AI lead follow up: what to automate and what to keep
How does AI lead follow up work after a quote request?
A written quote request is a warm lead that cools fast. The customer filled out a form and may already be asking the next company on their list. Whoever sends a real price first gives them a reason to stop looking.
The research on this is older but clear. A 2011 Harvard Business Review study sent a web lead to US companies. The ones that did reply took 42 hours on average, and 23% never replied at all. The same study found that a company reaching out within the hour was about 7x more likely to qualify the lead than one that waited even one hour more.
So the job has two halves. The first is the reply: a price and a booking link, sent while the customer is still looking at the screen. The second is everything after it, and on a full day that's the half that slips. You're on a ladder, the customer goes quiet, and nobody sends the second message.
AI does both halves well because both are repetitive. It quotes from your own rules, sends the booking link and keeps following up until the lead books. Aim for a reply inside the first minute. A nudge that's due while you're on a ladder still goes out.
Say a gutter cleaning request lands at 7:40 on a Friday evening, and your average gutter job is $180. If you get twenty written requests per month and three book elsewhere because your reply came the next afternoon, that's three times $180, which comes to $540 a month. Add two more that went quiet because nobody sent a second message. Three plus two is five jobs, and five times $180 comes to $900 a month, or $10,800 over twelve months.
Swap in your own job price and request count in the ROI calculator. Gutters are only the example. The same math fits a window cleaning route or a spring mulch install.
How many follow-ups should you send after a quote request?
Here's a sample cadence for a written quote request. It's one reply and three nudges, then it stops. A good AI lead follow up system runs some version of it.
| Touch | When it goes out | What it carries | What stops it |
|---|---|---|---|
| Reply | Right away | The price and the booking link | The lead books |
| Nudge one | About an hour later | A short check: any questions about the job? | The lead books or replies |
| Nudge two | The next day | One line about an open day on your schedule | The lead books or replies |
| Nudge three | Day three | A last note: book from the link, or say no and it closes | The lead books, replies or says no |
| Close | After nudge three | Mark the lead closed and record how it ended | Nothing more goes out |
Three nudges is a rule of thumb, not a law. A customer who hasn't answered the reply and all three nudges has probably bought elsewhere or put the job off. Another message starts to read as pressure, and your name is on it.
Stopping matters as much as timing. A nudge sent to someone who already booked is worse than silence. Before any message goes out, the system should check one thing: did the lead book, reply or say no?
Here's a copyable set for the three nudges. Swap in the customer's name, the job and an open day.
Nudge one: Hi NAME, your quote for JOB is ready and the booking link is open. Anything about the job you want to check first?
Nudge two: Hi NAME, room opened up on DAY for JOB. The link holds your spot if you want it.
Nudge three: Hi NAME, last note on your quote for JOB. Book from the link, or reply no and the request closes.
Keep each message to two sentences: one question or offer, one link. If a message needs a paragraph, it's a conversation, and conversations are yours.
What should AI handle, and what should you keep?
Run each job through three questions, in this order. Is the message the same for every lead, is a wrong answer cheap to fix, and does the customer need to hear from you personally? Automate it only on yes, yes, no. Anything else stays with you.
Here's a sample split for a lawn or exterior cleaning business:
| Job | Who handles it | Why |
|---|---|---|
| First reply with price and booking link | AI, from your rules | Same every time, and the first real number gets the lead's attention |
| The three nudges | AI, until the lead books or says no | Repeatable, and the first thing skipped on a busy day |
| A job outside your rules, like a steep bank or a four-acre lot | You | It needs your eyes and your judgment |
| Discounts and price exceptions | You | A discount is a decision about your margin |
| A reply with a complaint or a question the rules can't answer | You | The customer needs a person |
| A big job, like a full exterior wash on a large home | You | A larger price earns a personal word |
Screening comes before quoting. If a lead doesn't fit your rules, it shouldn't get a price at all. It should wait for you. That one habit keeps a bad quote from going out under your name at nine at night.
Workely works this way. It prices each lead from your own rules and gets the quote ready with a link to book. You approve it, or turn on auto-send and it goes out on its own. Leads that don't fit your rules wait for your OK, and follow-ups go out automatically until the lead books.
For the detail, see how follow-up and records work and how lead capture and pricing work. For the wider list of what to hand off first, read AI for a home service business: what to automate first.
Who is responsible when AI makes a mistake?
You are. The customer hired your business, so a wrong price or a clumsy message is yours, whether you typed it or a tool did. Set it up with limits and a mistake stays cheap.
Three limits keep mistakes small. The rules are yours, so a price can only be as wrong as a rule you wrote. Anything outside the rules waits for your OK. And for the first week, you approve every quote before it goes out.
Treat AI lead follow up like a new helper in week one: check the work, then give it more room. When a mistake happens, fix the rule behind it. If a quote came in low on a long driveway, add a driveway rule so the next one is right.
How do you set up automated lead follow up for your crew?
Do it in a week, one job type at a time.
- Pick the job you quote most, like a house wash or a weekly mow, and write down how you price it: size, access and add-ons.
- List what falls outside those rules, such as a steep bank, a commercial lot or a rush job. Those go to you.
- Write the reply and the three nudges in your own voice, using the template above.
- For the first week, approve every quote before it goes out, and check the first few nudges that follow.
- Turn on auto-send for jobs that fit your rules, and keep approving the rest.
- Each Friday, look at how the closed leads ended: booked, said no, or went quiet. Change one rule or one message, not five.
Your next step: write your pricing rules down
Pick the one job you quote most and write its pricing rules down tonight. Good AI lead follow up starts with your rules, not with the software.
When the rules are written, see whether the reply and the three nudges can run from them. Workely is $149 a month with a 30-day free trial, billed monthly with no contract, and we set it up with you. Start at Workely pricing.