Robert Farrell is a roofer. He has been doing it for 14 years. He runs a crew of six, serves the western suburbs of Chicago, and his work speaks for itself -- consistent five-star reviews, repeat customers, referral business that keeps the lights on.
But Robert had a problem that his skill with shingles could not fix: he was losing business every single day because nobody was answering the phone while his crew was on a roof.
This is the story of what happened when we fixed that problem. Not over six months. Not with a massive technology overhaul. In four weeks. You can see the full results breakdown at the Farrell Roofing case study page.
The Problem
When Robert came to us, Farrell Roofing was generating roughly 10 online inquiries per month. For a roofing company with a solid reputation and 14 years in business, that number was embarrassingly low. Here is why.
Missed calls were the norm. Robert's crew starts work at 7 AM and wraps around 5 PM. During those hours, nobody is sitting at a desk. Calls went to voicemail. And in home services, voicemail is where business goes to die -- 62% of callers to service businesses hang up when they hit voicemail and call the next company on the list.
Follow-up was an afterthought. When Robert did send an estimate, the follow-up was manual. He would jot a note on a Post-it, intend to call back in a few days, and then forget because he was busy running jobs. Estimates sat unanswered for a week or more. By then, the homeowner had already signed with a competitor who followed up the next day.
The website was doing nothing. The site had not been touched since 2018. It was slow on mobile, had no clear calls to action, and was invisible in local search for anything beyond the company name. It was a digital business card, not a revenue-generating asset.
After-hours capture did not exist. Homeowners researching roofing at 9 PM on a Wednesday had no way to get answers or book an inspection. They would move on to a company that made it easier.
What We Built
We did not hand Robert a stack of software logins and wish him luck. We built three systems, connected them, and operated them alongside his team. The whole thing went from kickoff to live in four weeks.
1. A New Website Built for Revenue
We rebuilt Farrell Roofing's site from scratch with one goal: turn visitors into booked roof inspections. This was not a cosmetic redesign. It was a strategic rebuild focused on three things.
Local SEO targeting every service area. Instead of one generic "roofing services" page, we built dedicated pages for each city and suburb Farrell Roofing serves. Each page targets the exact searches homeowners use: "roof repair in Naperville," "roofing company near Downers Grove," "emergency roof tarping Wheaton." Within 60 days, Farrell Roofing ranked on page one for 30+ local keywords they had never appeared for.
Mobile-first, fast-loading design. Over 70% of roofing inquiries start on a phone. The old site took 6+ seconds to load on mobile. The new one loads in under 2 seconds. Every page has a clear call to action and a tap-to-call button.
Conversion-focused layout. Review badges above the fold. Service area map. Before-and-after project photos. A booking form that takes 30 seconds to complete. Every element exists to move a visitor toward booking an inspection.
2. An AI Chat Agent That Books Inspections 24/7
We deployed an AI chat agent on the new site, trained specifically on Farrell Roofing's services, pricing ranges, materials, and service areas. This is not a generic chatbot that says "someone will get back to you." This agent handles real conversations. See how this works in our AI engagement systems overview.
The agent answers questions homeowners actually ask: "Do you do flat roofs?" "How long does a full replacement take?" "Do you work in Aurora?" "What brands of shingles do you use?"
When a visitor is ready, the agent books a roof inspection directly on the crew's calendar, collects the property address, and sends a confirmation to both the homeowner and Robert's team.
It works at 2 PM on a Tuesday and at 10 PM on a Saturday. That after-hours coverage alone accounted for 35% of new inquiries in the first month.
3. Automated Follow-Up That Never Forgets
Every estimate request now triggers an automated follow-up sequence that runs without Robert touching anything:
Instant confirmation (within 60 seconds)
The homeowner receives a text and email confirming their inspection request, with Robert's direct number and a link to check their appointment time.
Value touch (day 2)
An email with a short guide: "5 things to check before hiring a roofer" -- positions Farrell Roofing as the expert and keeps the conversation warm.
