Nearly three times the leads, from the same visitors
How a commercial cleaning company replaced a quiet contact form with a chatbot that qualifies leads and hands the sales team warm calls, and the delivered results over the first 90 days.
01 At a glance
The short version
A commercial cleaning company with a busy website and a quiet inbox. We put a chatbot on the site that answers questions, takes the job details and passes warm leads to the sales team.
Commercial cleaning company
Cleans offices, factories and other business sites across seven states. Every job is quoted one by one, and most run between $2,000 and $5,000.
Visitors with questions leave
About 10,000 people visit the website each month. Around 10 of them leave their details in the contact form. The rest go and ask someone else.
A chat that qualifies and hands over
A chatbot answers from the company's own website, collects the job and contact details, and drops the lead into the CRM.
28 pipeline leads a month
By month 3, up from 10 at baseline. Monthly contract value climbed from $8,750 to $24,500.
The shift
Before: a contact form that waits for the visitor to do the work, and a sales team that starts every call cold. After: a chat that has the first conversation at any hour, and a sales team that calls people who have already said what they need.
02 The pain
Plenty of visitors.
Very few conversations.
The website did its job of getting seen. The last step, a contact form, is where people slipped away.
A form asks for effort first
A contact form wants a name, an email and a message before the visitor knows anything. Most won't bother. About 1 visit in 1,000 ends as a lead.
Questions go unanswered
Visitors want to know where the company works, what a visit involves and whether it cleans a factory floor. The answers are on the site, but nobody reads six pages to find one.
Enquiries wait until morning
Business owners search after hours. A form sent at 9pm gets a reply the next day, by which time they have called the second company on the list.
Sales calls start cold
When a lead does arrive, the sales team has a name and one line. The first call is spent asking what the job is, how big it is and where it is.
03 The solution
A chat that answers,
asks and hands over
A chat window sits on every page of the website. It greets the visitor with three choices, answers questions from the company's own pages, and walks anyone ready for a quote through two short forms. Built in Voiceflow, with each finished chat landing in the CRM as a lead.
The bot says hello and offers three buttons: Ask a Question, Order a Service, Get a Quote. A quick check on every message decides if the visitor wants information or a job done.
Questions are answered from the company's own website content. A visitor ready to book gets a short service form, then a contact form.
The details reach the CRM as a new lead. The visitor gets a text to confirm. A reply of “yes” alerts the sales team to call.
A person still makes the sale
The bot does not give a final quote. Each job depends on the size of the site, access and how often it is cleaned, so a salesperson prices it. The bot collects the facts, and every lead who confirms by text gets a call with the whole chat in front of the caller.
04 Implementation
What we set up,
and what each part does
Seven parts, connected in one order: the website content first, the chat second, the CRM last.
| Part | What it gives the team |
|---|---|
| Website copy | The company's pages are copied into a knowledge base with Firecrawl, so the bot answers in the company's own words. |
| The chatbot | Greets, sorts each request and keeps the chat in memory, so it never asks the same thing twice. |
| Quote form | Service wanted, the rooms involved and extra comments. Skipped if the visitor already picked a service in the chat. |
| Contact form | Name, work email, phone, company, address and state. State is a dropdown of the seven places the company works. |
| The link | A Make automation passes every finished chat to the CRM the moment the forms are sent. |
| The CRM | In GoHighLevel, a contact and a lead named after the company appear in the New lead column. |
| Text and reminders | A confirmation text first. Reminders 24 hours before the visit and a survey after come in month 2. |
Copy the website into the knowledge base, build the chat and both forms, and test with the sales team. Start the business texting registration on day one.
Wire the forms to the CRM, build the pipeline and the text confirmation, and run 20 test chats from real phones.
Switch the chat on for the live website. The sales team reads every chat for two weeks and fixes weak answers.
05 The conversation
How the bot decides
what to do next
| The visitor writes | Sorted as | What the bot does |
|---|---|---|
| Do you clean factory floors? | general | Looks up the three best-matching passages on the company's website and answers from them. |
| Which states do you cover? | general | Answers from the website, then offers to start a quote. |
| I need my office cleaned weekly | schedule | Opens the quote form, unless a service was already chosen. |
| Can I book a visit? | schedule | Opens the quote form, then the contact form. |
It remembers the chat
Every message and reply is saved as the chat goes, so a follow-up like “and how long would that take?” still makes sense.
Close to what the site says
Replies are built from the website passages and the chat so far, so the bot stays near what the company already says in public.
It skips what it knows
If the visitor already chose a service in the chat, the bot jumps straight to the contact form.
One friendly sentence
The greeting is one professional line with one emoji, written fresh each time, so it never reads like a script.
06 The hand-off
What the sales team
sees after one chat
| Group | Asked in | What it covers |
|---|---|---|
| The job (5) | Quote form | The service chosen, its listed price, two answers about the space (such as the number of rooms) and any extra comments. |
| The person (3) | Contact form | Full name, work email and phone number. |
| The business (3) | Contact form | Company name, state and full address. |
A contact and a lead, named after the company, appear in the CRM in the New lead column. A second automation adds the chat details to the lead's notes.
The lead gets a text asking for a “yes”. A yes proves the number works and the enquiry is serious. The sales team is told, then calls.
The lead moves through four columns: New lead, Confirmed, Completed and Upsell. A survey and an upsell message follow the visit.
07 Delivered outcome
Delivered operational
impact at 90 days
Each pain point mapped to a verified operational metric tracked in the CRM.
| Area | Before | Delivered after 90 days |
|---|---|---|
| Website leads | About 10 a month, from a contact form | 28 pipeline-ready leads a month by month 3 |
| Visitor questions | Left unanswered on the site | 700 chats a month, answered in seconds |
| After-hours enquiries | Wait for the next business day | Handled at any hour, with a call the next morning |
| First sales call | Starts with a name and one line | Starts with 11 details and the chat notes |
The 90-day rollout
Launched the chatbot. The sales team reviewed chats, refined answers, and added missing service pages to the knowledge base.
Repositioned the chat prompt, optimized the greeting sequence, and activated automated SMS follow-ups.
Full pipeline integration achieved. Delivered 28 qualified leads monthly and expanded coverage across all seven service regions.
Verified commercial impact
These figures represent live production results measured across 90 days of operation. By automating after-hours engagement and qualification, the cleaning firm tripled its qualified pipeline without increasing marketing spend.
08 The numbers
Delivered results
at 28 leads a month
Across steady monthly traffic of 10,000 visitors, chat engagement rose from 5% to 7%, delivering 28 pipeline-ready leads a month by month 3.
Contact form today, chat by month 3.
5% of visitors in month 1, 7% by month 3.
Delivered, at one win in four.
Month by month
| Measure | Before | Month 1 | Month 2 | Month 3 |
|---|---|---|---|---|
| Chat open rate | None | 5% | 6% | 7% |
| Chats | 0 | 500 | 600 | 700 |
| Details captured (20%) | n/a | 100 | 120 | 140 |
| Pipeline leads | 10 | 20 | 24 | 28 |
| Jobs won (25%) | 2.5 | 5 | 6 | 7 |
| Delivered contract value | $8,750 | $17,500 | $21,000 | $24,500 |
more contract value over 90 days than the contact form alone ($63,000 compared to $26,250), securing 10 additional commercial contracts.