We build it.Then we teach it.

An AI lab in Auckland. We build apps, SaaS tools and browser extensions — and train New Zealand teams to build their own.

Scroll

How we think about this

Most AI advice comes from people who don't build.

The market is full of strategy decks written by people who have never shipped a feature with this stuff. We build software first. Everything we teach comes out of that.

A demo is not a workflow.

Nearly every AI pilot works in the meeting and dies in the month after. The hard part was never the model. It's the messy edges — your data, your process, and the person who has to trust the output on a Tuesday.

So we work in the open.

We publish what we build, what it cost, and what broke. If you would rather learn it than buy it, that is a good outcome for us too.

How we work

How a problem becomessomething your team uses

Every engagement runs the same loop, whether we are building the thing or teaching your people to build it. It is deliberately slow at the start, because that is where projects are usually lost.

Talk to us

Step 01
Frame
We start with the workflow, not the model. What takes too long, and who does it?
Step 02
Prototype
Something you can click, early — before anyone commits budget to it.
Step 03
Prove
We test it against real work, with your people, before it goes near production.
Step 04
Hand over
You get the thing, the documentation, and the training to run it without us.

What we do

Three ways we work

New Zealand businesses, startup teams, and people who want to learn this properly rather than be sold it.

BuildApps, SaaS tools and browser extensionsFrom first sketch to something running in front of real users.
TrainWorkshops and hands-on training for teamsFor organisations who want their own people doing this, not a vendor.
AdviseShort engagements for founders and product teamsDeciding what is worth building — and what is not.

From the lab

A warm glass attention matrix glowing diagonally through darkness — Attention, layer 2401

Clusters of translucent blue glass particles suspended in space — Embedding space02

A carved dark terrain descending toward an illuminated amber basin — Loss surface03

Rows of dark ceramic tokens with a moss-green path through them — Token trace, one query04

A branching metal search sculpture with one amber path highlighted — Search tree, depth six05

Frosted glass evaluation columns with one cyan result illuminated — Eval sweep06

Build

Start a project

Tell us the workflow that is costing you time. If AI is the wrong tool for it, we will say so.

Start a project

Train

Train your team

Hands-on sessions for teams and individuals, built around your actual work rather than generic slides.

Train your team

ThinkIt Through

We would rather explain it than sell it. If you have a problem you think AI might fit — or you are not sure it does — talk to us.