money to the owner.
The speaker runs automated audits on local markets using Claude Code and web scraping tools. In a Dallas test run of top-rated plumbers, four out of five landing pages failed to repeat the headline promise from their paid search ads. Three out of five pages hid their Google review ratings entirely. Most pages forced mobile visitors to scroll down three full screens before displaying a booking button. The speaker recommends pitching direct-to-consumer style funnels rather than generic web design. Effective setups include seventy-nine-dollar maintenance tune-up tripwires (low-cost introductory offers that immediately offset ad costs), instant price calculators, and repair-versus-replace quizzes that guide homeowners toward large furnace purchases.
Modern coding agents allow one operator to build agency-grade assets without hiring developers. Operators use Claude Code to scaffold responsive landing pages, wire up forms, and connect ad platforms. Claude Code uses a Meta MCP connector to construct targeted local ad campaigns inside a twenty-five-mile radius. For creative assets, the speaker points to Opus 5.5 to chop simple owner-recorded videos into paid social clips. According to the speaker, agencies should never pitch AI agents or automated workflow tools directly to blue-collar owners. Main Street owners often mistrust artificial intelligence buzzwords. The winning strategy uses AI as a silent production engine while selling guaranteed phone calls and booked appointments.
Client acquisition relies on blending software scraping with classic direct-response marketing. The speaker recommends compiling a Dream 100 list of prime local targets, inspired by sales strategist Chet Holmes. Instead of blasting generic cold emails, operators research each owner's online reputation and mail them physical packages. Dan Kennedy called this lumpy mail: sending a three-dimensional item like a brass plumbing pipe fitting alongside a physical letter that points to an interactive audit page. For newcomers without testimonials, the speaker advises building the first funnel for free. That initial free project gives the operator a verified local case study, which removes sales resistance for all future outreach.
🎯 Tech Expert Opinion
Repositioning AI tools as hidden backend leverage rather than client-facing deliverables is the strongest strategic advice in this presentation. Too many technical founders build thin wrappers around language models and try to sell conversational agents to trade contractors who do not want them. James Dickerson correctly identifies that trade businesses already invest heavily in customer acquisition, yet they lose twenty percent of their paid traffic to broken mobile pages. Using AI coding agents to close that specific commercial gap turns high-tech tools into practical cash-flow improvements without forcing business owners to change their daily habits.
The speaker's hands-on data with AI phone agents highlights an important boundary for field automation. While video demos of synthetic dispatchers look impressive on social media, real customers facing flooded basements or broken trucks demand immediate human empathy. Dickerson's finding that real callers hang up when they encounter automated voice receptionists confirms that developers should avoid replacing human touchpoints in urgent, high-ticket services. Generative software delivers its highest value when writing landing page code, analyzing paid search performance, and generating client proposals, not pretending to run a phone bank.
This business model makes sense for technically capable operators who understand basic performance marketing and want to monetize tools like Claude Code without building software startups. However, success depends far more on direct-response sales persistence and local relationship-building than on coding ability alone. The speaker warns that securing clients requires weeks of consistent value delivery, and agencies should wait on consumer AI agent optimization until local search habits actually drift away from Google Maps. For freelancers willing to show up in their local markets and speak the language of small-business revenue, this workflow provides a grounded blueprint for selling modern tech.
Kanal: Greg Isenberg