Is the AI Advisory Pilot a grant or a loan?

It is co-funding, not a loan. MBIE pays up to 50% of the cost of an approved AI advisory engagement, capped at $15,000 per business, directly to the advisory firm. There is nothing to repay, and you only ever pay your own share of the invoice.

Can I choose my own consultant under the AI Advisory Pilot?

Yes. You choose the advisory firm and include its quote in your application, but the firm must be registered with MBIE for the pilot. Ask any firm you are considering to confirm its registration before scoping starts, and remember that no firm can guarantee approval. That decision sits with the growth advisor assessing your application.

Who is eligible for the AI Advisory Pilot?

Three criteria are consistent across published sources: the business is registered in New Zealand, has been operating for at least 12 months, and has not received other government AI funding. Staff-count limits are stated inconsistently between published summaries, so confirm the current criteria directly with your Regional Business Partner before planning around eligibility.

How long does AI Advisory Pilot approval take?

Applications are assessed on a rolling basis rather than in funding rounds, and most decisions come back within three to four weeks. Once approved, the engagement must be completed within six months, so it pays to have your advisory firm and scope lined up before you apply.

Who is the Regional Business Partner for Queenstown and Otago?

Business South is the Regional Business Partner covering Otago and Southland, including Queenstown, Dunedin, Invercargill, Balclutha and Te Anau. They are the first call for confirming eligibility and starting an application in the region. The pilot runs until 31 January 2027 or until its 150-business quota fills, whichever comes first.

NZ AI Advisory Pilot: How It Works & How to Apply

New Zealand's AI Advisory Pilot co-funds up to 50% of AI advisory work, capped at $15,000 per business: up to $2,500 toward an AI roadmap and up to $12,500 toward implementation. You apply through your Regional Business Partner with a short proposal and a quote from an MBIE-registered advisory firm, before 31 January 2027.

That's the two-sentence version. The longer version is worth a few more minutes, because the summaries floating around contradict each other on details that matter, and at least one number being widely repeated doesn't trace back to any government source. Here's what's confirmed, what isn't, and how the application actually runs.

What the pilot is

MBIE launched the AI Advisory Pilot in January 2026 with an initial intake of 51 small businesses, and it framed that 51 as a floor, not a cap. Demand outran supply quickly. In May 2026 the government tripled the quota to a 150-business cap and extended the deadline a full year, to 31 January 2027.

The mechanics are simple. MBIE co-funds up to half the cost of a scoped AI advisory engagement, capped at $15,000 per business, split across two pools: up to $2,500 toward an AI roadmap and up to $12,500 toward implementing it.

It's co-funding, not a loan. Nothing gets repaid, and the approved share is paid directly to the advisory firm, so you only ever fund your portion of the invoice. The work must be delivered by an advisory firm registered with MBIE for the pilot, and the engagement has to complete within six months.

What the money is actually for

The two pools mirror how decent AI consulting works anyway. The roadmap money pays for the mapping phase: where the hours go in your business, which processes are worth automating, and in what order.

The implementation money pays for building whatever the roadmap says matters most. For a typical NZ small business that's the unglamorous admin: quoting, invoicing, customer follow-up, reporting, the jobs that quietly eat owner hours every week.

The deliverable MBIE expects is a documented AI plan for your business, followed by the implementation work. That's a useful filter when you're choosing a firm. If a proposal reads like a workshop series with no build at the end, be sceptical; I've written before about the strategy-deck trap in AI consulting and this funding deserves better.

Eligibility: what's confirmed and what isn't

Here's where this post earns its keep, because published summaries of the pilot genuinely contradict each other, especially on staff-count limits.

Three criteria show up consistently across every source I checked: your business is registered in New Zealand, it has been operating for at least 12 months, and it hasn't received other government AI funding. Plan around those with reasonable confidence.

Staff-count limits are another story. The summaries in circulation disagree by a factor of ten, and the ceiling most often repeated on consultancy blogs has no primary government source behind it that I could find. For context, the Regional Business Partner Network's general criteria published via business.govt.nz sit at fewer than 50 full-time staff, and that network is the channel the whole pilot runs through.

So do the boring thing. Ring your Regional Business Partner and confirm the current criteria before you plan around them, rather than relying on a blog post for eligibility, including this one. This piece is accurate as at early July 2026, and pilot rules move.

Is it still open?

Yes, as at July 2026. But the May expansion turned an open-ended floor into a 150-business cap, so places are finite and MBIE doesn't publish a live counter.

Your Regional Business Partner is the only reliable source on remaining capacity. One more reason to make that call early.

Rolling decisions also mean there's nothing to time. A complete application this month beats a perfect one in spring.

How to apply, step by step

Find your Regional Business Partner. The pilot runs through the Regional Business Partner Network, not MBIE directly. If you're in Otago or Southland, including Queenstown, that's Business South, which also covers Dunedin, Invercargill, Balclutha and Te Anau.
Choose an advisory firm and get a quote. You pick the firm, but it must be MBIE-registered for the pilot. Ask that question directly before anyone starts scoping.
Submit a short proposal. What your business does, what the engagement will cover, the outcomes you expect, and the firm's quote attached. This is a proposal, not a grant-writing exercise; short and specific beats long and impressive.
Wait for a rolling decision. There are no funding rounds to time. Decisions come back on a rolling basis, mostly within three to four weeks.
Complete the engagement within six months. Approved co-funding is paid straight to the advisory firm, so you only ever invoice-match your own share.

Before that first call, it's worth having three things roughly sketched: the processes eating the most hours in your business, what you'd want working at the end of the engagement, and a realistic budget range for your half. You don't need a polished document. A growth advisor can do far more with a clear one-pager than with a vague ambition to "do something with AI".

One honesty note before you shortlist anyone. No advisory firm can guarantee you funding, whatever their marketing implies; the decision sits with the growth advisor assessing your application. A firm that promises approval is telling you something useful about how it sells.

The co-funding maths, worked through

Say a scoped engagement comes to $20,000 NZD: a $4,000 discovery-and-roadmap phase and a $16,000 implementation build. At 50%, co-funding contributes $2,000 toward the roadmap (under the $2,500 cap) and $8,000 toward the implementation (well under the $12,500 cap). You pay $10,000 for $20,000 of advisory work.

Now the detail nobody spells out. The caps only bind at a $5,000 roadmap plus $25,000 of implementation, a $30,000 engagement all up, so drawing the full $15,000 takes a big project. Most small-business engagements will draw less than the maximum, and honestly that's fine; half of a right-sized project beats half of an inflated one.

If you run an accounting practice, I've walked through what a co-funded automation build looks like for a firm specifically, with this same maths applied to a practice's numbers.

Why the pilot exists now

Budget 2025 wound down Callaghan Innovation, and commentary at the time, the NZ Tech Podcast among it, traced part of this pilot's funding to those savings. That's the "why now": some innovation spend got redirected into direct, practical AI help for small firms.

Whether anything replaces the pilot after January 2027 is anyone's guess, and I won't pretend otherwise. The sensible read is to treat 31 January 2027, or the 150-business quota filling first, as a hard stop.

The takeaway

The pilot halves the cost of doing AI properly: map first, build second, delivered by a registered firm, done inside six months. And the application is genuinely light, a short proposal and a quote, with decisions landing in three to four weeks.

The step most businesses will skip is the first one. Ring your Regional Business Partner, confirm the current eligibility criteria, and only then start shortlisting firms. Ten minutes of confirmation beats three weeks of planning around a number nobody verified.

If you're weighing up whether there's a project in your business worth co-funding, a short call is the fastest way to find out what a scoped engagement would look like.