The Franchise Owner's Guide to AI That Actually Works
What AI can and can't do for your business — from someone who builds it
The Franchise Owner's Guide to AI That Actually Works
You've been pitched AI at least three times this year. Somebody at a conference, somebody in your inbox, somebody your buddy mentioned. They all promised it would "transform your business."
I build AI systems for service businesses. I'm going to tell you something those pitches won't: most of what they're selling you doesn't work for your business. Not because the technology is bad, but because it's solving the wrong problems.
Here's what actually works, what doesn't, and how to tell the difference.
What AI Is Good At (For Your Business)
AI is a machine that's very good at three things: responding fast, doing repetitive work consistently, and never forgetting to follow up. That's it. That's the whole pitch.
It's not magic. It's not going to rethink your business model. It's going to do the boring stuff that nobody on your team has time to do — and do it at 2 AM on a Sunday without complaining.
Things AI handles well for franchise operations:
Answering phones and texts, immediately, 24/7. A customer texts at 9 PM asking if you do transmission work. AI responds in 5 seconds with a real answer, asks about their vehicle, and books the appointment. No human involved. Your service advisor sees a confirmed appointment when they walk in Monday morning.
Following up after service. Customer picks up their car after a $400 repair. AI sends a "how's the car running?" text the next day. Three days later, asks for a Google review. Three months later, reminds them it's time for their next service. Every customer, every time, without anyone remembering to do it.
Appointment reminders. Confirmation when they book. Reminder 24 hours before. Morning-of text with the address. No-show rates drop because people actually get reminded — the same way their dentist reminds them.
Re-engaging dormant customers. AI identifies customers who haven't been back in 90+ days and sends a personalized message. Not a mass email blast — a text from your shop number that references their last service.
Notice the pattern. AI isn't doing anything creative here. It's doing what your best front desk person would do if they had unlimited time and a perfect memory. The value isn't intelligence — it's consistency and speed.
What AI Is Bad At (Don't Let Someone Sell You This)
Complex customer complaints. An angry customer who got their car back with a new scratch needs a human who can empathize, investigate, and make it right. AI escalates these, it doesn't handle them.
Upselling during service. "We found your brake pads are at 3mm — do you want us to replace them while it's on the lift?" That conversation requires trust, context, and judgment. Your tech and your service advisor handle this. AI doesn't.
Replacing your team. AI doesn't replace your service advisors, your techs, or your front desk person. It gives them their time back. Your front desk person stops playing phone tag with 30 customers and starts building relationships with the ones walking in the door.
Fixing a broken offer. If your pricing is off, your reviews are terrible, or your location is bad, AI won't save you. It amplifies what's already working. If nothing's working, you have a business problem, not a technology problem.
How to Evaluate Any AI Pitch
You're going to keep getting pitched. Here's a framework for deciding whether it's real or hot air.
Ask: "What specific task does this replace?"
If they can't name a specific, repetitive task that's currently being done badly (or not at all), it's vaporware. Good AI tools replace specific jobs — answering after-hours calls, sending review requests, reminding about appointments. Vague promises about "increasing efficiency" or "leveraging AI insights" mean nothing.
Ask: "What happens when it breaks?"
Every system breaks. Good AI systems fail gracefully — they transfer to a human, send an alert to the owner, or flag the interaction for review. Bad ones send customers gibberish at 3 AM with no fallback. Ask about the failure mode before you ask about the features.
Ask: "Can I see it work on a real call?"
Not a demo with a scripted scenario. A real call to a real number, with a real question your customers would ask. If they won't demo it live, ask why.
Ask: "What does month 2 look like?"
A lot of AI vendors charge a big setup fee and then... not much happens. The system goes live, nobody optimizes it, and six months later you're paying $1,500/month for something that handles 3 calls a day. Good vendors review performance monthly, tune the system based on real data, and proactively fix what's not working.
The Numbers That Matter
Forget impressions, engagements, and AI-powered insights. Here are the only numbers that tell you if your AI investment is working:
- Response time — How fast is the first response to an inbound lead? If it's over 60 seconds, you're losing people.
- After-hours capture rate — What percentage of after-hours inquiries turn into appointments?
- No-show rate — Did it go down after automated reminders? By how much?
- Review velocity — Are you getting more Google reviews per month than before?
- Cost per handled interaction — How much are you paying per call or text the AI handles? Compare that to the fully loaded cost of a human doing the same work.
If a vendor can't report on these numbers monthly, they're not measuring the right things.
What This Actually Costs
I'll give you real numbers because nobody else will.
A full AI phone + text + follow-up system for a single franchise location costs:
- Setup: $3,000-$5,000 (one-time). This covers building the system, training it on your services, integrating with your calendar and CRM, and testing with real calls.
- Monthly: $1,000-$2,000. This covers the AI infrastructure, phone/text costs, ongoing optimization, and support.
Compare that to:
- An additional front desk hire: $35,000-$50,000/year + benefits
- A traditional answering service: $500-$1,500/month (and they can't book appointments)
- Doing nothing: $5,000-$15,000/month in lost revenue from missed leads
The math isn't complicated. The system pays for itself if it captures even a few extra appointments per week.
The Bottom Line
AI for franchise operations is not about being cutting-edge. It's about doing the basics — answering phones, following up, reminding customers — with a level of consistency and speed that humans can't match during a busy day.
The shops that win over the next few years aren't going to be the ones with the fanciest technology. They're going to be the ones that never miss a call, never forget a follow-up, and never let a customer slip through the cracks.
That's it. No transformation required. Just the fundamentals, done consistently, at scale.
If you want to see where your business stands, I run a free response time audit where I mystery-shop your business and show you exactly where leads are being lost. Get yours here.
Software engineer. Former Spotify. Building AI agent security tools at Haun Lab.
More from the blog
Is your OpenClaw instance exposed?
Get a free exposure report. We'll scan public databases for your instance and tell you exactly what's visible from the outside.
Get your free audit