Replying to Trade Enquiries in 60 Seconds With AI

You can't type a reply in 60 seconds from a ladder, so the first response has to be automatic: an instant trigger (a web-form webhook, a missed-call text-back or an AI receptionist) that replies using the enquiry's own details and asks the two or three questions you need to price the job. You follow up personally within the hour.

The catch is that many automations aren't instant. Zapier checks most apps for new data every 15 minutes on its Free plan and every 2 minutes on Professional, and email-based triggers can lag further. To reply inside a minute, every link in the chain has to be instant. Speed also only helps if the reply is useful: "Thanks, we'll be in touch" in 30 seconds gives a customer ringing round three firms no reason to wait for you.

Follow me on Instagram@sagnikteaches

Where trade enquiries arrive, and which can be answered instantly

List every place a customer can reach you, then check each against this table. Most small trades firms have four or five channels and only ever thought about one.

Connect on LinkedInSagnik Bhattacharya

For most firms, missed calls and the website form cover the bulk of new work, so start there and leave the rest for later.

Subscribe on YouTube@codingliquids

Why a two-minute check can't give you a 60-second reply

Automation tools start in one of two ways. A polling trigger asks an app "anything new?" on a timer. An instant trigger waits for the app to push the new enquiry the moment it arrives, usually through a webhook, which is a web address the form sends data to. In Zapier, instant triggers carry an "Instant" label and a lightning-bolt icon; polling ones don't. If your form tool's Zapier trigger isn't instant, check whether the form can send a webhook instead. Webhooks by Zapier isn't available on the Free plan, so budget for Professional (from $29.99 a month billed monthly, $19.99 billed annually, for 750 tasks).

The second trap is the last step. SMS by Zapier only sends texts to phone numbers you've verified yourself, so it's fine for alerting you and useless for replying to customers. For customer texts you need an SMS provider such as Twilio connected to the automation, or your job software's own texting. Many providers require you to register as a business sender before you can text customers, which can take days, so start that paperwork first.

Three builds, cheapest first

1. Missed-call text-back

If you only do one thing, do this. When a call goes unanswered, the caller gets a text within seconds. Many business phone apps and VoIP providers include it; check your provider's settings before paying for anything. Jobber's Receptionist does it too, and more (it answers calls and texts, takes details and can book visits): it's an add-on on Jobber's Core, Connect and Grow plans at $29 a month for 30 conversations, then $0.79 for each one after that, and it's included on Plus. A good missed-call text is fixed wording, not AI:

Sorry I missed you, it's [name] from [firm]. I'm on a job and will call back by [5pm / first thing tomorrow]. If you text me what you need and a photo of the problem, I can give you an idea of price when I ring. If water or electrics are an emergency right now, reply URGENT.

One edge case to plan for: calls from landlines. The text-back goes nowhere, and the caller, often an older customer, hears nothing. If your phone app shows the missed number, have it flag numbers that can't take texts, or simply scan the missed-call list at lunchtime and ring the landlines first. They're the callers least likely to try a second firm online, and the most likely to stay loyal if you ring back the same day.

2. Website form with an AI-drafted reply

This is where AI earns its place, because the reply can use what the customer wrote. A typical Zap: form webhook (instant) → an AI step that reads the enquiry and drafts a reply plus a one-line summary for you → a Filter that holds anything marked spam → an SMS or email to the customer → a text to your own phone with the summary. Filter steps use no Zapier tasks; an AI by Zapier step uses 1, 3 or 5 tasks per run depending on the model tier. Budget roughly four to eight tasks per enquiry. The instruction for the AI step:

You write the first reply to a new enquiry for [firm], a two-person roofing firm.
Enquiry details: {{name}}, {{phone}}, {{job_type}}, {{message}}, {{photos_attached}}.

Write an SMS under 320 characters that:
- uses their first name and restates the job in their words
- says [owner name] will call today if it arrived before 3pm, otherwise tomorrow morning
- asks for the 2 most useful missing details from this list, only if missing:
  property type, roof type, which part is leaking, photos, access (scaffold, height)
- never mentions a price, a date for the work, or a diagnosis
- if the message suggests water coming through a ceiling now, say to reply URGENT

Then on a new line starting SUMMARY: give me a 15-word summary and a label:
URGENT / REPAIR / NEW ROOF / GUTTERS / SPAM.

What that produces for two different enquiries (illustrative):

Enquiry: "Hi, we've got a damp patch on the bedroom ceiling that's growing
since the storm. Terraced house. Can someone look this week?"

