I run a ~$900k/yr residential cleaning company. Here's the automation stack that runs the back office, what each piece does, and the one system that completely failed.
I own a residential cleaning company in Colorado. We're at about $75k/month with 209 recurring customers and a 5.0 rating from 80+ Google reviews. I'm not a developer. Over the past couple of years I've built out an automation layer on no-code tools plus AI that runs most of the back office, and since this sub is where I learned a lot of what I know about service businesses, here's the full inventory. Numbers included, and one expensive failure at the end.
The stack is Make.com, Airtable, Stripe, Slack, a texting service, and our booking software. Total tool cost is a few hundred a month. I direct the builds with AI doing the heavy lifting; I don't write code.
1. Billing runs itself. When a crew marks a job complete in our booking software, an automation calculates the final amount (base price plus a distance adjustment from Google Maps), writes the job to a ledger in Airtable, and either charges the card on file or sends a Stripe invoice, depending on the customer's setup. Commercial clients get swept into one monthly invoice automatically. I don't touch routine invoicing at all.
2. Speed-to-lead texting. Every lead from Thumbtack, Angi, or our website gets a personal text from my number in under a minute: "Just got your quote request. Okay to call in the next couple minutes?" If they don't reply within five minutes, a follow-up goes out. Every lead also posts a card into a Slack channel with their name, address, home details, and a tap-to-call phone number. When a lead texts back, it shows up as a thread in Slack and anyone on my team can reply from Slack; the reply goes out as a text from our number. Thumbtack runs about one new recurring customer a week for us, and speed is most of why.
3. Hiring without phone tag. Applicants from our job ads get an automatic text inviting them to book their own interview through a scheduling link. When they book, they get a confirmation text with a photo of our office so they can find it, and the interview lands on my manager's calendar and mine. No back-and-forth, no chasing people who were never going to show.
4. Post-service survey that pays customers. After each service, customers get a short survey. Completing it automatically applies a $5 credit to their next invoice, no coupon codes, no manual step. Low ratings ping a Slack channel so we can call before it becomes a bad review.
5. Daily ops accountability. Our operations tasks live in Airtable. At 4pm a bot posts anything not done to Slack. At end of day, another automation scores each task on-time, late, or missed, and posts a digest. It sounds small. It changed how consistently things get done more than anything else on this list.
6. Crew paperwork. Every day at 2pm, an automation builds a PDF appointment sheet for each of the next day's jobs, with customer notes and access details, and files them in a Drive folder for the crews. Mileage and route reports per cleaner generate themselves monthly.
7. SEO content. Twice a week an automation drafts a blog post from a topic queue, generates an image, publishes to our site with proper schema markup, and cross-posts to our Google Business Profile. Our organic leads have grown steadily since we cleaned this pipeline up.
The failure: our referral program. We built an automated referral email that sent happy customers a $50 coupon code to share. It ran for months. Zero redemptions. Not low, zero. When I finally audited it, the whole design was wrong: one-sided (nothing for the referred friend), email-only for something people share by text, no way to attribute who referred whom, and the coupon literally couldn't apply the way we'd configured it. We're rebuilding it two-sided over SMS. Lesson I'd pass on: an automation that runs is not an automation that works. Instrument the outcome, not the activity.
What actually mattered, ranked: speed-to-lead first, billing automation second (it's not glamorous but it's hours every week and zero missed invoices), the ops scoreboard third. The SEO pipeline compounds but took the longest to pay. The stuff I'd skip if starting over: anything I automated before the manual process was solid. Automation amplifies whatever's there, including the broken parts.
Happy to answer questions about any piece, costs, or how I'd sequence it starting from zero.