How to Automate a Cleaning Business with an OpenClaw AI Agent
Prefer to watch? Watch on YouTube · Subscribe for new builds · Read the transcript
In this video I walk through how Tony, a residential cleaning business owner who doesn't code, built one OpenClaw agent that runs his whole inquiry-to-review pipeline. If you run a service business, or know someone who does, this is a practical look at what an AI agent can take off your plate. I recorded this in April 2026. Tools change fast, so check the current OpenClaw docs before you build your own.
What I cover in this video
- 0:00 Tony's business and why the phone was eating 4 to 5 hours of his day
- 1:32 How his daughter set up OpenClaw and how a contact form kicks off the agent
- 3:49 The quote calculator logic built from Tony's own pricing rules
- 4:34 Booking straight into Google Calendar and the reminder sequence
- 5:19 The post-job follow-up that collects Google reviews and flags problems
- 6:04 A monthly re-engagement agent that wins back one-time clients
- 6:51 The results: time saved, faster responses, and a better rating
- 9:52 The pipeline you can reuse for any service business
The problem: a phone that never stopped ringing
Tony runs a residential cleaning business with about 12 cleaners covering a metro area. Most of his clients are on recurring weekly or bi-weekly cleans. The operation was solid, but Tony was spending 4 to 5 hours a day on the phone answering people who found him on Google. He was giving quotes, scheduling jobs, sending confirmation texts, following up on no-shows, and calling clients back. After all that he still had to actually run the business: managing his team, handling supplies, dealing with complaints.
He hired a part-time person to handle the phones, which helped until that person left. Finding and training someone new for a job that was mostly answering the same questions over and over felt like a waste. So his daughter, who's in her 20s and knows a lot more about AI than he did, set up OpenClaw for him. They built it together by describing what the business needed.
Inquiry, quote, and booking with no human involved
It all starts with the contact form on his website. Someone enters their name, address, the type of cleaning they need, how big the house is, and any special requests. That submission triggers the agent.
- Fast first response. Within about 60 seconds the person gets a text from Tony's business number, which is routed through the agent. It introduces itself as Tony's scheduling assistant, thanks them, and asks follow-up questions if the form missed anything.
- Answering questions. The agent answers what's included in a clean, what products they use, whether they do move-out cleans, and how pricing works. These are the questions Tony used to answer on the phone 30 times a day.
- The quote calculator. Tony spent a few hours with his daughter turning the pricing math he'd been doing in his head for years into a set of rules the agent follows. For example: a two-bedroom apartment of around 600 to 900 square feet falls in a set price range, add this much for an extra bathroom, add this much if it hasn't been cleaned in over a month, take this much off for recurring clients.
- Booking. The agent sends the quote, explains what's included, mentions the recurring discount for weekly or bi-weekly bookings, and sends a booking link. It's connected to Google Calendar, shows only the open slots, books the job, and updates the calendar automatically.
The whole thing can happen at 11 p.m. on a Sunday while Tony is asleep.
Reminders, reviews, and re-engagement
Once a booking is confirmed, the agent runs a follow-up sequence:
- 2 days before: a confirmation text.
- The day before: a reminder with the approximate arrival time.
- The morning of: a final reminder with the cleaner's name and any details the team needs, like a gate code or a dog to watch out for.
That sequence replaced roughly 30 minutes of manual outreach per job. At 40 jobs a week, that's about 20 hours of texting gone.
After each job, the agent thanks the client and asks if everything looked great. If the reply is positive, it asks for a Google review and sends a direct link. If the client mentions something that wasn't right, it flags it to Tony in Telegram right away so he can call before it turns into a complaint. His Google reviews went from 47 to 214 in a few months, and that kind of growth helps Google Maps ranking and organic traffic.
He also built a simple re-engagement agent that runs once a month. It finds clients who had a one-time clean and haven't booked again in 45 days or more, asks if they're ready for their next clean, and mentions the recurring discount. Tony said it brought back six clients in the first month, which at his prices was about $900 in recovered revenue.
