Answer after-hours emergency calls
An AI receptionist picks up nights and weekends, sorts real emergencies from tune-up requests, and books or texts the on-call tech instead of sending people to voicemail.
What AI can actually do for your trade: the tools, the real cost, the hours back, and when to skip it.
An AI receptionist picks up nights and weekends, sorts real emergencies from tune-up requests, and books or texts the on-call tech instead of sending people to voicemail.
When you can't pick up, the caller gets a text in seconds asking what they need, so the conversation keeps going instead of ending at voicemail.
Load your manuals, price book and job notes into an AI assistant so techs get answers on their phone instead of calling the office.
Every estimate that doesn't close gets a steady, polite run of follow-ups by text and email until the homeowner says yes, no, or not yet.
AI sorts your past jobs by equipment age, repair history and unsold quotes, and hands you a short list of homeowners most likely to need a new system soon.
Techs dictate or jot rough notes, and AI turns them into a clear invoice summary or an options estimate a homeowner can actually read.
A review request goes out automatically when a job closes, and AI drafts a reply to every review for you to approve.
Give AI your services, your town and the season, and get a month of Facebook posts and a customer email you can edit and schedule in one sitting.
Reminders go out to plan members and past customers when their tune-up is due, so the board fills before the first heat wave or cold snap.
AI drafts answers to the routine emails (scheduling, payment questions, warranty basics) so the office or owner just reads, fixes and sends.
Paste in the week's schedule, open estimates and unpaid invoices, and get a one-page rundown of what needs you before the trucks roll.
An AI receptionist answers nights and weekends, sorts real emergencies from routine jobs, texts the on-call plumber and books the rest.
When nobody can pick up, the caller gets a text in seconds asking what they need, so the job doesn't go to the next plumber on the list.
When a job runs long, the next customer gets a clear, polite heads-up without the plumber putting down the wrench to make a call.
Load your price book, fixture specs and company rules into an AI assistant so plumbers can check them from the job instead of calling in.
Water heater, repipe and sewer line estimates get a steady run of polite follow-ups until the homeowner decides.
AI sorts your job history to find customers whose water heaters are near the end of their life, so you can reach them before the tank fails.
Plumbers jot or dictate what they did, and AI turns it into a clear invoice description or a proposal a homeowner can follow.
A review request goes out when the job closes, and AI drafts a reply to each review for you to approve.
Give AI your services, towns and the season, and get a month of short posts and a customer email to edit in one sitting.
Customers who need regular drain cleaning or sewer rodding get a reminder when they're due, so the work comes back to you instead of becoming an emergency.
AI drafts replies to the everyday emails (scheduling, invoices, warranty questions) so the office just reads, fixes and sends.
After a hailstorm, an AI receptionist answers the overflow, takes the address, damage type and insurance status, and books inspections.
When someone fills out your website form, they get a personal-sounding reply right away with a way to book, instead of waiting until someone gets off the roof.
Homeowners get a short status update at each step (materials ordered, crew scheduled, weather delay) so they stop calling the office to ask.
Order a measurement report from aerial or phone images and bid from it, instead of sending someone up a ladder for every estimate.
Photos and quick notes from the roof become a clean, organized report the homeowner can read and share with their insurance company.
Every roof estimate that isn't signed gets a short, polite follow-up run until the homeowner decides, so jobs don't slip away during storm season.
AI gathers the photos, notes and carrier emails for an insurance job into one organized file and drafts the paperwork for your estimator to check.
A review request goes out when the job wraps, and AI drafts a reply to each review for you to approve.
Give AI your services and area, and get a set of storm-season posts and a past-customer email to edit in one sitting.
AI reads the order confirmations and delay notices from your suppliers and keeps a simple list of what's coming, when, and which jobs are affected.
A smarter intake form asks the questions you'd ask on a discovery call, so poor fits drop out and good fits arrive ready to talk.
Clients get a portal request and automatic reminders for the documents you need, so staff stop sending the same email five times.
New clients fill in one form that feeds your engagement letter, organizer and systems, instead of staff typing the same details several times.
