Skip to main content
AI setup1–2 weeksFrom $1,500

Your AI is not stupid.
It is badly briefed.

Most "AI hallucinations" are context bugs, not model bugs. Give a model a folder of contradictory files and it will guess, confidently. Give it a proper foundation and the same model writes in your voice, to your buyer, with your facts. Have a play below.

Try it

You can't arrange your way out of a messy context.

Shuffle these notes as much as you like. While the files contradict each other and half are out of date, the model has nothing solid to stand on. Structure has to come first.

16notes
3contradictions
2out of date
0sources of truth
Drag the notes about all you like. Nothing improves until you structure them.
You
Model

Once it's structured, try breaking it. Drag any card out of the model and watch the answer fall over. That is what each piece is holding up, and why half a foundation is not a foundation.

What gets built

Five context sets. That's the job.

The five groups the cards sort into are not a metaphor. They are what actually gets written, structured and loaded into your Claude Project, custom GPT or whatever your team already uses.

01

Brand

Who you are, in a form a model can apply rather than admire. Named colours, defined type, the logo rules nobody can find. It stops inventing a house style every time it writes.

02

Voice & tone

How you sound, with examples of good and bad output annotated so the model can tell the difference. Plus the banned-word list, so nothing ships saying "leverage best-in-class synergies".

03

Product

What you sell, what it costs today, what it will not do. Dated, so a 2017 feature never turns up in a 2026 proposal.

04

Audience

Who buys, what pressure they are under, what makes them reply. The difference between copy aimed at everyone and copy aimed at the person actually reading it.

05

Messaging & proof

The approved lines, the real numbers, the case studies you have permission to use. So claims come with evidence attached instead of being generated to sound good.

+

Guardrails

The rules about what it may never invent: statistics, customer names, prices, features, quotes. This is the one that stops a plausible fabrication reaching a buyer, and the one nobody writes themselves.

What changes

Same model. Same prompt. Different brief.

Without context

"In today's rapidly evolving digital landscape, our best-in-class platform empowers forward-thinking organisations to unlock transformative value across the entire ecosystem."

  • Invents statistics that sound plausible
  • Sounds like every competitor you have
  • Gets your own pricing wrong
  • Written for nobody in particular
  • Needs rewriting before anyone sees it
With context

"Finance teams lose the 20th of every month to reconciliation. Acme cut that to a morning. Here is what it cost them and what changed."

  • Refuses to invent a number it cannot source
  • Sounds like you, because the rules are written down
  • Quotes current pricing, dated
  • Written to the person who actually buys
  • Usable on the first pass
How it runs

One to two weeks. Then it's yours.

Week 1

Gather and sort the mess

You send me what exists, however scruffy. Brand files, old decks, the pricing spreadsheet, the Slack thread where the tone got decided. I read all of it and find the contradictions.

Week 1

Write what's missing

Most of it is not written down anywhere. The voice rules, the persona, the guardrails. That gets drafted, you correct it, and it becomes the source of truth.

Week 2

Build and load it

Structured into the five context sets and loaded into your Claude Project, custom GPT or both. Set up so your team uses it without thinking about it.

Week 2

Test it against real work

We run your actual tasks through it, find where it still drifts, and tighten. You get a short guide for the team and the files themselves, which are yours to keep.

AI setup & brand context, from $1,500 +GST
Fixed price once scoped. Everything produced is yours on delivery. Ongoing content built on top of it is what the retainers do.
Set up my AI engine →

AI setup questions

Because it has nothing solid to work from. Given a folder of contradictory files, a model fills the gaps with whatever is statistically plausible, and it does so confidently. That is a context problem, not a model problem. Give the same model a structured foundation with explicit guardrails and roughly 80% of the strange output disappears.
It is the work of writing down what your business actually is, in a form an AI can apply, then loading it into the tools your team already uses. It covers five context sets - brand, voice and tone, product, audience, and messaging and proof - plus guardrails defining what the model may never invent. It goes into a Claude Project, a custom GPT, or both.
One to two weeks, from NZ$1,500 excluding GST, fixed once scoped. Week one is gathering what exists and writing what is missing. Week two is structuring it, loading it into your tools and testing it against your real work.
No, and almost nobody does. Most of what is needed has never been written down at all - the voice rules, the persona, the guardrails. Send whatever exists, however scruffy, and the gaps get drafted for you to correct. The scruffiness is the normal starting point.
Claude Projects and custom GPTs most often, because that is what teams already have open. The same context files work in Gemini, Copilot, or any tool that accepts a knowledge base and system instructions. The files are plain text and portable, so switching tools later does not mean starting again.
An explicit guardrails file listing what it may never generate: statistics, customer names, prices, features, quotes. Instead of producing a plausible number, the model says it has no source and asks for one. In the demo on this page you can watch that happen, and watch it break when the guardrails card is pulled out.
You do, on delivery. They are plain text and yours to keep, edit and take elsewhere. Nothing is locked to me or to a particular tool.
A prompt is per-conversation and has to be re-explained every time. Context is loaded once and applies to every conversation, for everyone on the team, including people who are not good at prompting. Prompt engineering makes one output better. Context makes every output better by default.
Next step

Want one of these for real?

Tell me what your team is using AI for and where it keeps going wrong. I will tell you what the context foundation would need to cover and what it would take. No deck required.