Most developers building AI tools are building for other developers. That's the obvious market — we understand it, we hang out in the same spaces, and it feels comfortable.
But while we're shipping another dev tool, local businesses are drowning in missed calls, unanswered DMs, and leads that go cold overnight. A plumber doesn't have a "check Product Hunt" habit. A dentist isn't browsing GitHub. They have no idea AI chatbots exist for them — and almost no one is telling them.
I spent two weeks cold-emailing 50 local businesses to test this. Here's everything that happened.
Why Local Business Is the Hidden Goldmine
I've seen developers grind for months trying to sell a $15/month SaaS to other broke developers. Meanwhile, a plumber with 3 vans runs a $400K/year business and happily pays $300/month for software that makes his phone stop ringing at 11pm.
The math is different in local business:
- Budgets are real. Service businesses spend 10-15% of revenue on tools and marketing. A contractor doing $500K/year thinks $200/month is a rounding error.
- Competition is near-zero. No VC-funded startup is targeting "Roberts HVAC" in a city of 80,000. You have the field to yourself.
- They respond to email. Developer audiences are inbox-zero warriors who delete anything that smells like a pitch. A plumber who missed a lead last Tuesday will read your email carefully.
- The problem is visceral. Every local business owner knows exactly what a missed lead costs them. You're not selling a "nice to have" — you're solving a problem they complain about to their spouse.
The niche I focused on: service businesses with one to ten employees. Plumbers, electricians, HVAC contractors, dentists, law offices, pet groomers. Anyone who relies on inbound calls or contact forms and loses money when those go unanswered after hours.
The Exact Cold Email I Used
No fancy tools. Just a Google sheet of business names, emails from their websites, and this template:
Subject: Quick question about missed leads
Hi [Name],
I noticed you're a [plumber/dentist/contractor] in [City]. Quick question — what happens when someone tries to reach you at 10pm on a Sunday?
I build AI chatbots that answer customer questions 24/7 so you don't miss leads while you sleep. It knows your services, your prices, your hours — and it captures contact info from every visitor who asks.
Takes about a week to set up. Would a quick call make sense?
— Henry
That's it. No case studies, no features list, no pricing. Just the pain point and one question.
A few things I learned about why this worked:
- The subject line triggers a real memory (every business owner has lost a lead after hours)
- The "what happens when" framing is more powerful than "I can help you with"
- One question at the end beats a pitch — it opens a conversation instead of closing one
Results Breakdown
50 emails sent over 3 days to businesses in 6 categories. Here's what happened:
| Business Type | Sent | Replied | Interested |
|---|---|---|---|
| Plumbers / HVAC | 12 | 4 | 2 |
| Dentists / Clinics | 10 | 2 | 1 |
| Law offices | 8 | 3 | 2 |
| Contractors | 10 | 5 | 3 |
| Pet services | 6 | 2 | 1 |
| Restaurants | 4 | 0 | 0 |
Overall: 32% reply rate, 18% interested in a call.
That's significantly higher than typical cold email benchmarks (5-10% reply rate). The key insight: restaurants were a waste of time — their problem is operations, not lead capture. Service businesses with high ticket sizes and appointment-based models were by far the best fit.
Common objections:
- "We already have a contact form" — they mean a static form that goes to an inbox nobody checks
- "Our customers like talking to a real person" — fine, the bot captures the lead and you call them back
- "How much does it cost?" — this was actually a buying signal, not resistance
The businesses that replied were genuinely curious and sometimes even relieved. One HVAC contractor said: "I lost a $4,000 job last month because nobody picked up. How fast can you do this?"
How I Built the Chatbot
The stack: Python, Claude API, Railway for hosting. Two days of actual build time.
The core is simple — a conversation loop that loads a business's FAQ/service info as context and handles queries with Claude:
import anthropic
import json
client = anthropic.Anthropic()
def load_business_context(business_id: str) -> str:
with open(f"businesses/{business_id}.json") as f:
data = json.load(f)
return f"""You are a helpful assistant for {data['name']}.
Services: {', '.join(data['services'])}
Hours: {data['hours']}
Location: {data['location']}
Pricing: {data.get('pricing', 'Contact for quote')}
Always be friendly, answer questions from the info above, and if someone wants to book or needs more detail, collect their name and phone number."""
def chat(business_id: str, user_message: str, history: list) -> str:
system = load_business_context(business_id)
history.append({"role": "user", "content": user_message})
response = client.messages.create(
model="claude-haiku-4-5-20251001",
max_tokens=500,
system=system,
messages=history
)
reply = response.content[0].text
history.append({"role": "assistant", "content": reply})
return reply
Each business gets a JSON config file with their services, hours, and pricing. The bot is stateless — conversation history is stored in the frontend session. I wrapped this in a FastAPI endpoint and deployed to Railway in about 20 minutes.
The widget is 40 lines of vanilla JS that any business owner can paste into their website. No dependencies, no build step.
Total hosting cost: $0 (Railway free tier handles the traffic volume for a single local business easily).
The Starter Kit
After getting this working, I packaged everything into a ready-to-run kit:
- Lead scraper — pulls business emails from Google Maps results for any city + category
- Chatbot core — the Python + Claude API system above, fully configured
- Email templates — the cold email sequence that got 32% replies, plus follow-ups
- Business JSON configs — templates for 8 business types (plumber, dentist, lawyer, etc.)
- Railway deploy guide — step-by-step from zero to live in under an hour
If you want to run this exact system yourself — whether to sell chatbots to local businesses or just learn how it works — I packaged it all up.
Grab the full kit at https://payhip.com/b/GuGDX — $7 this week.
The local business market for AI tools is genuinely wide open right now. While everyone else is building for developers, the service businesses in your city are sitting on missed revenue every single night.
Go get it.
Top comments (0)