We consolidate the records your business already holds — sales, stock, customers, documents — into a database an AI can work from, then put it on one defined job that today costs a person hours every week. Two engineers work inside your business until it runs.
The jobs we are asked for most often. Each is running in a New Zealand business today — none is a demonstration, and each has been in daily use for months. Open any one to see precisely what the AI does, and what the business had to have in place first.
Answer the enquiries
Bring the customers back
Quote from a drawing
Plan the day’s run
Run the ad campaigns
Run the monthly review
“We’d been sitting on ten years of receipts and never did anything with them. Now there’s a list every Monday with a message already written next to each name. I go through it with a coffee and say yes or no. The sales desk got most of a day a week back.”
AlexOwner, IrelaxRead the study →SSA WHEELSTyre retail & wholesale · Auckland“Three branches, three different names for the same tyre. They spent time in our warehouse with the guys before building anything, which I didn’t expect. Now the pickers scan on their own phones and we hardly see wrong-tyre returns anymore.”
MelodyHead of Operations, SSA WheelsRead the study →ARCHIMAXAluminium cladding supplier“Pricing a drawing used to take me most of a day. Now it comes back before lunch, in our own item codes. And when a rule’s unclear it asks me instead of guessing. Honestly, that’s the bit that won me over.”
AndyQuantity Surveyor, ArchimaxRead the study →“The website was the easy part. The real change was the client system. Every enquiry, every property and every conversation now lives in one place, instead of in my inbox and three spreadsheets.”
VincentOwner, Mainstay Property Group“Working out how much timber and board a kitchen needs used to take ages, and a small mistake meant ordering twice. Now the materials come out of the drawings, and we check them instead of counting from scratch.”
SherryOwner, Popular Kitchens“Customers ask the same things all day: where’s my delivery, can you take this load. Now the website and WhatsApp answer them straight away, and anything that needs a person comes to us.”
AidenOwner, Dawn Transport“Enquiries now get an answer online straight away, even after hours, and the good ones come to us properly written up. Nobody is copying details out of chat windows anymore.”
CarolineOwner, Zealive Institute“We’re builders, not marketers. They run our website and use AI for the SEO and the ads, and every month we can see which of it actually brought in enquiries.”
TiffanyOwner, L’Abri Construction“We knew AI mattered but didn’t know where to start. They came in, looked at how we actually work, and gave us a clear, practical plan rather than a sales pitch.”
SamOwner, Proactive Fire Protection
A large language model on its own has nothing to work from. For it to take a task off someone’s desk and carry it out the way that person would — with your data, your rules and your systems — the business has to have seven things in place first. None of them is optional. Every scene above has all seven behind it, which is why they run.
One task, with “done” spelled out.
An AI cannot help with “the business”. It can carry out one task that has a clear start, a clear finish, and a result someone can check at the end. The more narrowly the task is defined, the more dependably it is carried out.
In a place it can read from and write to.
Customers, orders, products and documents usually live where an AI cannot reach them: a spreadsheet on someone’s laptop, the point-of-sale, a messaging thread, a folder of files. They have to be brought into a proper database the AI has access to, with the fields the task needs and nothing missing. Data the AI cannot reach, it cannot use.
How the task is done today, on paper.
At present the task lives in one person’s head: which case comes first, how a figure is arrived at, what is checked before anything goes out. That judgement has to be written down as rules the AI can follow. Without them the AI guesses; with them, it follows.
It reads and writes where you already work.
The AI has to take its data from the POS, the online store, the accounts and the messaging you already use, and return its results to the same places. Otherwise someone is re-keying again within a month, and the tools people know do not have to change.
Larger jobs are several agents working together.
An AI does its best work when it is given one task, one set of data and one set of rules, and nothing else. That is what we call an agent. A larger job is not one agent stretched wider, but two or three working together, each doing its part and handing on to the next. Narrow agents are what keep the whole dependable.
It runs on its own; the important calls wait for you.
The AI works daily or weekly without anyone starting it. But AI is not correct every time, and in a business a wrong message to a customer or a wrong quantity on an order costs more than the time saved. So the actions that matter — anything that reaches a customer, anything with money on it, anything that changes your records — stop in your queue first, and a person approves or edits before it goes. Everything else runs through. Every action, approved or not, is on record.
Data drifts, rules change, systems update.
A supplier renames a product, a system updates, a rule that was right in March is wrong in June. An AI nobody maintains stops being right within months. Someone has to own it after the launch week, read the numbers each month, and adjust.
Whether a business has five staff or fifty, these seven are a real challenge. There is rarely a data team, rarely anyone with a spare week to write the rules down, and rarely anyone to keep it running once a system changes underneath it. That is the work we do. You hand the seven over; we bring two engineers to your business and build them one at a time, in your office, until the AI is doing the job.
Nothing is paid before you have seen it running on your own data.
AI is changing how small and medium-sized businesses run, and you shouldn’t have to work it out alone. Start with a conversation, not a pitch: tell us the three jobs that take up most of your week, and we’ll come to you and tell you plainly which ones AI can take off your hands. It’s free — and the time you get back is yours, for your customers, your team, or the view.
Four steps, two minutes. Within two working days: can AI take them on, what it needs, roughly what it costs.