Reply: Hi [first name], thanks for getting in touch about the growing damp patch on
your bedroom ceiling. [owner name] will call you today. Could you text a photo of the
patch and, if you can, one of the roof from the street? If water is dripping
through now, reply URGENT.
SUMMARY: Damp patch bedroom ceiling since storm, terraced house, wants visit this week. REPAIR

Enquiry: "quote for new flat roof on garage approx 3x5m"

Reply: Hi [first name], thanks for asking about a new flat roof for your garage, around
3 by 5 metres. [owner name] will call you tomorrow morning. Is the garage attached to
the house, and could you send a photo of the current roof?
SUMMARY: New flat roof, garage approx 3x5m, attachment and condition unknown. NEW ROOF

When you test, watch for the model breaking its rules. An early version of prompts like this one often produces "we can pop round tomorrow" or "that sounds like a slipped tile", because the model is trying to be helpful. Both are promises you haven't made. The fix is the explicit "never mentions" line and a test set (below) that includes enquiries tempting it to diagnose.

The same prompt adapted for a plumber shows how the rule-breaking looks in practice. The missing-details list becomes boiler make and age, whether there's hot water, heating or neither, any error code on the display, and a photo of the boiler. An illustrative enquiry and first draft:

Enquiry: "Boiler making loud banging noises since yesterday, still
got hot water but heating is patchy. It's about 12 years old."

Draft:  Hi [first name], sorry to hear about the banging. That sounds
like kettling, which is often caused by limescale and is usually a
quick fix. [owner name] will call today. What make is the boiler?

The draft is friendly, quick and wrong to send. "Sounds like kettling" is a diagnosis, and "usually a quick fix" sets a price expectation you'll have to walk back if the heat exchanger turns out to be the problem. Worse, a customer who hears "quick fix" may put off calling anyone. After the prompt was tightened with a second line ("Do not name a cause or suggest how big the job is, even if the customer asks"), the reply became: "Hi [first name], thanks for the details about the banging and patchy heating. [owner name] will call you today. Could you text the boiler's make and a photo of the display, including any error code?" It's less chatty and much more useful to the person ringing back. There's more on building AI steps in adding AI steps to Zapier.

3. An AI receptionist inside your job software

If you already run Jobber, Housecall Pro or ServiceTitan, their AI call and chat answering can capture details and book directly into your calendar, which a Zap can't do as neatly. How the three compare on AI features covers what each includes and what's sold separately, and what an AI receptionist costs a tradesperson helps decide whether it pays.

How a two-person roofing firm wired it up

Suppose, as an illustration, a roofing firm run by two partners gets about 35 enquiries a month: roughly 15 phone calls they miss because they're on a roof, 14 website forms and 6 WhatsApp messages. Replies used to go out in the evening, three to six hours after the enquiry, and the partners suspected they were losing repair jobs to whoever called back first.

Their setup, in the order they built it:

1. Missed-call text-back through their phone app's existing setting, with the fixed wording above. Cost: nothing extra. Time: 20 minutes.
2. WhatsApp Business app greeting and away messages, fixed text, asking for photos. Cost: nothing. Time: 15 minutes.
3. Website form Zap on Zapier Professional with an SMS provider, built on a Saturday morning, about four hours including testing. Running cost: the Zapier plan plus a few cents per text. At about seven tasks per enquiry, 14 forms use under 100 of the 750 monthly tasks.

A typical enquiry now runs like this: form submitted at 14:02:10, customer's text arrives 14:02:31, summary lands on the partner's phone at 14:02:33, partner calls back at 15:10 between jobs with the photos already in hand. Whether it wins more work is something they'll only know from their own numbers, so they started logging, for every enquiry, the channel, time of first reply and whether it became a quote and a job. After two or three months that log answers the question better than any industry claim.

The log needs only five columns. A few illustrative rows from their first fortnight:

Even five rows point somewhere. The WhatsApp row lost the job in the gap between the instant greeting and a real reply, which suggests the greeting bought no time at all. The fix they tried next was rewording the greeting to say when they'll call back, and checking WhatsApp at lunchtime as well as in the evening, so no message waits more than a couple of hours for a real answer.

Whether the setup pays is a quick sum once you pick an assumption. Their running cost is about $30 a month for Zapier Professional on monthly billing, plus a few dollars of texts. If faster replies win them just one extra repair a month at around $400, the build has paid for itself many times over. If after three months the log shows no change in the share of enquiries that become jobs, the problem is probably price or availability rather than speed, and that's worth knowing too.

Emergencies need a different first reply

An automated reply to "water pouring through the kitchen light" that says "we'll call tomorrow morning" is worse than no reply. Give emergencies their own route: keywords in the form or message (leak, pouring, sparking, burning smell, no power, no heating with a baby at home) and the AI label URGENT both trigger an immediate call or loud alert to whoever is on call, and a fixed customer message you've written and checked yourself. Don't let the AI improvise safety advice. If the right advice is "turn off the electricity at the consumer unit if it's safe to reach", that sentence should be yours, word for word, and reviewed by whoever is qualified to give it.

