New Zealand is listed as available for self-service ChatGPT advertising in OpenAI’s Ads Manager guidance. That makes it a channel a qualifying NZ business can investigate now, rather than only a future announcement.

The useful question is whether it deserves part of your budget. Availability alone does not establish demand for your offer, a suitable audience or a profitable acquisition cost.

This article draws on OpenAI’s published guidance checked on 31 October 2026. It is not a report of M8 campaign results. The platform is in beta, and its capabilities and requirements can change.

What does a ChatGPT ad look like?

OpenAI describes an ad unit below a ChatGPT response, containing the advertiser identity, headline, description, landing page and image. Ads are labelled and separated from the answer.

Matching considers the conversation and the ad’s relevance. Advertisers can supply context hints, but these are not exact-match Search keywords. That makes a clear offer and a relevant landing page important parts of the brief.

The experience has its own eligibility limits. OpenAI says ads may appear on Free and Go plans and excludes accounts it identifies as under 18. You should not assume an ad can reach every ChatGPT user.

Can a New Zealand business advertise?

OpenAI’s country list says the legal entity advertising and being billed must be based in a supported country. New Zealand is on that list. Account creation, verification and ad review still apply.

Check the current eligibility and product rules in Ads Manager before planning a launch. An available country is not a guarantee that every advertiser, offer or category will be approved.

What does it cost?

OpenAI’s campaign guidance currently lists a minimum daily budget of NZ$25 for NZD billing. That is a platform minimum, rather than a suggested amount every small business should spend.

Buying options include impression and valid-click billing. Conversion optimisation does not mean that the advertiser pays only when a conversion happens. Check the billing model selected in the account.

Daily budgets are averages over the platform’s budget week. OpenAI says a day may spend up to twice the selected average, with the applicable weekly limit based on seven times that amount. At an unchanged NZ$25 average, that is up to NZ$50 in a day and NZ$175 over a full budget week. A campaign-total budget is also available for stricter overall spend control.

None of these settings tells you what an actual NZ customer will cost. We would need campaign evidence for that.

When might a test make sense?

Our view is that a test deserves consideration when the business has a clear offer, an appropriate destination page, a budget it can afford to learn with and a way to measure meaningful outcomes.

It may be worth exploring situations where customers ask detailed questions or compare options. That is a hypothesis to test for the specific business, rather than proof that conversational advertising will outperform another channel.

Before moving money, compare the opportunity with what your existing campaigns are doing. Keep an effective channel funded while deciding whether there is room for a controlled experiment.

What would we measure?

Start with spend, suitable visits, completed enquiries or purchases, and lead quality. Confirm the measurement setup before treating a dashboard total as a business result. Track links consistently and compare the same conversion definition across channels.

Set a review point and a decision rule before starting. What would justify continuing, adjusting or stopping? What would you need to learn even if the first test did not produce a customer?

A new platform can open an opportunity. It still has to earn its place alongside the Google and Meta options your business could use today.