Written by Kiwi Dynamics Team 1What the AI Advisory Pilot actually funds
The pilot is deliberately not a technology grant. It does not buy you software licences or GPUs. It co-funds expert time: identifying where AI would actually pay back in your operation, and then implementing something practical in your day to day work. That framing matters, because it means the money is best spent on the diagnostic and delivery work most businesses skip. If you have already decided what to build, you are using the pilot for the cheaper half of the problem. The expensive mistake is building the wrong thing well, which is what a scoping engagement is meant to prevent. Our own view on this is set out in how we work.
2Who qualifies and how many places exist
Eligibility runs to New Zealand businesses that want to identify and implement AI in their operations, and the expanded pilot supports up to 150 of them. The number is the important part. One hundred and fifty places across the whole country is not a broad subsidy, it is a queue. Businesses that already know which workflow is costing them the most, and can describe it in a paragraph, will get through that queue faster than businesses arriving with a general interest in AI. If you are not sure which workflow that is for you, the time savings calculator is a blunt but honest way to put a number on it before you apply.
3Why 50/50 co-funding rewards a narrow scope
A 50/50 split changes the economics of scope in a way most people miss. On a $30,000 project you are funded to the cap and paying $15,000. On a $10,000 project you pay $5,000 and the return arrives inside a quarter. The second is almost always the better first move, because the thing that kills AI projects in small businesses is not cost, it is the gap between spend and evidence. A tight first build that demonstrably saves hours gives you the internal permission to do the next one. We size engagements this way regardless of funding: find the one workflow costing the most time or leads, ship it, prove the return, expand. There are worked examples of that in our case studies.
4What to spend it on if you only get one shot
If you get one funded project, spend it on the workflow that touches revenue or the phone. In practice that means one of four things for most New Zealand businesses. Missed calls and after-hours enquiries, which is what a voice AI agent or an AI receptionist is for. Quotes going out slowly, which is AI quoting. Paperwork that has to be read and re-keyed, which is AI document processing. Or knowledge trapped in people and files nobody can search, which is a knowledge base problem. Every one of those has a measurable before and after, which is what you will need when the pilot asks what the money achieved.
5What the funding will not cover
The pilot will not fund a general AI strategy document, and you should not want one. It will not cover ongoing licence costs, and it will not fund a project with no defined outcome. It also will not fix a data problem you have not acknowledged: if your job records live in three systems and a shared inbox, the first honest piece of advice any expert gives you is that the integration work comes before the AI work. That is not a reason to wait. It is a reason to include it in the scope you submit.
The wider context is worth holding onto. Generative AI alone is estimated to be worth up to $76 billion to the New Zealand economy by 2038, and 45% of businesses surveyed described AI as the biggest opportunity since the internet. Neither of those numbers helps you on Monday. What helps on Monday is one workflow, scoped tightly, half funded, running in production by the end of the quarter.
If you want help working out which workflow that is before you apply, get in touch. We will tell you honestly if we think the project is too small to be worth the paperwork, or too large to finish inside the cap.