Testing it like a customer would

Before relying on it, send ten enquiries yourself from a friend's phone and email: at 7am, at 10pm, on a Sunday, with typos, with no phone number, with a photo, one obvious spam message, one emergency, one asking for a price outright, and one that also calls. For each, note the time sent and the time the reply arrived, and read the reply as the customer would. A pass is a reply under 60 seconds, nothing promised you can't keep, and a summary on your phone you could act on from the van.

Expect a first run to fail some of these. The roofing firm's first ten tests, for illustration, produced seven passes and three failures. The enquiry with no phone number made the Zap error at the SMS step, so nothing reached the customer or the partners; they added a path that emails the customer instead when the phone field is empty. The 10pm test got an instant text, which prompted the quiet-hours rule below. And the "how much for a new roof?" test got a reply that said "most roofs cost between..." before listing the questions, a price hint the prompt had forbidden. A second run after those fixes passed all ten, and they now rerun the same ten whenever they edit the prompt or the form.

Then keep an eye on it. Automations break quietly when a form changes a field name or an SMS provider's balance runs out. Turn on error alerts and check the task history weekly for the first month; stopping automations breaking silently covers what to set up.

Quiet hours, spam and double replies

Three small problems catch most setups in their first month. Night-time texts: an instant text at 23:40 is fine for an emergency and irritating for a gutter quote, so send non-urgent replies by email overnight and hold SMS until 7am. Spam: contact forms attract bots, and each bot submission costs tasks and, worse, a text to a fake number; add the form tool's spam protection and let the AI's SPAM label stop the run at the Filter. Double replies: a customer who fills in the form and then calls gets two different messages, so have the Zap check whether that phone number has already had a reply today. None of these are hard to fix. They're just easier to fix before a customer points them out.

For the broader version of this approach across any service business, see replying to every enquiry in under five minutes, and for the text-back step in detail, setting up missed-call text-back.

Further reads

• Can AI Triage Emergency Plumbing Calls Out of Hours? — Handling urgent calls that arrive out of hours.
• WhatsApp Business App vs API: Which Do You Need for AI Replies? — Which WhatsApp setup you need for automatic replies.
• How Much Do Missed Calls Cost a Heating and Cooling Business? — Put a figure on the enquiries you're losing now.
• The First 30 Days of AI in a Trades Business — Where instant replies fit in a first month of AI.
• Seven AI Mistakes That Cost Tradespeople Jobs — Automation errors that lose trades work.
• What Is the Cheapest Way to Handle Enquiries Out of Hours? — Compare the low-cost options for evenings and weekends.
• AI Tools for Roofing Contractors in 2026: Estimating and Leads — Roof measurement, instant estimates, lead replies and photo documentation tools for roofers, with current prices and three stacks by firm size.
• How a Plumber Can Save Five Hours a Week on Admin With AI — An illustrative time log for a sole plumber, six changes that add up to five hours a week, a four-week rollout and what it costs.
• Instant Roof Estimates on Your Website: Do They Win More Jobs? — How satellite-measured instant roof estimators work, why they raise lead volume but change lead quality, what they cost, and how to test one on your own site.
• How Painters and Decorators Can Quote From Photos With AI — Send a photo brief, let AI read surfaces and condition, get sizes from a tape or scan, price from your rates and word the range so it holds up on the day.
• Following Up Silent Quotes With AI: A Decorator's Playbook — Four follow-up plays for decorating quotes that go quiet, from the day-two check-in to the day-thirty close-out, with AI drafting each one from your notes.
• Following Up Unapproved Repair Estimates With AI — A step-by-step way for repair trades to chase open estimates with AI-drafted messages that explain the fault honestly and allow partial approval.
• What Tradespeople Actually Use AI For: 12 Real Jobs It Handles — Twelve everyday trade jobs AI handles well, from voice-note job sheets to invoice chasers, each with a real-looking example and the mistake to watch for.
• Nine Everyday AI Uses for Electricians Beyond Estimating — Nine electrician-specific jobs AI handles, from explaining inspection findings to landlords to logging intermittent faults, each with a worked example.
• Can a Roofer Quote From Satellite Images? How AI Measurement Works — How aerial roof measurement reports are made, where they go wrong, and a two-stage way to quote from them without losing margin.
• Can AI Give Customers a Ballpark Price for Decorating Work? — How a decorator can let AI give honest ballpark prices from a room-by-room grid, and which jobs it should refuse to price.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Zapier help pages on instant and polling triggers, plan polling intervals and SMS by Zapier limits; Jobber pricing page and Receptionist help article; Google Business Profile help on the end of chat; Meta announcements on WhatsApp pricing.