Give AI your deadlines and what's changed this year, and get reminder emails and a short client newsletter to edit in one sitting.
Clients get a short feedback request after a return or project, and the answers land in your practice management system.
AI-based bookkeeping tools categorize transactions and flag the odd ones, so staff review instead of keying everything by hand.
Scan or upload W-2s, 1099s and K-1s and let the software fill in the return, so preparers review instead of keying.
Paste the notice (with identifying details removed) and get a plain-English explanation and a draft client email to check and send.
AI rewrites the tough emails (late payments, missing documents, bad news) so they're clear and calm, and you review before sending.
A meeting assistant records client calls (with permission), then writes the summary and action items so nobody types notes afterward.
An AI research tool finds and cites the rules, and AI drafts the client-facing summary, so a day of research becomes a couple of hours.
An AI or virtual receptionist answers after hours and during court, runs your intake questions, and books consults or passes urgent matters on.
An AI chat on your website asks the screening questions, filters out matters you don't take, and sends good leads to intake with the details filled in.
After each step in a matter, the client gets a short, plain-English update, so they stop calling to ask what's happening.
Prospective clients who had a consult but didn't sign get a short, respectful follow-up run, instead of hearing nothing.
On estate plans, divorces and other flat-fee work, clients get clear requests and polite reminders for what you need, so the matter doesn't stall.
Turn the questions clients ask every week into short, plain-English website answers, drafted by AI and checked by an attorney.
At the end of a matter, satisfied clients get a simple review request, and AI drafts replies that don't reveal anything confidential.
AI drafts the routine letters, agreements and forms you write over and over, from your own templates, and an attorney reviews every word.
AI reads medical records, depositions and document sets and pulls out what matters, so attorneys and paralegals review a summary instead of every page.
Rough time entries become clear, client-friendly billing descriptions, and activity tracking catches time you'd otherwise forget.
With the client's consent, a meeting assistant records the consult and writes the summary and next steps for the file.
A former HVAC owner describes adding AI voice agents that book into ServiceTitan after hours and when the CSRs are busy.
What they did. Put an AI voice agent on after-hours and overflow calls. It schedules sales and service calls in ServiceTitan.
How it went. The first calls were rough, and it improved over the following months. He kept his human CSRs for daytime calls.
Worth knowing. He suggests piloting it in the dead hours, like midnight to 4 a.m., before trusting it with busier times.
“Initially, there were painful conversations with clients”
Bill Joplin's Air Conditioning and Heating put ServiceTitan's AI voice agent on its phones and reports most calls booked without a person.
What they did. Turned on ServiceTitan's AI voice agent for inbound calls and texts.
How it went. The company says it handled more than 1,300 calls in about three months.
Worth knowing. This is a vendor press release, and it doesn't say how many of the calls came after hours. The revenue figures in it are the company's own estimates.
Their number. More than 90% of calls booked and 72% handled with no person involved (vendor-reported).
Leonard Splaine Co. uses AI to handle estimate follow-ups and membership inquiries so its CSRs don't have to.
What they did. Picked an AI vendor (UpSmith) after shopping options at an ACCA event, and handed it routine estimate follow-ups and membership questions.
How it went. The COO reports follow-up sales on unsold estimates went up after going live. He frames it as taking work off CSRs, not replacing them.
Worth knowing. His advice is crawl, walk, run: test, expect some failures, and structure contracts so the payments match the return.
Their number. Follow-up sales on unsold estimates up 25% since going live (company-reported, no baseline given).
“There's going to be some failures with it, because it's technology.”
Bluon's equipment database is built into ServiceTitan's AI assistant so techs can pull warranty, documents and parts for the unit in front of them.
What they did. A tech scans the nameplate to see the unit's age, warranty status, documents and parts, then asks model-specific questions like how to wire a defrost board.
How it went. No measured results yet. The interview describes expected benefits, not outcomes.
Worth knowing. This is an interview with the vendor's CEO in a trade magazine, so treat it as a vendor claim. Integrations with Housecall Pro and BuildOps were announced, not live.
