AI Automation for Tradies: Quote Follow-Ups That Run Themselves

Send three short messages to every quote that goes quiet: a check-in on day 2, something useful on day 6 and a polite close-out on day 14. Let your job software or a Zapier link send them, let AI write the wording once, and stop the whole sequence the moment the customer replies or the quote is accepted.

Two details decide whether this wins work or annoys people. The sequence must re-read the quote's status before every send, because a chaser for a job you won by phone yesterday costs more goodwill than the quote was worth. And bigger quotes deserve a phone call, so the automation should flag those to you instead of texting them.

Follow me on Instagram@sagnikteaches

Check what your job software already chases for you

Several popular tradie apps list a quote follow-up feature on their own pages, so open the quoting settings of the one you use before building anything. This is what the vendors say, read in October 2026. Plans and features move, so confirm each line in your own account.

Connect on LinkedInSagnik Bhattacharya

Built-in follow-ups are scheduled templates. They drop a name and a job title into wording you wrote and send it on a timer. They don't read the quote, understand a reply or decide that a $14,000 job needs a phone call. That is where AI earns its place: writing wording that sounds like you and fits the job, then sorting the replies.

Subscribe on YouTube@sagnikteaches

Two limits are worth knowing before you rely on a built-in sequence. Jobber's automations stop at two reminders, so a third message has to be a manual task or part of your own build. And because Jobber sends each follow-up at the same time of day as the original quote, a quote you finish and send at 9 pm gets its chaser at 9 pm too. Send quotes in the morning, or build the sequence yourself with a fixed send time.

Three messages, three jobs: days 2, 6 and 14

The rhythm below is a starting point, not a rule, and I don't know of a published study that proves one best interval for trade quotes. Give it a month, then adjust from your own reply times. What matters more than the exact days is that each message has a different job, because three copies of "just checking in" read as nagging by the second.

Day 2: confirm the quote arrived

Quotes land in spam folders, sink under other messages or sit unread. The first message is a low-pressure check that it reached them, with an easy way to ask a question. Keep it to one idea. A text version:

Hi [first name], it's [your name] from [business]. Just checking the quote for [job] reached you OK. Any questions, reply here or ring [number]. Reply STOP to opt out.

The email version carries the same message under a subject line that names the job, such as "Your quote for [job]", with your phone number in the signature.

Day 6: answer the question they haven't asked

By day 6 the customer has probably talked it over with a partner or collected a second quote. Use this message to remove the likeliest doubt and say something true about the job. For a rewire that might be what the price includes; for a bathroom refit, when you could start. If the total is the sticking point and you really would split the work into stages, say so here and nowhere else.

Hi [first name], [your name] from [business] again. People usually ask about [the likeliest question for this job type]. For [job]: [one-line answer]. If the total is the sticking point, I can price it in two stages. Reply or ring [number]. Reply STOP to opt out.

Day 14: close the loop kindly

The last message gives the customer an easy way to say yes, later or no. It also tidies your list, because a quote that stays open for months distorts your win rate. Only mention a closing date you will honour, and never invent a price rise or a shortage of slots.

Hi [first name], I'm tidying my quote list on [date], so I'll close the quote for [job] then. Still keen? Reply YES and I'll book you in. Timing changed? Reply LATER and I'll check back in [month]. Nothing needed otherwise. Reply STOP to opt out.

SMS tools bill by message segment. Twilio's documentation says plain GSM-7 text holds 160 characters a message, and that any character outside that set, including the angled smart quotes some word processors insert, makes the whole message fall back to an encoding with a 70-character limit. Emoji sit outside the set too. So type straight apostrophes and skip emoji. With real names and numbers filled in, my three sample messages come to roughly 180, 265 and 255 characters, so each is billed as two segments. Trim them if that bothers you.

Let AI draft all three from the quote itself

Writing the wording once per job type takes a few minutes in any chat assistant. The prompt below works in ChatGPT, Claude or Gemini. Paste in the facts you want used and nothing else, because text a model invents about availability or inclusions is the main risk, so the prompt forbids it.

You write follow-up texts for a small trades business.
Business: [business name], a [trade] business. Owner: [your name].
Quote: [one-line job description]. Value: [amount]. Sent: [date].
Facts you may use: we can start from [date]; the quote includes [list]; it excludes [list].
Write three SMS messages for a customer who has not replied:
1. Day 2: check the quote arrived and invite questions.
2. Day 6: answer the question people most often ask about this kind of job, using only the facts above. Offer to split the work into stages only if the value is over [amount].
3. Day 14: say I will close the quote on [date]; reply YES to go ahead or LATER to be checked in [month].
Rules: British spelling, no emoji, plain straight apostrophes, no discounts, no claims about availability or price rises that I have not given you, under 300 characters each, every message ends with "Reply STOP to opt out."
Return the three messages only, numbered.

