The Missed-Call Playbook: What Every Service Business Should Automate This Year

Most home-services owners I talk to are losing more business to missed calls than to bad marketing, slow websites, or competitor underbids combined. The math is brutal once you sit down with it. A typical local-services business missing 20% of inbound calls, with a $400 average ticket and a 35% close rate on calls that do get answered, is leaving roughly $28,000 a year on the table per 1,000 inbound calls. Most owners have never run that calculation. The phone just rings, or it doesn't, and the rest of the day moves on.

The good news is this is one of the most fixable problems in the entire small-business stack. The infrastructure is cheap, the workflows are simple, and almost every missed call can be turned into either a recovered booking or a high-quality lead that goes into your CRM the same day. This post is the playbook we use for TruLight and the same one we set up for Front Door Digital clients during a build.

The actual cost of a missed call

Quick math, because most owners haven't done it. Take your monthly inbound call count (from your phone system, your call-tracking number, or even a rough estimate). Multiply by your missed-call rate. Multiply by your close rate on answered calls. Multiply by your average ticket. That's the revenue you're losing per month, in the most direct possible terms.

For a typical local-services business: 200 inbound calls a month, 20% missed (40 calls), 35% would have closed (14 jobs), $400 average ticket. That's $5,600 a month, $67,200 a year. Every category of home services I've seen runs in this range or worse. Plumbing and HVAC missed-call costs tend to skew higher because emergency tickets average closer to $700. Lawn care and pest control skew lower per ticket but higher in call volume.

The reasons calls get missed are mostly mechanical. The owner is on a job. The office is closed. Two calls come in at once. The dispatcher steps away. The phone is on vibrate and nobody hears it. None of these are character failures. They're just unsolved process problems.

The three layers of the playbook

The system that actually works has three layers, deployed in order. Each one catches calls the previous one didn't.

1. Layer one: real-time AI answering. Every inbound call is answered within two rings by an AI receptionist trained on your business. It books appointments, captures lead data, and escalates the small share of calls that genuinely need a human.
2. Layer two: instant SMS auto-reply. For the rare call that slips past layer one (mid-conversation drop, hung-up-after-one-ring), an automated text lands on the caller's phone within sixty seconds.
3. Layer three: structured return-call SOP. A documented protocol for the calls the AI escalates and the SMS replies that need a human follow-up.

The net effect: what was previously a black hole becomes either an immediate booking or a fully tracked lead that doesn't get lost. Layer one handles the bulk of the volume. Layer two recovers most of what slips through. Layer three is for the small share of calls that genuinely need a human voice.

Layer one: real-time AI answering

This is the highest-leverage piece, and it's the one most owners haven't tried yet. The idea is straightforward: an AI receptionist picks up every inbound call within two rings, twenty-four hours a day. It speaks naturally, knows your services and your pricing, takes the address and the problem, and either books the appointment directly into your calendar or escalates to a human when the call genuinely needs one.

The reason this beats every other layer is friction. A caller dialing a plumber at 11pm doesn't want to leave a voicemail and doesn't want to text. They want to hear another voice on the line telling them help is on the way. Real-time AI is the closest thing to that experience that scales below the cost of staffing a 24/7 dispatch team. The text-back layer below catches the few calls that slip through, but the goal of layer one is that very few calls slip in the first place.

Three things that make this work:

• The AI sounds like a person. Old phone-tree IVR systems trained a generation of callers to hang up the second they hear "press one for". Modern voice AI is conversational, listens, and adapts. Callers who'd hang up on an IVR will talk to a good AI receptionist all the way through to booking.
• It knows your business. The AI is trained on your services, your pricing, your service area, and your scheduling rules. It doesn't take generic messages the way a live answering service does. It captures the data your dispatcher actually needs.
• It escalates cleanly. For calls that genuinely need a human (a complex repair quote, an upset existing customer, a press inquiry), the AI hands off to a phone number or a Slack message with the full transcript and a summary already in hand.

This is the layer we ship as Front Door Digital's AI Receptionist. $999 one-time setup, $199 a month, live in about a week. Typically less than half the cost of a live answering service, and it converts substantially better because the AI doesn't disappear after fifteen seconds of pleasantries.

Layer two: SMS auto-reply for the calls AI didn't catch

Even with real-time AI answering, a small share of calls won't get through. The caller hung up after one ring before the AI picked up. The internet connection at the truck dropped a call mid-conversation. A spam-flagged number routed to voicemail before AI could greet it. For these calls, SMS auto-reply is the safety net.

The mechanic: when a call rings through and isn't connected within four rings, an automated SMS lands on the caller's phone within sixty seconds. The text doesn't pretend to be human. It tells the caller exactly what's happening and gives them a way to keep the conversation going without sitting on hold.