The results and the pattern you can copy
Tony went from 4 to 5 hours a day on the phone to about 45 minutes reviewing what the agent flagged. His response time went from hours to under a minute, and people shopping for cleaners usually book whoever responds first and professionally. His Google rating went from 4.1 to 4.8. When he did hire a new part-time person, they moved into quality control and client relationship management instead of answering the same questions all day.
The most important part is that Tony isn't a tech person. He described his business in plain language: what happens when someone calls, what they ask, how pricing works, what his team needs to know, and what makes a good job versus a bad one. That description basically became the agent. He built it in about a week with his daughter's help. OpenClaw works as a business tool here. You don't need to write code, you need to know your operation and review what the agent does.
This works for lawn care, HVAC, roofing, dog grooming, tutoring, bookkeeping, or any business that handles inquiries, quotes, scheduling, or follow-up. The details change, but the structure stays the same:
- Inquiry comes in
- Agent responds
- Agent qualifies the lead
- Agent gives the quote
- Agent books the appointment
- Agent confirms the appointment
- Agent follows up
If you want to see more of what these setups can do, my AI agents page covers the bigger picture.
Key takeaways
- One OpenClaw agent handles Tony's inquiries, quotes, Google Calendar bookings, reminders, and review requests, and he doesn't code.
- The quote calculator is just Tony's pricing knowledge written down as rules the agent follows.
- Responding in under a minute matters because customers often go with the first business that gets back to them.
- Asking happy clients for reviews and sending unhappy feedback to Tony in Telegram raised his rating from 4.1 to 4.8.
- The inquiry-to-follow-up pipeline carries over to almost any service business.
FAQ
Can a non-technical business owner build an OpenClaw agent?
Yes. Tony doesn't code. His daughter helped him set up OpenClaw, and the agent was built from his plain-language description of how the business runs. It took about a week.
How does the OpenClaw agent generate cleaning quotes?
It follows a set of pricing rules Tony and his daughter wrote down, like a set price range for a two-bedroom apartment, extra charges for added bathrooms or homes not cleaned in over a month, and a discount for recurring clients.
Can OpenClaw book appointments in Google Calendar?
In Tony's setup, the agent connects to Google Calendar, shows only open slots, books the job, and updates the calendar automatically, with no human involved.
How does the agent help get more Google reviews?
After each job it asks the client if everything looked great. Happy clients get a direct link to leave a Google review, and any complaints go straight to Tony in Telegram. His reviews went from 47 to 214 in a few months.
Chapters
Transcript
Read the full transcript
0:00In this video, I'm going to be talking about a person who built an Open Claw agent that manages his entire cleaning business. Inquiries, quotes, booking confirmations, reminders, review quests, pretty much all his business has been automated. And he is not a technical person. He does not code at all. He barely knew what an AI agent was even 6 months ago, but he had a very clear problem and he built the simplest thing that solved. His name is Tony. He runs a residential cleaning business. About 12 cleaners on his team covering a metro area. Mostly recurring weekly and bi-weekly clients. It's a really good operation. The problem was the phone. Tony was responding to like 4 to 5 hours a day just on the phone answering inquiries from people who found him through Google, giving quotes,
0:46scheduling jobs, sending confirmation texts, following up when someone didn't show up, calling clients back when they had questions. And at the end of all those calls, he still had to do the actual job of running the business. Managing his team, handling the supplies, dealing with the complaints. He hired a pipe part-time person to handle the phones. That helped for a bit, but the person had left. And finding and training someone new for a role that was basically just answering the same questions over and over pretty much felt like a waste of time. He knew there had to be a better way and there is with AI agents. That's why I make these type of videos. I like to deep dive into the use cases that are more practical for people to understand and implement into their own business. This same type of strategy can be applied on multiple niches. That's why we're
1:32covering it. I have a link down below if you want to check out further on what he actually built out more detailed. But basically this guy hit got his daughter who is in his 20s. Basically this guy got his daughter who she is in her 20s and knows way more about AI than this guy did. She sets up an Open Claw for him and helps him and describes basically what he needs. And what they built together is what I want to walk through. The whole thing starts with a contact form website. Someone fills out the form, gives it their name, their address, what kind of cleaning they need, how big the house is, and maybe any special requests. That form submission triggers the agent. Within about 60 seconds, the person gets a text message not from a number that says unknown from Tony's business number
