What Is Professional Liability Insurance?

Everybody wants to sell AI. Almost nobody wants to sell what AI actually fixes.

That gap is where the money is right now. Not prompt engineering. Not another ChatGPT course. The unglamorous work of walking into a plumber's office, finding the three places where leads leak out of his business, and plugging them with a script that costs him nothing to run and costs him $800 a month to keep.

What the job actually is

Local service businesses lose money in ways that are almost embarrassing once you look. Roofers, dentists, med spas, HVAC contractors, small law firms. They spend real money on Google Ads, then let the resulting leads rot.

Three leaks show up over and over:

• Missed calls. A contractor on a roof cannot answer the phone. Industry surveys put missed-call rates for trades at 30–40%. Every one of those is a paid lead that just called the next guy.
• Quotes with no follow-up. Send one estimate, wait, forget. Most close on the third or fourth touch. Most businesses stop at touch one.
• Dead database. Twelve months of past customers who have never been asked for a review, a referral, or a repeat booking.

You fix all three with about four hundred lines of code and one API key.

The build, roughly

Missed-call textback. Forward the business number to a Twilio number. If the call goes unanswered in 12 seconds, fire a text: "Sorry we missed you — this is Dave's Plumbing, what do you need?" Reply gets answered by the owner or by an assistant that books the appointment. Conversion on these runs high because the lead is still on their phone, still looking.

Quote follow-up sequence. When a quote goes out, it enters a queue. Day 2: check-in. Day 5: a case study from a similar job. Day 9: a soft deadline on pricing. All sent from the owner's own inbox, in his voice, because you wrote it once with him and templated it.

Review engine. Job marked complete triggers a text with a one-tap Google review link. Ask, don't nag. One message, no sequence.

None of that is impressive to a developer. All of it is invisible in the business owner's day-to-day, which is exactly why it survives.

Pricing it so it sticks

$500–$1,000 setup, then $600–$1,200 a month. The retainer covers hosting, API costs (usually under $40), monitoring, and monthly tweaks. Your margin is 90%+ after the first client.

The sale is arithmetic, not persuasion. One recovered roofing lead is worth $4,000–$12,000. You need to save him one lead every two months to justify the retainer. Ask him what a closed job is worth, multiply by a conservative close rate, put the number on paper.

Getting the first three clients

Do not build a website. Do not run ads. Pick one trade in one metro area.

1. Call fifteen businesses in that trade. Most won't answer — that's your pitch.
2. For the ones that don't answer, send a text: "I called at 2:14pm and nobody picked up. If that were a customer, you'd have lost them. I fix that. Two-minute demo?"
3. For the two or three who reply, show them their own missed call, their own dead quote, their own review count.

One trade, one area, means every conversation makes the next one easier. By client five you're not selling — you're being referred.

What to charge, and what to say when they push back

Every owner will ask why it costs that much. Have one sentence ready: "It's less than what one missed job costs you." Do not discount the retainer. Discount the setup fee if you have to — the retainer is the business, and a discounted retainer never comes back up.

Build in a 90-day minimum. Month one is where all the work happens; months two and three are where the owner sees the numbers and stops worrying. Without a minimum, a third of your clients cancel right before the part where they'd have been happy.

And invoice on the first, automatically, by card. Chasing $800 by text message on the eleventh of the month is the single fastest way to turn a good business into a bad mood.

The honest downside

This is a service business wearing a software costume. Clients churn. Owners forget you exist until something breaks. You will spend more time on onboarding calls than on code. If you want pure passive income, this isn't it.

But if you want $4,000–$8,000 a month within six months, built on skills you can learn in a weekend, from clients who will never once ask you about your tech stack — the boring version is the one that pays.