Here is the kind of output you get back for a made-up kitchen rewire quote of $3,850. The two phrases in italics are the ones I'd delete:

1. Hi [first name], [your name] from [business]. Just checking the quote for your kitchen rewire reached you OK. Any questions, reply here or ring [number]. Reply STOP to opt out.
2. Hi [first name], [your name] again. Most people ask when we could start. We can start from [date]. Making good the plaster is included. If the total is the sticking point I can price it in two stages. Reply STOP to opt out.
3. Hi [first name], I'll close your kitchen rewire quote on [date]. Reply YES to go ahead or LATER and I'll check back in [month]. Slots are filling fast. Reply STOP to opt out.

Both italic sentences are the model improvising. I never said plastering was included or that slots were filling, so they go. That is the standard fix with AI-written follow-ups: read every line against the quote, delete anything you did not supply, and save the corrected version as your template. You review once per job type, not once per customer. For more prompts in this vein, see ten ChatGPT prompts for tradies covering quotes, texts and follow-ups.

Rules that keep a sequence helpful instead of pushy

These are the settings I'd decide before switching anything on. Most are one line in your job software or one filter in a Zap.

• Stop on any reply. A reply by text, email or phone call pauses everything and alerts you. Even a "thanks" means a person has joined the conversation.
• Stop on a status change. Accepted, declined, expired or booked ends the sequence. This is the rule that prevents the embarrassing chaser, and it only works if you keep statuses up to date.
• One channel per customer. Use the one they used to contact you. Jobber does this by design, sending follow-ups the same way as the original quote.
• Daytime only. Send between about 9 am and 5 pm on working days, then look at your reply times after a month before widening that.
• A phone-call threshold. Pick a value, say $6,000, above which the day-6 message becomes a task for you ("Ring [customer] about [job]") instead of a text.
• A do-not-chase flag. Repeat customers, friends of friends and anyone who said "give me till spring" get no automatic messages.
• A hard cap. Three messages and the sequence ends. A fourth is a decision you make by hand.

Build the chaser yourself with a sheet, a Zap and three delays

If your quotes live somewhere without a built-in follow-up, or you want the third message and the call threshold, a spreadsheet and Zapier will do it. Zapier's Free plan allows only two-step Zaps, so this needs a paid plan. Professional lists at $19.99 a month billed annually or $29.99 billed monthly, for 750 tasks a month (Zapier pricing page).

Start with the log. One row per quote, with columns for quote ID, first name, mobile, email, job, value, date sent, status (Quoted, Accepted, Declined, Closed), replied, do not contact and scope line. The sheet is the single source of truth, so whoever sends a quote adds a row. If your job software has a Zapier connection, a second Zap can add the rows for you.

Then build one Zap with the shape below. Task counts follow Zapier's task-usage help page, which says triggers, Filter, Paths, Formatter and Delay steps don't count, while action steps do.

Trigger  New spreadsheet row in the Quotes sheet (status = Quoted)        [no task]
Step 1   AI step: write a one-sentence plain-English scope line          [1 or more tasks]
Step 2   Delay for 2 days                                                [no task]
Step 3   Find the row again by quote ID                                  [1 task]
Step 4   Filter: status = Quoted, replied blank, do not contact blank    [no task]
Step 5   Send message 1 by SMS or email                                  [1 task]
Step 6   Delay for 4 days, then repeat steps 3 to 5 with message 2       [2 tasks]
Step 7   Delay for 8 days, then repeat steps 3 to 5 with message 3       [2 tasks]
Step 8   Update the row: status = Sequence finished                      [1 task]

Step 1 is what makes the messages read as written for this job. The scope line becomes the [job] in each text, so a long quote description turns into "your kitchen rewire". Then the Find step after each Delay matters more than anything else in the build. A Zap carries the data it saw when it started, so without a fresh look at the sheet it would happily chase a quote you accepted yesterday. Choose the new-row trigger rather than the new-or-updated-row one, too, or every edit to a row starts a second sequence for the same quote.

Count the tasks before you trust the price. That is seven tasks per quote plus the AI step. Zapier's AI steps can use more than one task depending on the model option, and I couldn't confirm the current counts, so plan with 1 to 5. That makes 8 to 12 tasks per quote, or 320 to 480 at 40 quotes a month, inside the 750-task tier. A sequence that stops early costs less, because filtered steps never reach the send.

If you would rather approve each message than let it fly, there is a lighter variant. A formula in the sheet flags which quotes are due message 1, 2 or 3 today, and a daily digest email lists them with the text ready to paste. It keeps a person on every send, which is the same idea as ServiceM8's approval-first design. For a longer email-only version of the idea, see a quote follow-up email sequence AI can chase for you.