The text I use as a fallback at TruLight reads roughly: "Hi, this is TruLight SLC. We missed your call. Reply here with what you're looking for and we'll text you back within 15 minutes during business hours. For urgent issues, call again and we'll prioritize."

Three things that make this work:

• Speed. The text has to land within sixty seconds, ideally faster. After two minutes, the prospect has already moved on.
• Honesty. Don't pretend a human is going to call back immediately. Owners overestimate how much customers expect instant phone response and underestimate how much they appreciate honest text communication.
• A clear next step. Reply with what you need. That's it. The lower the friction, the higher the response rate.

Platforms that handle this layer well: GoHighLevel, Twilio with a simple webhook, RingCentral with its built-in SMS feature, or any modern voice-over-IP system with an SMS module. Cost is typically $20 to $100 a month.

If the caller still leaves a voicemail despite the SMS, transcription matters. RingCentral, OpenPhone, Google Voice, and GoHighLevel all transcribe voicemail to text and forward it to email or SMS. The transcription is good enough to triage a message in five seconds: emergency or non-emergency, lead or existing customer, address mentioned or not.

Layer three: the structured return-call SOP

The AI handles the booking-ready calls. The SMS catches the voicemail-leavers. The structured return-call SOP is for the conversations that need a human, every time: complex quotes, upset customers, the small share of calls the AI flagged for escalation. This is the human layer, and it's where most systems break down without a written protocol.

The basic SOP looks like this:

1. Every missed call and voicemail creates a task in the CRM with a timestamp. Automated, no human input.
2. The task is assigned to a specific person on duty. Not "the team." A named individual with a return-call queue.
3. The window is 15 minutes during business hours, one hour outside business hours. If the assigned person can't make the call, the task escalates to the owner or office manager.
4. Every return call gets a disposition logged in the CRM. Booked, quoted, not interested, no answer, follow-up scheduled. No call ends in a void.
5. "No answer" return calls trigger a second attempt within four hours. If still no answer after that, an SMS goes out: "Hi, this is [name] from [business] returning your call. Reply here or call back at your convenience."

The reason this works is that it removes the cognitive load on the team. Nobody is deciding in the moment whether a missed call is worth chasing. The system says it is, the timer starts, and the protocol runs. Owners who deploy this typically see their missed-call-to-booking conversion rate go from 5 to 10% to 35 to 50%. That's the difference between losing the call and recovering it.

The compounding benefit

The missed-call system doesn't just recover the calls it catches today. It also generates better data over time. Every missed call gets logged, tagged, and analyzed. You can see your missed-call rate by day of week, by time of day, by service category, by source (Google ads vs organic vs direct). Patterns emerge that tell you where to staff up, where to add hours, where to invest in better call-handling training.

For TruLight specifically, the data showed our missed-call rate was concentrated in two windows: Saturday mornings (when most leads come in but nobody's by the phone) and weekday lunchtime (when the office is briefly empty). Knowing this lets us either staff those windows or accept them and lean harder on the SMS auto-reply during those periods. Either way, the decision is informed instead of accidental.

Frequently asked questions

Will customers actually respond to an automated text?

Yes, in large numbers. The data on this is consistent across home-services categories. Roughly 50 to 60% of recipients reply to a well-written missed-call SMS, compared to roughly 10 to 15% return rates on voicemails and 0% on calls that simply disappear. The reason is friction: replying to a text takes five seconds, returning a phone call takes commitment, and ignoring a voicemail is easy.

What if I'm worried the auto-reply makes us look like a chain?

Tone matters more than format. A text that says "Hi, this is Tom from TruLight, we missed your call, what can I help with?" reads as personal even though it's automated. A text that says "Thank you for contacting us. A representative will be with you shortly. Reference number 4892" reads as corporate. Write the auto-reply the way you'd talk to a neighbor.

How much does the full system cost to run?

The AI Receptionist runs $199 a month with a $999 one-time setup. The SMS auto-reply layer adds $20 to $100 a month depending on your phone platform. Voicemail-to-text is usually included in any modern phone system at no extra cost. The recovered revenue from even one extra booking a month covers the whole stack many times over.

Can I just route everything to an answering service instead?

You can, but the economics rarely work. A live answering service costs $150 to $500 a month and the operators don't know your business, your pricing, or your scheduling. They tend to take messages rather than convert leads. An AI receptionist costs $199 a month, is trained on your business specifically, and books appointments directly into your calendar instead of dumping a message in your inbox. For most home-services operators, AI plus the SMS fallback is cheaper and converts better than either a live answering service or an SMS-only stack.

The phone is going to ring whether you're ready for the call or not. The question is what happens in the sixty seconds after it stops ringing. Build that part of your business with the same care you'd put into the rest, and you'll stop bleeding money you didn't even know you were losing.