A homeowner with water coming through the ceiling does not leave a voicemail and wait. They call the next plumber on the list, and the one after that, until someone picks up. Whoever answers usually gets the job.
That is why the phone is still the most important piece of software in a trades business, and why "should I get an AI receptionist?" has become one of the most common questions I hear from plumbing, HVAC, electrical and roofing owners. The honest answer is: it depends on how many calls you are missing, what those calls are worth, and how complicated your calls are. This post walks through the options side by side so you can make the call yourself.
First, find out how many calls you are actually missing
Most owners guess low. The calls you miss never show up on an invoice, so they are easy to forget.
Industry data gives a rough range, but it swings a lot depending on how it is measured. CallRail's 2025 benchmark report, based on 1.1 million leads, put the missed call rate for home services at 14%, and noted that up to 85% of people whose call goes unanswered will not call back. Invoca's 2026 home services benchmarks found that only 52% of inbound calls were answered by a person, with answer rates across sub-industries ranging from 32% to 74%. The two numbers count different things, but both point the same way: a meaningful share of people who try to call you never reach you.
Your own number is what matters. Before you buy anything, pull these from your phone system or call tracking for the last 30 days:
- Total inbound calls
- Calls that went to voicemail or rang out
- How many of those came after hours, on weekends, or while the crew was on a job
- How many missed callers you called back, and how many of those you booked
- Your average job value for a typical first visit
If your phone provider can't show missed calls, a basic call tracking number on your website and Google Business Profile will, usually within a couple of weeks.
The four realistic options
Almost every trades business ends up with one of these, or a mix.
1. Voicemail and call back
Free, and fine if your volume is low and your customers are mostly repeat clients who will wait. It falls apart with new customers and urgent jobs, because they are calling around. A voicemail greeting that promises a callback "within the hour" helps a little, but only if someone actually keeps that promise at 9pm on a Friday.
2. Owner or technician on call
Calls forward to whoever is on rotation. Callers get a real expert, which is great for emergencies. The cost is your team's evenings, a phone that rings during jobs, and calls answered from a noisy van by someone who can't check the schedule. It also doesn't scale: two busy weekends in a row and people start letting it ring.
3. Human answering service or virtual receptionist
Trained operators answer in your company name, take details, and follow a script you agree on. Some can book into your calendar. Quality depends heavily on the provider and on how well you brief them.
Cost is usually per minute or per call. OnceHub's 2026 pricing comparison puts typical small-business answering service plans at $175 to $700 a month, with per-minute overage on published plans running from about $1.75 to $5.40 depending on the provider. Busy months get expensive quickly.
4. AI receptionist (voice agent)
Software answers every call instantly, in a natural voice, any time of day. A well-set-up one can answer common questions (service area, hours, whether you handle a certain job), collect the caller's details and problem, book a slot, and text your team a summary. The same OnceHub comparison lists AI receptionist tools from roughly $25 to $300 a month, with most full-featured options between $50 and $200.
The weak spots are real. AI handles routine calls well but can struggle with a caller who is upset, rambling, or describing an unusual problem. It is only as good as the information and rules you give it. And it needs a clear path to a human for anything it shouldn't handle.
Side-by-side comparison
| Voicemail | On-call staff | Answering service | AI receptionist | |
|---|---|---|---|---|
| Typical monthly cost | Free | Staff time, possible on-call pay | About $175-$700+ (per minute/call) | About $25-$300 (flat or usage-based) |
| Answers instantly 24/7 | No | Usually, not always | Yes, may queue at peak times | Yes, handles many calls at once |
| Handles complex or emotional calls | No | Best | Good, depends on operator | Weakest |
| Books directly into your schedule | No | Sometimes | Some providers | Yes, when integrated |
| Setup effort | Minutes | Low | A script and briefing | Knowledge base, rules, testing |
| Cost when call volume spikes | None | Burnout | Rises with minutes | Usually small |
Prices above are ranges reported in late 2026 and change often, so confirm current pricing with any provider before you budget.
