AI Follow-Up Sequence for Service Businesses
A practical 24-hour, 72-hour, and 7-day follow-up workflow so service business leads hear back fast — without the owner tracking every touch manually.
If you run a service business — HVAC, law firm, salon, dental practice, landscaping — you know that most booked jobs don’t come from the first call. They come from the second or third touch. The lead called, you answered (or didn’t), and then nothing happened for three days until they booked with someone else.
That lag is the whole problem. Not the ads. Not the website. The follow-up gap.
Most owners solve this by trying to remember who they need to call back. That works until the schedule gets busy, which is when the leads are also arriving fastest. The follow-up system I describe here removes that memory dependency entirely. The AI tracks the sequence; you only step in when a lead is warm.
Short answer: A three-touch AI follow-up sequence sends an immediate first reply at contact, a check-in at 24 hours if there’s no response, a soft nudge at 72 hours, and a final low-pressure close at 7 days. Each touch is logged to your CRM automatically. You get notified when a lead replies or books — not on every outbound touch the AI sends.
Why most follow-up breaks before it starts
The default setup at most small service businesses: a lead calls, maybe leaves a voicemail. The owner sees it when they get out of a job. They mean to call back. They don’t. By Thursday, the lead has cooled, the owner doesn’t know whether this person is still interested, and they skip it to avoid the awkward “are you still looking?” call.
Even businesses with CRMs have this problem. Logging a lead is different from following up with one. Most CRMs will hold the contact forever — they won’t send the text.
The fix is a sequence that fires automatically based on lead status, not on the owner’s mental queue. Once it’s built, you set the rules once. New leads flow in, the AI handles the outbound touches, and you get notified only at decision points.
The three failure modes I see most:
- No follow-up at all. Lead comes in, gets one response (or none), goes cold.
- Manual follow-up that depends on memory. Works when it’s slow, breaks when it’s busy.
- Follow-up that’s too aggressive. Three messages in 48 hours reads as spam; the lead blocks you.
A calibrated sequence avoids all three.
The sequence map: trigger → AI → CRM → escalation
Here’s the full chain and what the AI does at each step.
Touch 0 — immediate (under 5 minutes from contact)
A lead submits a website form, texts the business number, DMs on Instagram, or calls and hits a voicemail. The AI fires an immediate acknowledgment:
- If text or DM: reply within 30 seconds, confirm you got their message, ask one qualifying question (“What’s the address?” or “What are you looking to get done?”)
- If voicemail: send an automatic SMS to the caller’s number within 2 minutes
- CRM action: create the lead record, log the contact channel and initial message
- Escalation: if the lead says “urgent” or asks to speak to someone right now, ping the owner via Telegram or SMS immediately
The goal here isn’t to book the job. It’s to prove you’re available before the lead opens another tab.
Touch 1 — 24-hour check-in
If the lead hasn’t replied or booked within 24 hours, the AI sends one follow-up. Short, not pushy:
“Hey [first name] — just circling back to see if you had any questions about [service]. Happy to answer over text or set up a quick call.”
- CRM action: log the outbound touch, update lead status to “follow-up sent”
- Escalation: none unless they reply with something outside the AI’s scripted answers
- Skip condition: if the lead replied to Touch 0, no check-in fires — the AI is already mid-conversation
Touch 2 — 72-hour soft nudge
Still no reply. The AI sends a second message, this time with a light time hook:
“I have a couple openings this week if timing works — let me know and I’ll hold one for you.”
Specifics make this work. “I have openings” is vague. “I have a Thursday afternoon and a Friday morning” is concrete. Pull available slots from your calendar if possible; if not, a general reference still outperforms silence.
- CRM action: log the touch, mark lead as “low engagement”
- Escalation: if the lead replies with any interest — even “maybe” — the AI notifies the owner immediately and pauses the sequence
Touch 3 — 7-day last touch
One final message, three to five days after the 72-hour nudge:
“I’ll leave you be after this — if your timing changes, just reply here and we’ll pick it up. Thanks for reaching out.”
This one does two things. It closes the loop cleanly so the lead doesn’t sit in “maybe” indefinitely. And it sometimes converts. People who had every intention of booking but got distracted respond to the “last one” message more often than the check-ins.