Social proof (day 5)
A follow-up with a recent project photo and a customer testimonial. Real work, real results, from a homeowner in their area.
Final follow-up (day 10)
A direct message: "Hi [Name], just following up on your roofing estimate. We have availability opening up in [month]. Want me to hold a spot for you?"
Before this system, Robert followed up on maybe 30% of estimates. Now every single estimate gets four touchpoints over 10 days. No Post-it notes. No forgotten callbacks.
The Results
Four weeks after launch, the numbers told the story.
Here is the full breakdown:
- Online inquiries: 10/month to 50+/month
- Response time: Hours (sometimes days) to under 2 minutes
- Monthly revenue: Up 75% within 90 days
- After-hours capture: Zero to active 24/7
- Estimate follow-up rate: ~30% to 100%
These are not hypothetical projections. These are actual numbers from Robert's business, tracked from the month before launch through the first 90 days. See the detailed results with month-by-month data.
What Robert Says
“We went from missing half our calls to never missing a single inquiry. The AI agent books inspections while we are up on a roof. Best investment we have made in the business.”
What Went Wrong (And How We Fixed It)
I am not going to pretend this was flawless from day one. Here is what tripped us up.
The AI chat agent gave wrong service area answers in week one. We trained it on Farrell Roofing's website, but the old site listed service areas that Robert had dropped two years ago. A homeowner in Plainfield asked if they could get an inspection, the AI said yes, and Robert had to call back and explain they no longer serve that area. We retrained the agent the same day and added a service area verification step before booking.
Follow-up email three had a 9% open rate. The original day-five email was a generic "just checking in" message. Nobody opened it. We rewrote it as a before/after project showcase with the subject line "What a new roof looks like in [their city]" — open rate jumped to 34%. The lesson: automated does not mean generic. Every message in the sequence needs to earn the open.
The first week of call data showed a gap we did not anticipate. About 15% of callers asked about financing options, and the AI had no training on that. Robert offers financing through a third-party provider, but we had not included those details. We added financing FAQs to the agent's knowledge base and saw those conversations start converting immediately.
The point is not that everything went perfectly. The point is that we were monitoring closely enough to catch problems in days, not months, and we had the flexibility to fix them without Robert lifting a finger.
What Made It Work
I want to be direct about what made this project succeed, because it was not the technology alone.
Speed. We went from kickoff to live in four weeks. Not four months. Not a six-week discovery phase followed by a three-month build. Four weeks. Robert did not have time for a drawn-out process, and honestly, he did not need one. We knew what to build because we have done it before for home services companies.
Integration, not isolation. The website, the AI agent, and the follow-up sequences are connected. A visitor who chats with the AI agent at 9 PM gets a confirmation within 60 seconds, enters the follow-up sequence the next morning, and shows up in Robert's calendar with full context. Nothing falls through cracks because everything talks to everything.
We run it. Robert does not log into five dashboards. He does not manage the AI agent's training data. He does not write follow-up emails. We do that. His job is roofing. Our job is making sure every inquiry turns into a booked inspection and every estimate gets followed up on.
The Lesson for Contractors
If you are a contractor reading this, here is the takeaway: you do not need to become a tech company. You do not need to learn prompt engineering or spend your weekends watching software tutorials.
You need systems that work while you are on a job site. Systems that answer the phone at 7 AM when you are loading the truck. Systems that follow up on estimates at 6 PM when you are eating dinner with your family. Systems that book inspections at 11 PM when a homeowner is researching roofers after a storm.
That is what we build. If you want to see what it would look like for your business, the fastest way to start is our plan builder -- it takes two minutes and gives you a custom scope based on your trade, your service area, and your goals.
For more on how AI is changing home services, read our full guide: AI for Contractors.
See What AI Can Do for Your Contracting Business
Answer a few questions about your business and get a custom plan with pricing. Takes 2 minutes.
Build Your Plan