2:17which is routed through the agent. The message introduces itself as Tony's scheduling assistant. Thanks them for reaching out and ask a few follow-up questions if the form didn't capture everything it needed. Real quick, before we keep going, if you're watching this and you want to actually build with some of these tools, not just watch videos about them, then you're going to want to check out our community down below Shipping School. We have a full Claude Code course, a full Open Claw course, and four live bootcamps every single week where we actually help you get set up from scratch. Like actually set this thing up, not just watch a tutorial and figure it out by yourself. And we also provide one-on-one coaching so you could book a call with me. We could share screens and I can help you get Claude
3:02Code or Open Claw running on your machine. That's it. No fluff. I built this community because watching YouTube's only gets you so far. We launched it just 3 days ago and we have over 55 members. You need people around you who are actually building, people who hold you accountable, and coaches who can help you when you get stuck. I'll put the link in the description down below. Get in now before the price goes up. Most people text back within a few minutes. The agent handles the conversation naturally. It answers about what is included in the cleaning, what produces and what products they use, whether they do move out cleans, how the pricing works, all the things Tony used to answer on the phone 30 times a day. Once it has enough information, it generates a quote. And here is where the
3:49setup gets crazy. Tony spent a few hours with his daughter building what they call the quote calculator logic. Basically, it's a set of rules that the agent follows. A two-bedroom apartment up to like 600-900 square feet in this set price range, you know. Add this much for extra bathroom. Add this much if it hasn't been cleaned in over a month. Deduct this much for recurring clients. Tony had been doing this math in his head for years. They just turned it into agent instructions. The same thing can be applied to multiple niches, especially if you have domain knowledge in a specific niche. That's where you could basically train out your OA agent to handle a lot of the tasks that you're doing manually. This takes the next step further that Tony did here. The agent
4:34generates the quote. It explains what is included. It mentions the recurring discount if the person books weekly or bi-weekly. And then sends a booking link for the available slots on the calendar. Tony uses Google Calendar, nothing fancy. The agent is connected to Google Calendar and it only shows the open slots. And then it books directly and it updates the calendar automatically. From inquiry to booked appointment, no human involvement required. The whole thing can happen at 11:00 p.m. on a Sunday while Tony is asleep. After the booking is confirmed, the agent handles the follow-up sequence automatically. 2 days before the appointment, the client gets a confirmation text. The day before, they get a reminder with the approximate arrival time. The morning of the
5:19appointment, they get a final reminder with the cleaner's name and a note about any details Tony's team needs to know about like a gate code or a dog to watch out for. That sequence replaced probably about 30 minutes of manual outreach per job. Tony has 40 jobs per week. That's about 20 hours of texting every week that just vanished. And then after each job, the agent sends a follow-up message. Thanks the client, ask if everything looked great, and if they give a positive reply, it asks them to leave a Google review and sends them a direct link. If they mention something that was not perfect, it flags it to Tony in Telegram immediately so he can call them before it comes a complaint. His Google reviews went from 47 to 214
6:04in just a few months since they turned on that follow-up agent. You know, this is exactly the kind of thing that drives Google Maps ranking and brings in more organic traffic. Those jobs compound. He also built a simple re-engagement agent that runs once a month. It looks at clients who had like a one-time clean and have not booked again in 45 days or more. It sends them a message asking if they are ready to schedule their next clean and mentions the recurring discount. Tony said this one agent brought back six clients in the first month alone, which at his prices was about $900 in recovered revenue from one automated message. I want to talk about the results here because this is cool in theory, but does it actually work? Tony went from spending 4 to 5 hours a day on the phone