Let AI label the replies, then answer them yourself

Replies are where automation usually goes wrong, because they arrive as messy human sentences. The safe split is that AI labels each reply and you write the answer. This prompt runs in a Zap step, or in a chat window if replies are few:

You read one customer reply to a quote follow-up from a trades business.
Choose exactly one label: ACCEPT, QUESTION, PRICE, LATER, DECLINED, STOP, UNCLEAR.
If the reply asks for messages to stop in any wording, the label is STOP, whatever else it says.
Return JSON only: {"label": "", "summary": "", "next_check_date": ""}
Use YYYY-MM-DD for next_check_date, or null. Today is [date].
Reply: "[paste the reply]"

Run against six invented replies, a good model gives you results like these:

Two guardrails. Never let ACCEPT trigger a booking on its own, because "yep" can mean "yep, I'll think about it". And read the UNCLEAR and STOP pile daily for the first month, since those are the labels where a mistake costs you a customer or a complaint.

A 40-quote month, worked through

Take a two-van electrical firm that sends 40 quotes in a month at an average of $2,400. Every figure below is my assumption, there to show the shape of the sum. Replace them with numbers from your own log.

Suppose four of the 29 quiet quotes were above $6,000. Those got a phone call from the owner on day 6 instead of the text, and two of the five sequence wins came from those calls. The table still values the two extra wins at the $2,400 average, which keeps the sum cautious, because a won job above $6,000 would only add to it. At an assumed 30% gross margin, two extra jobs add about $1,440 a month in gross profit. Against that sits roughly $20 a month for Zapier on annual billing, your SMS provider's per-message charge (small at under 90 messages, but look up your rate) and about an hour a month reviewing the log, which at $60 an hour of your own time is $60. Call it $100 a month all in.

At 30% margin an average $2,400 job leaves $720, so $100 a month is covered by one extra win every seven months or so. Halve my assumptions and it still pays. The real cost is the three or four hours of setup and the habit of keeping statuses current, which is worth deciding before you start. If you are still choosing software, ServiceM8, Tradify or Fergus by business size compares them, and Jobber's plans and pricing for a small crew shows where its follow-ups sit in the price ladder.

Opt-out lines and spam-law wording

Here I can offer practical wording, not legal advice. Rules on unsolicited messages differ by place and by channel, and email and text are often treated differently, so ask a qualified adviser what applies where you and your customers are. What follows is a set of habits that holds up under most rules.

• Message only people who asked for the quote. The sequence is about the job they requested. Don't import an old list of past enquiries and chase them, and keep offers for other services out of all three messages.
• Say who you are. Every message names the business and gives a way to reach a person.
• Give a plain opt-out and honour it at once. Texts end with "Reply STOP to opt out." Emails carry a line such as: "If you'd rather not hear from us about this quote, reply 'no more' and we'll stop straight away." When someone opts out, tick the do-not-contact box in the same minute and let your Zap's filter do the rest.
• Treat an opt-out as total for the sequence, whichever channel it came through, including a phone call to you.
• Don't send consent-request messages to people who haven't agreed. A "may we email you?" message is itself a message, and a regulator has fined businesses for sending them.
• Remember that an opt-in doesn't last for ever. A box someone ticked years ago is not a licence to message them today.

A STOP reply deserves a confirmation, so the customer knows it worked. A short one is enough: "You're opted out. We won't message you again about this quote. [business]". Check whether your SMS provider already sends its own opt-out confirmation, so nobody gets two.

One rhythm, three trades: signage, security and solar

The three-message rhythm suits a typical home job decided within a fortnight. Other trades need different timings and different content. Here is how the same build bends.

A signage company with artwork to approve

A sign quote can turn on whether the customer can picture the finished result. Swap the day-6 "likeliest question" for an offer of an artwork proof if you produce them, and keep the quote validity honest, because material prices move. The day-14 close-out then says exactly which date the price holds until.

A security installer who surveys before quoting

The decision follows a site visit, so the day-2 message should come from the surveyor's name and mention the visit. If the quote includes monitoring or a monthly fee, the day-6 message states the running cost plainly. A running cost the customer has to ask about is an easy reason to wait.

A solar installer with a decision that takes weeks

Stretch the rhythm to days 3, 10, 24 and 45, and make every message add something new, such as a short note on how to read the proposal. Solar quotes tend to be large, so the call threshold will catch most of them and the sequence becomes a set of prompts for your phone calls.

Small repair jobs under $500

Send one reminder on day 3 and stop. A three-message sequence on a $300 job risks annoying a customer you may want for years, and the margin doesn't justify the nagging.

What to measure after four weeks and the mistakes to expect

Keep the scorecard small enough to read in ten minutes each week.

The mistakes below are the ones I'd expect to see, each with its fix.