Tersh Blissett, who owns an HVAC company and hosts a trade podcast, uses a custom ChatGPT setup that asks about each service call and writes the invoice in plain English.
What they did. Built a custom GPT that asks questions about the service call, then writes the invoice summary at an 8th-grade reading level with little jargon.
How it went. He uses the same approach for SOPs and quote summaries, and for reviewing his P&L.
Worth knowing. He uploads documents with logos and sensitive data removed to limit the risk.
An HVAC business coach describes techs snapping photos and jotting notes, and AI turning them into a clean options estimate and a homeowner summary.
What they did. Recommends feeding job photos and rough tech notes to AI to produce an options-based estimate and a plain-English summary for the homeowner.
How it went. No client results are given; it's advice from a coach who works with contractors.
Worth knowing. Her guardrails: ground the AI in your SOPs or numbers, and keep a person on anything customer-facing at first.
“Keep a human in the loop for anything customer-facing during the first 90 days of a new workflow.”
Intelligent Design's software administrator uses ServiceTitan's AI to write review responses and approves each one before it posts.
What they did. Turned on AI-written review responses in ServiceTitan. Each reply is a little different, and he approves every one.
How it went. He says the replies are almost always right, and review volume was well ahead of the previous year.
Worth knowing. The same article notes an agency that drafts AI replies but skips one- and two-star reviews so management can handle them.
Their number. 446 reviews by early July 2024 versus 418 in all of 2023 (company-reported; not all because of AI).
“almost always spot on”
At an HVAC trade show session, an owner and an agency founder walked through the automations they run with Zapier and ChatGPT.
What they did. Automated Google review responses, posting new reviews to social media, invoice summaries, and a lead follow-up run on day 1, 3, 7 and 14 that stops when the customer replies.
How it went. Presented as working setups. The hours-saved figures in the article don't add up, so I've left them out.
Worth knowing. Drafts land in Gmail for a person to check before anything sends.
Harley's Heating & Air in Lincoln, Nebraska ran an automated campaign pushing members to book seasonal maintenance.
What they did. Ran an automated marketing campaign to plan members, and used an AI agent that booked a before-hours call and slotted it as the first job of the morning.
How it went. The campaign filled the board fast, and techs went from four calls a day to six or seven.
Worth knowing. Vendor case study. The shop also dropped plumbing, changed equipment lines and switched to arrival windows at the same time, so not all of the gain is from the software.
Their number. 250 tune-up appointments booked in three days (vendor-reported).
SD Plumbing & Heating uses an AI receptionist that looks callers up in its job software and emails the office a summary of each call.
What they did. Connected an AI voice receptionist to their existing number and job software. It logs the issue, location and urgency, and passes anything it can't handle to a person.
How it went. The vendor reports 226 calls answered in the first 11 weeks, about 90% of them outside office hours.
Worth knowing. This is the builder's own case study, with no before-and-after numbers and no figures on jobs booked.
Their number. 225 of 226 calls handled with no person needed (vendor-reported).
“It answers the phone better than we manage to on a busy day”
A Nottingham plumbing and heating firm's AI receptionist asks qualifying questions, texts emergencies to the on-call engineer and books routine jobs.
What they did. Rolled out a custom AI phone receptionist tied to their schedule over about four weeks, including a week of testing alongside staff.
How it went. After six months the builder reports every call answered and after-hours bookings going from none to 34 a month.
Worth knowing. The firm isn't named, and the numbers come from the company that built the system.
Their number. Calls answered went from 63% to 100% and call-handling cost from about £1,800 to £250 a month (vendor-reported).
Oak Creek Plumbing & Remodeling's technicians use ChatGPT to write invoices and proposals and to help troubleshoot.
What they did. Put ChatGPT on the tablets techs carry, for invoices, work proposals and troubleshooting.
How it went. The company's president told CNN it's been worth the investment and has made the company more efficient, with more time for customers and less for paperwork.
Worth knowing. No hard numbers were given. The same story cites a Housecall Pro survey that over 70% of tradespeople have tried AI tools.
A&A Roofing's CEO, quoted by the software vendor, credits an AI phone agent with catching every call after a storm.
