Most conversations about AI for tradies NZ start in the wrong place. Someone tells you to try a chatbot, you type a few questions into it, nothing changes, and you go back to quoting at 9 pm with a beer going warm beside the laptop. The tool was never the problem. The problem is that the admin side of your business only exists inside your head and your phone.
This post is about the practical version. Four specific jobs in a Kiwi trade business that AI handles well right now, the ones it handles badly, and how to start without taking a week off the tools to set it up. No theory, no future predictions. Just what actually works on a Wednesday when three clients are ringing, and you are up a ladder.
Key Takeaways
- AI for tradies in NZ works best on four jobs: answering missed calls, sending quotes faster, chasing follow-ups, and reviving dead leads.
- Most trade businesses lose more money to unanswered phones than to underpricing. The phone is the leakiest part of the operation.
- Roughly 78% of deals go to whoever responds first, so a 90-second reply beats a better price sent four hours later.
- Voice AI answers every call, qualifies the job and books it in. One clinic I worked with went from missed calls to 44% more bookings.
- Start with one task, not a full rebuild. One automation running properly beats five half-built ones nobody trusts.
- 95% of AI projects deliver nothing because they start with tools instead of the process the tool is meant to run.
- A quoting backlog is a memory problem, not a pricing problem. The system should know which 40 quotes are still open.
What AI for Tradies in NZ Actually Looks Like on a Wednesday
Picture a five-person plumbing outfit in West Auckland. The owner is on the tools four days a week. His phone rings maybe twenty times a day. He answers eight of them. The rest go to voicemail, and about half of those callers ring the next plumber on the Google results instead.
He quoted forty jobs last month. He knows roughly eighteen came back. He could not tell you which of the remaining twenty-two are still live, because that information lives in a mix of his memory, his text messages, and a notebook in the ute.
Nothing about that business is badly run. He is good at the work; his team turns up; the reviews are strong. The gap is that every piece of coordination runs through one head, and that head is under a house half the day.
This is what AI actually fixes for a trade business. Not the trade. The layer around the trade. The calls, the quotes, the reminders, the follow-ups, the “did we ever hear back from the Henderson job?” question that never gets asked because nobody has time to ask it.
The useful mental model is an AI brain that holds everything your business knows, plus an AI workforce that does the repetitive admin on top of it. The brain knows your pricing, your service area, your job types and your team. The workforce answers the phone, sends the quote, chases the reply and tells you what changed. You stay the tradesman. You stop being the receptionist.
The Four Jobs Worth Automating First
If you only ever do four things with AI in a trade business, do these. In this order.
1. Answer every call, including the ones you cannot take
The phone is where trade businesses bleed the most money and notice the least. A missed call is not a delayed call. It is a lost job, because the caller has a problem right now and four other numbers in the search results.
A voice agent picks up on the second ring, every time, including at 7 pm and on Saturday. It knows what you do, where you travel to, and what you do not touch. It takes the job details, works out whether it is urgent, books the callback or the site visit, and drops a summary into your CRM with the caller’s number attached.
I built this for a dental practice where 47% of calls were going unanswered despite two people on reception. Missed calls went to zero, and booked appointments rose 44%. Trades have the same shape of problem with worse coverage, because there is no reception desk at all.
The thing that surprises most owners is that customers do not mind. They mind voicemail. They do not mind a straight answer at 6:40 pm. More on the after-hours side in after-hours call answering.
2. Respond to new enquiries in under two minutes
Speed beats almost everything else in the sales process for trades. Harvard Business Review’s research on lead response found that firms responding within an hour were seven times more likely to have a meaningful conversation than those responding an hour later, and most companies took far longer than that (HBR, The Short Life of Online Sales Leads).
The practical version: a form fill on your website, a Facebook lead, a Builderscrack enquiry. Instead of sitting in an inbox until smoko, an automated response goes out within 90 seconds by text and email, asks the two or three questions you would have asked, and offers a booking time.
You are not competing on price with the other three quotes. You are competing on being the one who replied while the customer still had their phone in their hand.
3. Chase the quotes you have already sent
Every trade business I have looked at has the same silent leak. Quotes go out. Nobody follows them up. Not through laziness, just because following up forty quotes properly is a part-time job on its own.
Automated follow-up sends a short, human-sounding check-in at day three, day seven and day fourteen. If the customer replies, it stops and hands the conversation to you. If they say they went with someone else, it logs the reason so you learn something.
The conversion lift from this is usually the biggest single number in the whole exercise, and it costs nothing extra in advertising, because you already paid to get those leads.
4. Wake up the database you forgot you had
Every business more than two years old is sitting on a list of past customers and dead enquiries. Old quotes, one-off jobs, people who went quiet. That list has money in it.
A structured reactivation campaign, conversational SMS and email rather than a mailout, reopens those conversations at a rate that catches most owners off guard. For one finance client, 319 contacts the team had written off produced $49,000 in recovered revenue. Different industry, same principle: those people already know you and already decided once that they liked you.
For trades, this looks like maintenance reminders, second-stage work on a job you did eighteen months ago, or a simple check-in on a quote that went cold in winter.
Why Most Tradies Get Nothing Out of AI
MIT research on business AI projects found that around 95% deliver no measurable return. The pattern behind the failures is consistent, and it is not about the technology.
The failures start with the tool. Someone buys a subscription, plays with it for a fortnight, and nothing sticks because the tool was never connected to an actual process. It sits there as another login you do not open.
