The NZ Trades AI Checklist
Eight practical checks to know whether your plumbing or electrical business is ready for an AI admin rollout. Built from real NZ trades-system rollouts — graduate when you have three green ticks and the rollout will actually stick.
Why this list
The eight things that decide whether AI sticks
- It’s the eight things we see NZ trade businesses fix before any AI workflow ships. Skip ahead and the rollout stalls.
- Some checks take 20 minutes. The whole list fits in a single afternoon with the owner and one lead technician.
- You don’t have to pass every check to start. Three green ticks across the list is enough; the rest you improve while the AI runs.
Capture job notes in one place
Every technician writes the brief note after a job in the same place — a phone app, a shared form, a notebook that gets typed up. If today the note lives on a sticky note and gets retyped on Monday, the AI has nothing to read and the workflow stops at the first step.
Use one consistent job taxonomy
Pick a short, plain-English list of job types (toilet install, hot-water cylinder replacement, RCD trip…). Apply it consistently across every technician. AI works best when the same kind of job is described in the same words every time.
Standardise one pricing template
One quote template — even a basic one — that lists labour rate, materials margin, callout fee, and GST separately. The AI fills the template; you don’t re-key the math. If pricing today varies per quote in a way that nobody can articulate, that’s the work to do first.
Decide your follow-up rhythm
Pick how many days after a quote to nudge a client, and when to give up. A documented nudge rhythm lets the AI run the same sequence for every quote without you making the call each time. "Just ring them" is fine as a fallback but it doesn’t scale.
Move invoices into one system
Whatever invoicing tool you use, make sure the quote-to-invoice link is one click. AI can draft an invoice from a quote in seconds; if the handoff is a printout and a manual re-type, the win collapses and the team stops trusting the output.
Know your turnaround target
Write down — even on a whiteboard — how fast you want a quote after the technician’s note. Sub-60s is a realistic target for AI; same-day is realistic with semi-automation. Without a target the workflow feels like a science project and the team disengages.
Make the dashboard the answer place
Quote turnaround time, follow-up conversion rate, jobs in flight. Pick three numbers. Make them visible every Monday morning. The AI optimises what the team can see; numbers nobody opens are invisible to the model.
Choose a human escalation path
Even on a fully autonomous system, name the one or two people who step in when a quote is unusual, a client is escalated, or a payment fails. Without an escalation rule the team either ignores the AI’s output or chases every edge case by hand.
Where to next
Once you have three green ticks, see the packages
All three passes usually arrive inside a single afternoon. When you’re ready to see what the rollout looks like, the pricing page walks through the diagnostic, the 30-day sprint, and the ongoing managed plan — same checklist logic, just delivered for you.