What they did. Used an AI agent to answer overflow and storm calls, capturing name, address, damage type, insurance status and urgency, and booking inspections.
How it went. The CEO says they didn't miss a call when the storm hit.
Worth knowing. This is a testimonial on the vendor's own page, with no numbers.
“The AI Agent handled the surge and we captured every opportunity, without adding a single person to the team.”
Knights Roofing in Edmonton bids from aerial measurement reports on about 250 jobs a year.
What they did. Switched to bidding from EagleView reports instead of sending staff to every home first.
How it went. The company says estimating costs about half what it did 10 years ago.
Worth knowing. Vendor case study. The 10-year comparison includes other changes, not just the reports.
Their number. $60,000 to $80,000 a year saved on insurance, vehicles and labor (vendor-reported).
Infinity Roofing's owner told Roofing Contractor that aerial reports changed who he hires and how fast proposals get done.
What they did. Used aerial measurement reports instead of having salespeople climb roofs to measure.
How it went. He says proposals are 90% done before he meets the customer.
Worth knowing. This is an older article (2011), and the 90% is the owner's own estimate.
“We now hire salespeople who can focus on selling, not climbing the roof to get measurements.”
On the Roofing Contractor podcast, a roofing marketing agency owner said replying to reviews is the simplest first use of AI.
What they did. Recommended using AI tools, such as Podium and Birdeye, to respond to reviews because volume is high and time is short.
How it went. Advice from someone who markets roofing companies, not a measured result.
Worth knowing. She warned that AI doesn't always get facts right, so its claims need checking.
On a roofing podcast, the founder of a supplement company described using AI trained on past claims to speed up supplements.
What they did. Described AI built on tens of thousands of past claims, emails and call recordings to speed supplement turnaround and improve carrier communication.
How it went. No results are given in the episode notes.
Worth knowing. The guest sells this service, so treat it as a vendor claim. He says ChatGPT on its own isn't enough.
Agate CPA, with offices in Atlanta and Rogers, Arkansas, rebuilt its website form with AI tools so prospects answer the screening questions up front.
What they did. Built a form asking whether prospects own a business, their industry, size, revenue and CPA history, with answers flowing into their project management system.
How it went. The cofounder says the old form brought lots of sales calls but few conversions.
Worth knowing. They used Lovable to build the form and Claude to embed it in the website.
Their number. Conversion rate up about 25% (firm-reported).
At High Rock Accounting, an eight-person Scottsdale firm, the head of tax built a working onboarding and tax organizer prototype with Claude.
What they did. Used Claude and Claude Code to prototype a client onboarding flow and tax organizer.
How it went. A working prototype in a few hours, built by someone without a technical background.
Worth knowing. It's a prototype, not a finished system. The firm runs on Karbon for practice management.
John Eiduk of Eiduk Tax & Wealth says moving to an AI-based ledger shortened onboarding and let him lower bookkeeping prices.
What they did. Moved bookkeeping clients from QuickBooks to an AI-native accounting platform (not named).
How it went. Onboarding fell from two to four hours per client to about one, and bookkeeping package prices dropped by more than a third while staying profitable.
Worth knowing. The article doesn't name the platform, and the numbers are the owner's own.
Their number. Onboarding from 2 to 4 hours to about 1 hour per client (owner-reported).
“Once the model learns the transactions, I'm mostly reviewing instead of manually processing everything”
High Rock Accounting's COO built an app that asks clients for feedback after big projects and feeds it into Karbon.
What they did. Built the app with Claude Code and a Claude team account in four or five hours.
How it went. It replaced satisfaction and NPS programs the firm couldn't justify, and feedback comes in faster and cheaper.
Worth knowing. It collects private feedback, not public reviews.
Mana Accounting and Advisory's founder writes that moving to Digits cut cleanup time on client books.
What they did. Adopted Digits after QuickBooks and Zoho couldn't keep pace as the firm grew.
How it went. Cleanups that took entire afternoons now take minutes, he writes.
Worth knowing. It's an opinion piece that recommends the tool, with no disclosure either way. The numbers are the firm's own.