A worked example (hypothetical numbers)
To make this concrete, here is a made-up HVAC company. These are not real client figures, just round numbers to show the math. Plug in your own.
- 300 inbound calls a month
- 15% missed, mostly evenings and while techs are on jobs: 45 calls
- Half of those are genuine new jobs (the rest are spam, suppliers, existing customers): about 22
- If the business would normally book 1 in 3 of those, that is roughly 7 lost jobs a month
- At a $350 average first-visit ticket, that is about $2,450 a month in work going to competitors, before counting any repeat business
Against that, a $150 a month AI receptionist or a $400 a month answering service both look cheap, as long as they actually convert a decent share of those calls into booked jobs. If your own math shows only one or two missed jobs a month, a better voicemail and a strict callback habit may be enough.
What separates a good AI setup from a frustrating one
Most complaints I hear about AI phone agents come from rushed setups, not the technology itself. Before you switch one on, make sure it has:
- Your real service area by town or ZIP code, so it doesn't book a job 90 minutes away
- A clear list of what you do and don't do, including brands, job types, and minimum call-out fees if you publish them
- Emergency rules: which words trigger an immediate call or text to your on-call tech, and any safety line you want it to give (for example, telling callers who smell gas to contact the utility or emergency services)
- A human handoff for upset callers, existing jobs, and anything outside the script
- Honesty up front: a short line saying it is the company's automated assistant
- Booking rules: job lengths, buffer time, and which slots are off limits
- A summary to your team by text or email after every call, with the caller's number and recording
Then test it the way a customer would. Call it from your mobile with a vague problem, a rude tone, a job you don't do, and an address outside your area. Listen to the first 50 real call recordings and fix what it gets wrong. That first week of tuning makes most of the difference.
One more thing worth fixing while you are at it: Invoca's report found that 55% of home services businesses don't ask callers to buy or book the job. Whatever answers your phone, human or AI, the script should end with a clear offer of a time slot.
Calls aren't the only leads you miss
The same logic applies to your website forms, Google Business Profile messages, and Facebook inquiries. A form submission at 8pm that gets a reply the next afternoon is a missed call by another name. The same automation that answers your phone can send an instant text reply to form leads, offer booking times, and log everything in one place. Estate agencies feel this most: a 2026 Pied Piper study of large US real estate brands found 68% of website enquiries got no reply within 24 hours, which is why fast follow-up sits at the centre of our work for real estate agencies. Law firms are no different: when Clio had a researcher pose as a potential client and contact 500 US law firms for its 2024 Legal Trends Report, only 33% replied to the email, which is why enquiry handling is part of everything we do for independent law firms.
It also matters for your ad spend. If you run Google Ads or Local Services Ads, every unanswered call is a lead you already paid for. Fixing call handling often improves your cost per job more than any change to the ads themselves, which is why we look at it early in any paid advertising engagement. And if customers are increasingly finding you through AI tools, our guide on getting your business recommended by ChatGPT covers the other end of that funnel. You can also see how we put the website, local SEO and call handling together for plumbers, HVAC companies and other home service businesses.
A simple decision guide
- Fewer than 5 missed calls a month: improve your voicemail greeting, set up missed-call text-back, and call back within the hour.
- Lots of after-hours and on-the-job misses, mostly routine calls: an AI receptionist is usually the best value.
- Complex, high-value or sensitive calls (large installs, insurance work, unhappy customers): a human answering service, or AI that hands these straight to a person.
- High volume and a mix of both: AI answers first, books the simple jobs, and routes the rest to your team or a human service.
Whatever you choose, measure it. Compare booked jobs from phone leads for the month before and the month after. If the numbers don't move, the problem may be your script, your pricing, or your website sending the wrong people to call.
If you'd like help working out your missed-call numbers or setting up an assistant that books into your existing tools, that's the kind of AI automation work we do. Book a short call and we'll look at your setup together, no pressure either way.