Google AI, built properly.
Maintain AI designs, builds and runs business systems on Google’s AI stack — Gemini, Gemini Enterprise, Google Workspace and Antigravity — plus the infrastructure underneath them. Google first, because for most New Zealand businesses it is the strongest all-round platform. When it isn’t, we say so before you spend.
Fixed scope, fixed price. No licence commission, no vendor lock-in pitch.
Why we lead with Google
Most AI projects fail on plumbing, not on models. Google’s stack removes the most plumbing.
One stack, end to end
Models, agent runtime, workspace apps and cloud infrastructure from one vendor. Fewer integrations to break, and one place to look when something does.
Meets your team where they work
Gmail, Docs, Sheets, Drive and Meet are already open on their screens. No migration project, no new app to adopt.
Enterprise controls, not consumer toys
Gemini Enterprise and Google Cloud give you access control, audit logging and data handling your IT people can sign off.
Costs you can forecast
Per-seat Workspace licensing plus usage-based Gemini API pricing. You can model the bill before you commit to the build.
What we build
Three ways we put Google’s AI to work — plus the part most AI consultancies skip.
Google Workspace automation
The unglamorous work that pays for itself: inbox triage, quote and invoice drafting, reporting packs, document generation, meeting notes turned into actions.
Private assistants on Gemini Enterprise
An assistant that answers from your own documents, with permissions that match your Drive — not a chatbot bolted onto a website.
Custom agents and tools
When the job needs real code: prototype in AI Studio, build with Antigravity, host on Cloud Run, ground it in BigQuery.
Bring us the task you think Google’s AI can’t do
One real business task, one hour live in Google’s stack. Acceptance criteria agreed in writing before you pay — that document is the only test. If Google’s AI cannot meet them, you get a full refund by bank transfer within five business days and we tell you which platform will.
Credited in full against any build started within 60 days. Four challenges per month.
How it works
Email hello@maintainai.co.nz with one paragraph: what you want done, what data is involved, and what 'done' looks like.
Before you pay, we send written acceptance criteria and constraints. If we already know Google can't do it, we decline and charge zero.
60 minutes, screen-shared, live. You watch it happen in Gemini, Workspace, AI Studio or Antigravity — not slides.
Written within two business days: what worked, what did not, what we would build, and alternative recommendations if Google is not the answer.
The refund, exactly
| Term | What it means |
|---|---|
| Price | NZ$495 per challenge, one task per challenge. |
| The test | The written acceptance criteria and constraints we agree before the session. That document is the only test. |
| Fail | If we do not meet them using Google’s AI products, we refund the full fee by bank transfer within five business days and give you a written recommendation of an alternative. |
| Pass | The fee is credited in full against any build engagement started within 60 days. |
| Declined up front | If we discover before the session that Google’s AI cannot do it, we refund before we start and tell you what can. |
| Limits | One task per client per month, four challenges per month in total. |
| Your rights | Nothing in this offer limits your rights under the Consumer Guarantees Act 1993 or the Fair Trading Act 1986. |
Where Google AI is the wrong answer
A specialist knows the edges of the platform. If your job sits in one of these, we tell you before you pay.
Air-gapped or fully on-premise
Google’s AI is a cloud service. If the data legally cannot leave your building, the answer is a local model, not Gemini.
Deep in Microsoft 365 or AWS
If your estate is Entra, Exchange and Azure, the integration cost usually beats the licence saving. We say so.
Specialist media pipelines
Frame-accurate video work, voice cloning at scale or unusual 3D pipelines usually belong with a dedicated tool.
Narrow, high-volume classification
At millions of calls, a small fine-tuned model can cost a fraction of any frontier API.
Four ways to work with us
Every engagement is scoped and quoted at a fixed price before work begins — no surprises, no lock-in.
Assess
- • Infrastructure audit
- • Automation opportunity map
- • Written recommendations, including where Google’s AI is the wrong tool
Challenge
- • One real business task, 60-min live session in Google’s AI stack
- • Acceptance criteria agreed in writing before session
- • Full refund if Google’s AI cannot meet them
- • Credited in full against any build started within 60 days
Build
- • Google Workspace automation
- • Private assistants on Gemini Enterprise
- • Custom agents on Gemini API, AI Studio and Antigravity
- • Infrastructure, monitoring and documentation
Maintain
- • Ongoing autonomous monitoring
- • Proactive maintenance and self-healing
- • Monthly reporting and roadmap
Built on Real Operations, Not Slide Decks
Vaughan Taylor is an AI practitioner and infrastructure automation specialist based in Hamilton, New Zealand. With deep experience across network infrastructure, cloud platforms and autonomous AI systems — including running 24/7/365 network environments at Spark NZ — Vaughan founded Maintain AI to bring the same operational discipline to AI-powered automation.
We work with New Zealand businesses, usually 5 to 200 staff, that want AI doing real work inside the tools their team already uses. We reduce manual overhead so your team can stop firefighting and get on with the work that matters.
Questions we get asked
Start with the free assessment
We audit your current setup, map what can be automated, and tell you honestly whether Google’s AI is the right tool for it. You get the written recommendations either way.