6:51to spending 45 minutes reviewing what the agent flagged for his attention. His bookings went from about 30% because of the response time dropped from hours to under a minute. People searching for cleaning services shopping multiple options at once, the first business to respond professionally almost always wins the job. He is now always first. His Google rating went from 4.1 to 4.8 because of the review collection agent in just a few months. He recovered recurring clients he would have otherwise lost. And his part-time phone person, when they eventually found someone new, shifted to a completely different role doing quality control and client relationship management instead of just answering the same question on repeat. Here's what I think is the most important thing about Tony's story. He
7:36is not a tech person. He did not learn to code. He did not build anything complicated. He sat down with his daughter and described his business in plain language. What happens when someone calls, what questions when they ask, how prices work, what his team needs to know, what makes the good job versus a bad one. And that description became the agent essentially. That is what I keep saying. If you're struggling to keep up with content, to save you about 40 days worth of work. I built [music] something called Content Machine. It's 10 AI agents that run on the Open Claw orchestration and they handle everything. Scripts, thumbnails, X posts, [music] blogs, outreach, clips, newsletters, all of it. So, I went from 1,000 subscribers to 4,000 subscribers
8:22on YouTube in 7 days using this exact system. Every single morning I wake up and the content's already done. I spend maybe 15-20 minutes reviewing and approving them and I move on with my day. It works for [music] any niche. Fitness, finance, real estate, marketing, whatever you are building and it is 100% completely customizable to your use case. >> [music] >> So, you get the mission control dashboard, all of the cron jobs, everything I've built over the last 40 days helping me gain more and more people to subscribe and join the community. So, you plug in your own thing and it molds it to you. It learns how you talk and it writes so it doesn't sound like AI slop. [music] $97 one-time. It's not a subscription. I'll put the link down below and you'll thank
9:07me later. Open Claw is not a developer tool. It's a business tool. The people building the most useful things on it are the ones who understand their business operations the best. They do not need to learn code at all. You don't need to understand code. You don't even need to touch it. You don't need to look at it. You need to review it. You just need to be an operator in your business and know the ins and outs. You can train an agent to replace 80% of the workflow that you do manually. The redundant tasks. Tony knows his business. That knowledge is what makes his setup work. This pattern applies to any service business. You have a lawn care business? Dial it up. Get an agent going. You have an HVAC business? Do it. You have a roofing business, a dog grooming business, tutoring, bookkeeping, whatever. If your business involves
9:52receiving inquiries, giving quotes, or scheduling appointments, or any type of follow-up, you can build this. The specifics are different, but the structure is the same. Inquiry comes in, agent responds, agent qualifies, agent gives the quote, and then it books the appointment. Agent then confirms that appointment, and then the agent follows up. That is the whole pipeline. Tony built it in about a week with his daughter's help. If you run a service business, or know someone who does, this is worth paying attention to. The barrier to entry has basically disappeared. You do not need a developer, you do not need a software budget. You would need to describe how your business works, and let the agent run the repeatable parts. We have a community based on taking your next step with AI, building these same exact workflows. We do nine live calls every single week, covering all various topics
10:39from Open Claw, getting Claw code installed, and building your first app. So, I'll put the link to it down below. I launched the community 26 days ago, and we have 219 members. And we focus on human-to-human interaction, and helping you take the next step in your AI journey. You know, we have people in there building setups of all kinds around services businesses, from real estate to finance to fitness. And the live calls is where you can bring
Ready to start building with AI?
Join the Shipping Skool community free. Paid members get live bootcamp calls every week and weekly-updated courses.
Join Shipping SkoolKeep reading
This Restaurant Cut Response Times From 4 Hours to 2 Minutes With One AI Agent
A restaurant was taking four hours to respond to customer messages. They deployed one OpenClaw agent. Response time dropped to under two minutes. 80% of inquiries now handled automatically.
OpenClaw for Bookkeeping and Taxes: Automate Invoices With AI
An OpenClaw agent can handle your bookkeeping intake: it watches an inbox, parses invoice PDFs, pulls out amounts, dates, and vendors, and organizes everything for your accountant for roughly $10 to $20 a month in API usage. You still need a human review step and a CPA for final decisions.
How to Build an AI Product Photography Workflow in OpenClaw Without Spending a Month Coding
A builder spent a month hand coding an AI product photography workflow, then rebuilt it in OpenClaw in two days. Here is how the workflow works, why the speed gap matters, and how to turn it into a real service.