- CRM action: move lead to “dormant” or “closed/no response”
- Escalation: none — if they reply, the AI restarts with Touch 0
The full chain:
| Time from contact | Touch | Goal | CRM status |
|---|---|---|---|
| Under 5 min | Acknowledge + qualify | Prove you’re available | Lead created |
| 24 hours | Check-in | Restart conversation | Follow-up sent |
| 72 hours | Time hook | Light urgency | Low engagement |
| 7 days | Last touch | Clean close or late convert | Dormant |
What I would automate first
Start with Touch 0. That one is purely a speed problem — humans can’t reliably reply in under five minutes across every channel, all day, including evenings and weekends. AI can.
The immediate reply alone changes the conversion math. Leads that get a response within five minutes are far more likely to stay engaged than leads that wait 30 minutes. If you’re already handling Touch 0 manually and doing it well, skip ahead to Touch 2 — the 72-hour nudge is the one most businesses skip and most regret.
Forty-eight hours in, owners feel like the lead is gone and don’t want to be annoying. The AI doesn’t have that reluctance. It sends the message on the schedule you defined, with the language you approved.
The 7-day last touch is optional. Some businesses skip it for short-purchase-cycle services (same-day HVAC, emergency plumbing) and keep it for longer cycles (legal, renovation, high-ticket dental work). Match the length of the sequence to how long your buyer typically takes to decide.
The tools that connect this
Three things need to be in place before this runs:
A CRM that accepts writes. HubSpot’s free tier works. Jobber, Housecall Pro, and GlossGenius all have integration paths. If you’re running a spreadsheet, you can write to Google Sheets via Zapier or Make — not ideal, but functional for a first deployment.
A communication channel the AI controls. SMS via Twilio is the most universal option. If most of your leads come in through Instagram or Facebook, the Messenger API handles the same function. If you want all escalations and summaries routed through your phone without a new dashboard, a Telegram bot is the cleanest command center — owner approvals, voice notes, and stale-lead summaries, all from the app you already use.
A trigger that starts the clock. Website form submission, Google Business profile message, inbound SMS, voicemail transcript. The AI needs to see the event to start the timer. If you’re capturing leads from multiple sources, the setup described in AI lead generation covers how to structure the intake layer so every lead enters the same sequence regardless of where it came from.
Decision checklist before you build
Run through this before committing deployment time:
- Your CRM data is clean — no duplicate contacts, no stale records from 2023
- You know your one qualifying question for each service type
- Your average sales cycle is long enough to justify a three-touch sequence (most service businesses: yes)
- You’ve confirmed SMS compliance requirements for your industry (healthcare and legal have specific rules)
- You have someone who can handle escalated replies within 1 business hour
- Your call volume is above ~10 inbound leads/month (below that, do it manually)
If three or more of these aren’t checked yet, fix them first. The sequence amplifies what you already have — it won’t rescue a broken intake process.
When this isn’t the right move yet
Skip the automation if:
Your CRM isn’t clean. Automating follow-up on a dirty database means the AI sends messages to wrong numbers, outdated contacts, and leads that already converted six months ago. Run data hygiene first.
You haven’t settled on your qualification questions. The 24-hour check-in only works if Touch 0 gathered something useful. If your initial message asks a question the lead can’t easily answer (“What’s your budget range?”), you’ll get no replies and no data.
Your business has a compliance constraint around automated SMS. Healthcare businesses and attorneys need to verify applicable rules before running automated follow-up campaigns. This isn’t a reason to skip the sequence — it’s a reason to check first.
Your sales cycle runs longer than 14 days by design. Some consulting and high-custom-work businesses have a two-to-four week evaluation window. A 7-day sequence fires too aggressively for that buyer. Extend the windows to 3 days / 10 days / 30 days instead.
One honest note: the sequence doesn’t fix the problem if the original leads weren’t qualified. If you’re running broad ads and converting low-intent inquiries, the AI is following up on a lot of people who never seriously considered you. Garbage in, follow-up sequence out.
Next step
If leads are going cold between first contact and your first real conversation, the immediate reply and the 72-hour nudge recover a meaningful share of what you’re losing. Start with one channel — inbound SMS or website form — run it for 30 days, then expand.
If you want to see how I’d wire this for your specific business type, book a 20-minute audit and I’ll map the trigger-to-CRM flow before we talk about anything else.