Their number. Up to five hours saved per client project (firm-reported).
Nick Boscia of Boscia & Boscia PC told an Intuit practitioner panel he uses ChatGPT to rephrase emails that come out too emotional.
What they did. Uses ChatGPT to rework emails about late payments and missing documents.
How it went. He says it helps him communicate clearly in tough situations, and he checks every draft before sending.
Worth knowing. He advises redacting anything confidential before using an AI tool. The recap is on Intuit's site.
H. Randy Hughes III of Counting Pennies LLC described using Fathom to capture client meetings and build proposals from the summaries.
What they did. Uses Fathom to record and summarize client meetings, then works from the summaries.
How it went. The same panel recap says his firm's tax-planning GPT turns a 12- to 14-hour task into under two hours, using anonymized client profiles.
Worth knowing. He reviews AI output for accuracy and says the setup should hold up to IRS scrutiny.
Their number. Tax plans from 12 to 14 hours to under 2 (practitioner-reported).
One Stop CPA's founder used an AI research tool and ChatGPT to answer a client's M&A and state-move question.
What they did. Used BlueJ for cited tax research and ChatGPT Enterprise to organize findings into follow-up questions and a client summary.
How it went. He produced the answer in about two hours, which he says would have taken much longer before.
Worth knowing. He's built reusable prompt templates, and treats AI output as a draft.
“AI is a starting point, not the final answer.”
Clio's lawyer in residence, discussing Clio's 2025 report on small firms, said they've barely dipped their toes into AI.
What they did. Clio studied solo and small firms. Its lawyer in residence suggested intake chatbots, promotional emails and first drafts for flat-fee matters as places to start.
How it went. Utilization rates were 27% for solos and 32% for small firms, versus about 50% at mid-sized firms, per Clio's report.
Worth knowing. These are suggestions and industry data, not one firm's results.
Their number. 59% of solo and small firms name referrals as their top lead source (reported by Clio).
A Practice Technology column surveys virtual receptionists and AI chat for small firms, including AI-powered website chat that captures and qualifies leads.
What they did. Describes tools like Smith.ai's AI-first receptionist, Intaker's AI website chat and LawDroid's virtual intake.
How it went. A survey of options, not one firm's results.
Worth knowing. The author works at MyCase, a practice management vendor. She warns that outsourcing intake means sharing firm data, which triggers confidentiality duties.
Lorraine Gingery of Lorraine Law, PC uses an AI tool to draft demand letters and search medical records by body part or procedure.
What they did. Uses Anytime AI for demand drafts, record summaries and plain-language searches of the case file.
How it went. The vendor says demands that took at least 4 hours now take under 30 minutes.
Worth knowing. This is the vendor's case study, and its dates don't match. Treat the numbers as vendor claims.
Their number. Demand drafting from at least 4 hours to under 30 minutes (vendor-reported).
“I can take 456 pages and immediately find what relates to a specific body part or procedure.”
In Mata v. Avianca, a federal judge in New York sanctioned two lawyers and their firm for a brief that cited cases ChatGPT made up.
What they did. A lawyer used ChatGPT for research, and the brief cited decisions that didn't exist.
How it went. The court found bad faith, imposed a $5,000 sanction, and ordered them to notify their client and the judges whose names were falsely used.
Worth knowing. This is why every AI citation and fact has to be checked against the source.
A short, curated list of practical AI jobs for local service businesses, organized by trade. Each use case shows the tools that do it, what they cost, how hard they are to set up, how many hours they save, and when to skip them.
Charlie Riske, who runs AIMention's AI consulting for small businesses in the Chicago suburbs. Tools are listed in alphabetical order, and no tool pays to be listed.
Prices come from each vendor's own pricing page and show the date they were checked. Real examples come from trade press, owner write-ups and labeled vendor case studies, and every one links to the original.
Yes. That's what AIMention's consulting does. Book a free 30-minute call and we'll pick the one or two use cases that pay off first for your business.
Tell me how your week actually runs. In 30 minutes I’ll tell you the one or two use cases worth doing first, and which ones to skip.
We may help you set these up. Tools don’t pay us for placement.