We use AI in our content. Here's what that actually means.

How We Actually Use AI in This Content

Let's be direct about something. This content, including this post, is produced with AI assistance. We're not going to dress that up or hide it in a footnote.

We think the right move is to explain exactly what that means, because vague disclosure is its own form of dishonesty. "AI-assisted" means nothing on its own. Here's what it actually means in practice.

What AI Does in Our Content Workflow

AI is a research and drafting tool in our process. Here's the specific role it plays.

Research and pattern recognition. When we're writing about NZ SMB AI adoption, we use AI tools to pull together relevant data points, identify gaps in public research, and map out what's already been said on a topic. That's faster than doing it manually. It also helps surface angles we might otherwise miss.

Structure and argument flow. We use AI to pressure-test whether an argument holds together. If we have a position, we'll often run it through an AI tool and ask it to find the weakest points. It's a faster version of having a sceptical colleague read a draft.

First draft prose. Once we have a clear brief, an angle, and specific points we want to make, we use AI to produce a first draft. That draft gets edited, often heavily. The prose that comes out of an AI tool is rarely wrong but it's often flat. It needs a human read to put personality and precision into it.

Fact-checking prompts. We ask AI tools to flag claims that need verification. They're not reliable enough to be the final check, but they're useful as a first pass.

What AI Does Not Do in Our Content Workflow

AI does not decide what we think.

Every position in this content, every claim, every recommendation, comes from people who have worked with NZ businesses on technology and operations problems. When we say "most NZ SMBs buying AI tools right now are solving the wrong problem," that's not an AI-generated opinion. It's what we've seen in practice.

AI does not know what a Wellington hospitality business actually struggles with when they try to implement a new booking system. It does not know what questions NZ accountants ask before they'll sign off on a new software spend. It does not know how NZ SMB owners actually talk about their businesses, what they're worried about, what language they use.

That knowledge comes from human experience. AI helps us write faster. It does not help us know more.

Why We're Telling You This

Partly because we think the current conversation about AI content is stuck on the wrong question. The debate about authenticity is fixated on production method. We think that's a distraction.

But also because we'd be in a contradictory position if we published content arguing for transparency about AI use while being vague about our own. That would be exactly the kind of hedging we argue against.

What This Means for Trusting Our Content

You should read everything we publish with appropriate scepticism. That's true of all content, AI-assisted or not.

The claims we make about NZ market conditions, AI adoption rates, and SMB pain points are grounded in real data and real experience. When we cite specific figures, we've checked them. When we state opinions, they're our actual opinions, not a language model's best guess at what sounds authoritative.

If you read something here and think we've got it wrong, we'd genuinely like to hear that. The only content that can afford to be unchallenged is content that nobody reads.

The Practical Question This Answers

Some businesses we talk to worry about whether their audience will care that AI helped write their content. The honest answer is: a small number will, and most won't.

What audiences respond to is whether the content is useful, whether it's clear, and whether it says something real. A reader who finds a post genuinely useful does not become less satisfied when they learn AI helped structure it. A reader who found a post vague and hedging is not going to feel better learning a human wrote every word.

The quality signal is the content itself. The production disclosure is just honesty.

The Version of This We'd Be Embarrassed By

Here's how this could go wrong, and why we're careful about it.

If we used AI to generate generic opinions, published them under our name, and presented them as hard-won expertise, that would be a problem. Not because AI helped with the prose. Because we'd be claiming knowledge we don't have.

The obligation on any content producer is that when you state a position, you actually hold it. When you make a claim, you've checked it. When you say "here's what NZ businesses should do," you've thought seriously about whether that's true.

That obligation doesn't change based on which tool helped you write it down. It sits with the humans whose name is on the work.

That's us. And we mean what we've written here.

Work with an NZ AI agency that's transparent about how it operates.