• The chaser that follows an acceptance. A customer rings on Tuesday to accept, the status stays Quoted until Friday, and a "Still keen?" text lands on Wednesday. Change the status during the call, and add a five-minute Friday check of the log.
• "Hi , it's [your name]". A blank first-name cell produces an awkward greeting with a gap in it. Make the filter require a first name, or use a fallback of "Hi there".
• The midnight text. A late-night quote gets a late-night chaser under Jobber's same-time-of-day rule. Send quotes in the morning, or use a fixed send time in your own build.
• Two sequences for one quote. The new-or-updated-row trigger restarts the Zap whenever a row is edited. Use the new-row trigger.
• Doubled SMS bills. A curly apostrophe pasted from a word processor drops every message into the 70-character encoding. Paste into a plain-text field first.

If you want a wider view of where AI can take work off a trade business, the AI guide maps the other jobs, and a quote sequence is one of the safest first ones because every message is about a job the customer asked for. For bigger jobs with long decisions, an HVAC replacement follow-up sequence shows a longer cadence in detail, and the quote template wording guide helps you write quotes that give the follow-up something specific to refer to.

Questions tradies ask before automating quote chasers

Is it worth automating if I only send ten quotes a month?

Probably not yet. At ten quotes a month, three saved messages and a daily calendar reminder do the same job without a subscription. Automation starts to pay once you cannot remember who has been chased, which tends to happen at twenty or more quotes a month or when several people write quotes. Start with the spreadsheet log either way, because the log is what any later automation will run from.

Do the follow-ups built into job software use AI?

The scheduled ones usually do not. They are templates you write, sent on a timer. Some vendors sell a separate assistant on top, such as ServiceM8's Office Agent, which drafts quote follow-ups and waits for your approval. Use AI where it adds something the timer cannot: writing wording that fits the job, summarising the scope in a sentence, and labelling the replies that come back.

Should follow-ups go out at weekends?

For homeowners comparing quotes in their own time, a Saturday morning text can work. Early mornings, late evenings and Sundays rarely do. For builders, property managers and other trade customers, stay with weekday working hours. Neither is a rule, so log the day and time of every reply for a month and let your own customers show you when they answer.

How should I follow up a quote sent to a builder or a property manager?

Ask at quote time who decides and by when, and record that date in the log. These customers often hold your price while their own client decides, so a chaser on day 2 is noise. Schedule the first message two days after their decision date, then one more a week later, and move to a phone call if the job is large.

Further reads

• AI Automation for Contractors: Qualify, Book and Confirm Site Visits — The step before the quote: qualify enquiries and book site visits.
• Getting More Reviews After Every Job With AI Follow-Ups — Ask for reviews after the job using the same stop rules.
• AI Apps for Contractors: A Phone-Only Setup for One-Truck Firms — Which AI apps a one-truck contractor can run from a phone on site, in what order, what each costs, and how to check every step is working.
• Best AI for Tradies: Seven Tools by Job, Price and Setup Effort — Seven AI tools ranked for quoting, calls, invoices and marketing, with vendor-page prices and a realistic setup time for each.
• Six AI Automations an Electrician Can Set Up in a Weekend — Missed-call texts, invoice reminders, quote chasers, review asks, job notes and certificate reminders, ranked by weekend effort, cost and payoff.
• Skip Hire Software: A Buying Checklist for Quotes and Permits — Eight quote inputs, a permit tracker, a run sheet and waste paperwork, with what five skip hire vendors print on their pages and a 45-minute demo.
• Tradie Apps: Quote, Schedule and Invoice Tools Compared, With AI Features — Tradify, ServiceM8, Fergus and Jobber scored on the AI each one documents for quotes, schedules and invoices, with list prices and a trial script.
• 10 ChatGPT Prompts for Electricians: Quotes, Emails and T&Cs — Ten prompts for quotes, customer emails and terms and conditions, each with a filled example and the check to run before anything is sent.
• AI for Construction Contractors: Bid Invitations and Follow-Up — A working routine for a contractor's inbox: triage invitations, draft acknowledgements and clarifications with AI, then chase each bid on a set cadence.
• Air Duct Cleaning Business: Quotes, Reports and Follow-Ups With AI — A per-vent rate card, a job-report routine built on before-and-after photos, and follow-up texts that offer only what the job showed.
• CCTV Installation Quote Template: A Worked AI-Drafted Walk-Round Quote — One six-camera site, from messy walk-round notes to a checked quote: the prompt, the faults in a first draft, storage maths, exclusions and follow-up.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Jobber pricing page and help centre (Automations), ServiceM8 pricing page and help centre (Introducing Office Agent), Tradify pricing page, Fergus pricing page, Zapier pricing page and task-usage help page, Twilio documentation on SMS segments and encoding. Pages read October 2026.