New Zealand's government is now co-funding AI adoption for small businesses through the AI Advisory Pilot — a real, current programme, not a marketing angle. But a funded AI plan and a working AI agent are two different things. This guide covers what the pilot actually funds, and what it takes to go from a plan to something running in production.
What the pilot funds — and what it doesn't specify
MBIE's own language covers both ends of the work: businesses get support "to identify opportunities for AI and implement practical solutions in their day-to-day operations." What isn't spelled out publicly is exactly which suppliers can deliver the implementation stage, or whether funding can flow to a development partner outside the Regional Business Partner Network. If you're already working with a Regional Business Partner advisor, that's the right place to confirm the specifics before you commit a supplier.
The gap between an AI plan and a working AI agent
An AI plan tells you where AI fits and what to prioritise. It doesn't ship code. Most AI-focused providers active in New Zealand right now are positioned as consultants and automation specialists — strong on strategy, less often set up to design and build a production agent end to end. That's the specific gap this article is aimed at: what it takes to turn a funded AI plan into a shipped, working agent.
How we approach an AI agent build
99 Francs builds AI agents the same outcome-based way we build MVPs and websites: split into phases, no deposit, and you pay at the end of a phase only once you've seen it work — if you're still not satisfied after revisions, you owe nothing for that phase. For an agent specifically, that pairs naturally with sandbox validation: the agent proves itself in a safe environment before it touches anything that matters, and before you pay. The full process — feasibility checks, behaviour prototypes, tested action maps, provider fallbacks — is laid out in how we build AI agents.
Our track record, shown honestly
We haven't yet shipped a dedicated AI-agent project for a New Zealand client — we'd rather say that plainly than imply otherwise. What we do have, kept as two separate, real things: our AI-agent build process (above), and our New Zealand delivery record — 4 Madeira Place and The Pacifica Penthouse, both shipped remotely for the Auckland market in collaboration with Glasshouse Digital. If an AI agent is the next step for your business, weigh those two track records together, not a combined case study that doesn't exist yet.
What an AI agent build typically covers
- Customer-facing support agents that handle common questions and escalate the rest to a human.
- Internal ops agents that automate a specific repetitive workflow — triage, data entry, scheduling.
- Lead-qualification agents that pre-screen inbound enquiries before they reach sales.
- Agent-assisted internal tools built on your existing data and systems, not a generic chatbot bolted on top.
Getting started
99 Francs is not a registered Regional Business Partner provider, so we're not part of the pilot's advisor network directly. If you've already gone through the AI Advisory Pilot and have a plan, we can scope the first build phase around it. If you haven't, that's fine too — an AI agent doesn't need government co-funding to make sense, since the outcome-based model already removes most of the upfront risk. See AI agents and bots for the service, or web development for New Zealand if the agent is part of a wider website or product build.
Frequently asked questions.
Need this done instead of just read about it?
99 Francs is a subscription-based design studio: one flat monthly rate, unlimited requests, first delivery in 1–2 days. Start with pricing or book a free intro call.
