A missed call in the trades is not a missed call. It is a missed install. One Dallas HVAC operator told Fortune in April that a single missed phone call can be a $30,000 to $40,000 system replacement evaporating in real time. The homeowner with a dead compressor in July does not leave a voicemail and wait. They hang up and dial the next company on the list.
Most trades shops answer 60 to 70 percent of inbound calls during staffed hours. After 5 p.m., on weekends, and through the lunch rush, that rate falls off a cliff. Voice AI vendors now report booking rates between 70 and 90 percent on the calls they pick up. The gap between your after-hours answer rate and theirs is revenue you hand a competitor every night. Call answer rate is the one field metric to fix before June, and the tools to fix it got cheaper and better over the past year.
Count Your After-Hours Misses Before You Buy Anything
Do not buy software to solve a problem you have not sized. Before you talk to a single vendor, pull 30 days of call data from your phone provider or your CRM. Every major VoIP system and every field service platform logs call time, duration, and outcome. Sort for calls that arrived outside your staffed hours and went to voicemail or rang out with no answer. That list is your leak.
Now put a dollar figure on it. Take your missed after-hours calls, multiply by your historical book rate on answered calls, then multiply by your average ticket. A plumbing shop missing eight calls a weekend at a 50 percent book rate and a $480 average ticket leaves roughly $1,900 on the table every weekend. That is close to $100,000 a year. An HVAC shop in replacement season can run that math into six figures before August.
That number does two jobs. It sets your budget ceiling, because no answering tool is worth more than the revenue it recovers. And it sets your urgency, because a six-figure leak is not a next-quarter project. Write the figure down. You will use it again when a vendor quotes you a monthly price.
The Built-In Option Is Cheap and Already in Your Stack
If you run ServiceTitan, Jobber, or Housecall Pro, you may already own an after-hours answering tool and not know it. All three shipped voice features over the past year, and the pricing spread is wide enough to matter.
Housecall Pro folded HCP Assist into its Essentials and MAX plans at no add-on cost. It answers calls, books appointments, and writes customer details straight into the account. Jobber sells its AI Receptionist as a $99-per-month add-on and bundles it free on the $599 Plus plan. ServiceTitan routes after-hours calls through Voice Agents inside its Titan Intelligence layer, paired with Dispatch Pro, which scores jobs by predicted revenue and assigns the tech with the right skill and drive time.
The built-in route works when call volume is moderate and trade questions are simple. It books the easy jobs: a tune-up, a drain clear, a quote request. It will not hold a careful conversation with a panicked customer describing a gas smell, and it will not run outbound follow-up on the ones it cannot close. For a 3 to 8 truck shop testing the idea, that tradeoff is fine. You are capturing calls you were losing outright, using a tool already on the invoice.
The Dedicated Layer: Avoca and the Standalone Front Office
When call volume climbs or conversations get complex, operators move to a dedicated voice AI front office. Avoca is the loudest name in that category right now. The company closed a $125 million Series B on April 27, 2026, reached a $1 billion valuation, and serves more than 1,000 service businesses. Its platform splits into three parts. Convert answers inbound calls, texts, and chats around the clock. Nurture runs outbound follow-up with speed-to-lead under 15 seconds. Coach scores recorded calls and trains your human CSRs from the transcripts.
Pricing is custom, but third-party reviews place a mid-market HVAC operation in the $1,500 to $2,500 per month range, depending on call volume and which of the three parts you switch on. Standalone contractor answering services built on newer voice models, including vendors like OnCrew, sit lower and often bill per call or per minute. Industry figures show these agents booking around 90 percent of calls when schedule capacity is open, and about 70 percent once you count jobs declined purely for scheduling reasons.
The dedicated layer earns its price when after-hours and overflow volume is high enough that the built-in tool would drop real money. Hold the monthly fee against the dollar figure you calculated earlier. If your missed-call math says $8,000 a month and Avoca quotes $2,000, the decision is not close. The Coach feature is a quiet bonus here: the same transcripts that train the AI also show you exactly where your human team fumbles live calls.
What Voice AI Will Not Fix
Be honest with yourself about the limits before you sign anything. A voice agent books the call. It does not get a truck to the house. That 90 percent booking rate drops to 70 percent the moment your schedule is full, and no software closes that gap. If you book an emergency call at 11 p.m. and the soonest slot is six days out, you have a worse problem than a missed call: a booked customer who now feels lied to.
Voice AI also does not fix a weak callback habit, a dispatcher who buries leads, or a tech who quotes inconsistently. It surfaces those problems faster, because every call is now recorded and transcribed. Treat the rollout as a measurement tool as much as an answering tool. If the transcripts show you booking jobs you cannot staff, the next hire is a tech, not more software.
A Two-Week Rollout That Will Not Burn Your Brand
The fastest way to sour customers on voice AI is to point it at your main line on day one and let it fumble a real emergency. Stage it instead, and keep the test small enough to read.
Week one: route only after-hours and weekend calls to the agent. Those calls go to voicemail today, so any booking is pure upside and no live caller gets a worse experience than they already get. Write your trade-specific intake questions before you switch it on. An HVAC script needs system age, the symptom, and whether the house has no heat or no cooling. A plumbing script needs water shut-off status and whether the leak is active. Build one hard escalation rule: any mention of gas, sewage backup, or active flooding triggers an immediate text and call to the on-call tech.
Week two: read every transcript from week one. Listen to the calls the agent lost. You will find two or three script gaps, and closing each one is a five-minute settings change. Once the after-hours numbers hold steady, decide whether to extend the agent to lunch-hour coverage and hold-time overflow during staffed hours. Keep your dispatchers on the calls that need a human voice. Let the agent absorb the volume they could never reach anyway.
Track one metric through the whole test: booked jobs from calls you used to miss. Not call count, not minutes, not transcript volume. Booked jobs and the revenue attached to them. That is the only figure that tells you whether the tool paid for itself, and it is the figure to report when you decide whether to keep it.
Here is the decision to make before June. Pull your last 30 days of call logs this week and put a dollar figure on your after-hours misses. If that number sits under a few hundred dollars a month, switch on the built-in tool in the platform you already pay for and move on with your day. If it runs into the thousands, book demos with Avoca and one standalone answering vendor, then start the two-week staged rollout. The phone is still where trades jobs are won or lost, and for most shops it is ringing into the dark every night after five.
At Atlas Unchained we build the AI front office for operators who would rather book the job than chase the callback. If you want help sizing the missed-call math for your shop, or deciding between the built-in and dedicated routes, that is the work we do. Subscribe to The AU Brief for the Friday digest of what actually moved for operators this week.