The ones that work start with the process. What exactly happens when a call comes in on a Saturday? Who decides whether it is urgent? What does the customer get told? Once that is written down clearly, pointing AI at it is straightforward.
There are three failure modes I see repeatedly in trade businesses.
Automating the hard thing first. Owners try to get AI to price jobs or make judgment calls about site access. That is the wrong end. Price and judgment stay with you. Answering, logging, reminding and chasing go to the system.
Building five things at once. One automation running reliably beats five half-built ones your team stops trusting. Once the crew stops trusting a system, they work around it, and you now have two processes instead of one.
No context. Generic AI gives generic answers. If the system does not know your service area, your minimum callout, your job types and how you talk to customers, it will sound like a call centre in another country, and your customers will hear it immediately. The context is the actual work. The software is the easy part. I have written more about that in AI Automation NZ.
How to Start Without Coming Off the Tools
You do not need a project. You need one job crossed off permanently.
Week one: count the leak. For five working days, log every inbound call and enquiry, whether it was answered, and how long the reply took. Do it in the notes app if that is easiest. Most owners find the missed-call number is somewhere between two and four times what they assumed.
Week two: pick the single biggest hole. For most trade businesses, it is the phone. For some it is the quote backlog. Take the one that costs you the most jobs, not the one that annoys you the most.
Week three: write down what should happen. Plain English, half a page. When someone rings about a blocked drain after 5 pm, here is what we ask, here is what we tell them, here is when we get back to them. This half page is what turns a generic AI tool into something that sounds like your business.
Week four: put one automation live and leave it alone. Watch it for a fortnight. Read what it said to people. Fix the wording where it is off. Do not add a second thing until the first one is boring.
That sequence works because each step gives you something on its own. Even if you stop after week one, you now know a number you did not know before, and that number usually changes how the owner thinks about hiring.
The other thing worth saying plainly: this does not replace anyone on your crew. It replaces the two hours of admin that were never going to get done properly anyway, sitting between the last job of the day and dinner. If you are weighing this against putting on an admin person, the honest comparison is admin overload in small business.
Where This Ends Up
The trades businesses that pull ahead over the next couple of years will not be the ones with better tools. They will be the ones where the coordination layer stops depending on the owner’s memory, which is where AI for tradies NZ can make a practical difference.
Right now your business works because you work it. Every quote that gets chased, every call that gets returned, every customer who gets remembered, all of it runs through one head. That is why you cannot take a fortnight off without the backlog eating the following month.
An AI brain that holds what your business knows, with an AI workforce doing the repetitive admin on top of it, changes that arrangement. Not overnight and not all at once. One task at a time, starting with the phone, until the parts of the business that used to need you no longer do.
The number that tells you it is working is simple. How many hours can you be unreachable before something breaks?
Ready to Find Your Biggest Leak?
If you want a straight read on where your trade business is losing jobs, book a 30-minute Discovery Call. No pitch deck, no jargon. We go through your call volume, your quote backlog and your follow-up process, and I will tell you which one is worth fixing first, even if you go and do it yourself.
Book a 30-minute Discovery Call
Frequently Asked Questions
What is the best AI for tradies in NZ to start with?
Start with call handling. For most trade businesses, the phone is the biggest leak, because missed calls turn into jobs for whoever answers next. An AI receptionist picks up every time, takes the job details, checks whether it is urgent and books the callback. It runs after hours and on weekends, which is exactly when a tradesman cannot answer.
How much does AI cost for a small trade business?
It varies with what you automate, but the useful comparison is not against software. It is against the cost of a part-time admin hire, or the value of the jobs you lose to unanswered calls each month. Most trade businesses find the missed-call figure alone covers a system several times over. Ask for a tailored quote based on your call volume.
Will an AI receptionist annoy my customers?
Customers dislike voicemail far more than they dislike a straight answer. A properly configured agent introduces itself, asks the questions you would ask, and books the next step rather than pretending to be human. The bad experiences people remember come from generic scripts with no knowledge of the business. Context is what makes the difference.
Can AI write quotes for my jobs?
Not pricing, and you would not want it to. Job pricing depends on site access, materials, and judgment you have built over years. What AI handles well is everything around the quote: collecting details, formatting and sending it faster, then chasing it at day three, day seven and day fourteen so it never sits forgotten in a text thread.
Do I need to be technical to use AI in my trade business?
No. The build work is done for you, and the part you contribute is the part only you can do: explaining how your business actually runs. If you can describe what happens when a customer rings about an urgent job on a Sunday, that is the input required. The technical setup is not your problem to solve.
How long before I see results from AI automation?
Call handling shows up within days, because you can count the calls that used to go to voicemail and now get booked. Quote follow-up takes three to four weeks to show a conversion change, since it depends on your quote cycle. Database reactivation usually produces replies in the first week. Give any single automation a fortnight before judging it.
Is AI worth it for a two-person trade business?
Often more so, because there is nobody to answer the phone at all. The smaller the team, the more the owner is doing admin between jobs. Below roughly $300k in revenue, the maths gets tighter, but for an established two- or three-person outfit turning away work because nobody picked up, the case is usually clear.
About Octavius
Titus Mulquiney is the founder of Octavius AI, where he builds AI brains and AI workforces for founder-led businesses stuck running everything out of their own head. Twenty years in marketing, ex-Sony product manager, ex-GM Zeal NZ. Based in Auckland, working with operators across NZ, Australia, and the US. Connect on LinkedIn.