Field Notes 7 June 2026 8 min read
Why AI-generated social content underperforms in NZ, and what to do instead
Generic content is invisible in a market where everyone knows everyone. A field report on what's landing.
Generic content is invisible in a market where everyone knows everyone. A field report on what's landing.
NZ’s social media active user base sits at roughly 3.4 million. Australia’s is about 20 million. The US has over 300 million. That scale difference isn’t interesting as a statistic; it’s interesting because of what it means for how content performs. In a market where your total addressable LinkedIn audience in a specific industry might be 8,000 people, the content dynamics are fundamentally different from a market where the same category has 800,000.
After running and supporting social media marketing programmes for NZ and Australian businesses across professional services, trades, and consumer categories, the patterns are consistent enough to write down. Most of what the generic social playbooks say is correct. Most of it also produces mediocre results in small markets because it was designed for scale.
AI-generated social content has a tell. Not always a visible one, but a structural one. It is optimised for form: the right hook structure, the right call-to-action placement, the right length. What it lacks is the specific texture that makes content land in a market where the audience is small enough to notice when nobody’s home behind it.
When a Wellington commercial property manager posts something generic about “market conditions in 2026,” it reads like ten thousand identical posts from everyone else in the category. The NZ professional community is small enough that people notice, consciously or not, when the content could have come from anywhere. When the same person posts about the specific challenge of CBD retail tenancy renewals in a market where vacancy rates are climbing but landlord expectations haven’t adjusted, that specificity reads as credibility. People who know what that problem feels like stop scrolling.
The gap between AI-generated content and AI-assisted content is not about the tools. It is about who provides the substance.
AI-generated means prompting a tool to produce five LinkedIn posts about your accounting practice. The output will be grammatically correct, correctly structured, and completely unusable in a market where the reader can tell you don’t mean any of it.
AI-assisted means a business owner speaks or writes a raw observation about something real: a pattern from a client call, a problem they have been thinking about, a change they have noticed in their market. AI takes that input and structures it into something postable with the right length, the right hook, and enough tension to earn a comment. The substance is human. The production is handled.
Across the accounts we run or support, posts grounded in specific local content produce two to four times the engagement rate of structurally similar generic posts. That ratio holds across industries and platforms and is consistent enough to treat as a baseline rather than an exception.
Instagram in NZ performs best when the visual is doing real work. Reels consistently outperform static posts in reach, but not always in conversion. The pattern that holds across accounts: a Reel with high organic reach from a genuine hook, followed by a static post or carousel in the same week with practical depth, produces the best combination of discovery and downstream engagement.
For professional services, authentic visuals outperform branded templates. Phone photos taken on location, real team moments, actual work in progress, read as signals of credibility in ways that polished studio photography sometimes doesn’t. AI helps with scripting Reels, structuring captions, and iterating on hooks. What AI cannot do is identify what the camera should be pointed at. That requires someone who understands the business and has a feel for what the local market finds interesting.
Facebook in NZ is still relevant for regional markets and demographic groups that are underserved by Instagram. A Hawke’s Bay vineyard, a Southland agricultural supplier, a Bay of Plenty building company. The content that performs here is practical and community-facing: project completions, team recognition, local event tie-ins. For consumer brands targeting NZ women over 40, Facebook regularly produces better return on ad spend than Instagram in categories where trust and familiarity matter more than reach. Clients who assume Facebook is universally declining in NZ are often surprised by what the numbers show when they actually look.
Short-form video has the most consistent upside of any format in NZ and Australian markets, and the most production friction. The reluctance to appear on camera is more pronounced in NZ than in markets where video-native personal branding is culturally established. Most NZ business owners will say they plan to start posting video, and most of them don’t.
AI avatar video removes that friction for businesses where the presenter is delivering structured information rather than selling their personal style. A weekly educational video rendered through an avatar and posted consistently builds visible content presence without requiring anyone to stand in front of a camera repeatedly. For professional services businesses where thought leadership is a genuine sales channel, that consistency over twelve months is worth more than a higher-production-value video made twice a year. An AI automation pipeline that converts rough scripts into rendered avatar content can get the turnaround from idea to posted video below four hours, which is what consistency actually requires.
The most common failure mode in NZ social programmes is not bad content. It is inconsistency. A strong month, three weeks of silence, a flurry of posts when something needs to be sold. Platforms respond to consistency, and so do audiences, particularly in small markets where absence is noticeable.
AI tools change the production economics without changing the substance requirement. The bottleneck has never been platforms or ideas. It is the time required to convert genuine thinking into postable content on a regular schedule. When production is assisted, a business owner who currently posts once a fortnight can realistically post three or four times per week without meaningfully increasing the time they personally spend on social.
Three posts per week over twelve months produces roughly 150 pieces of content indexed on a profile. At one post per fortnight, it is 26. The compounding is not subtle: 150 pieces of content reaching overlapping but distinct segments of a small, defined audience builds recognition faster than any single piece of content could, regardless of how good it is.
Follower count and aggregate engagement rate are the metrics most clients watch. They are also the least correlated with commercial outcomes in small-market social programmes.
The numbers that matter more: profile visits generated per post. This indicates whether content is making people curious enough to investigate the person behind it. A post that earns 40 likes and generates eight profile visits is performing differently from one that earns 20 likes and generates 30. Curiosity is the upstream signal for connection requests, website visits, and, eventually, enquiries.
Inbound enquiry attribution matters more than it gets credit for. Asking new enquiries where they found you, and tracking which reference social content, is coarse over small sample sizes but accurate across a quarter. In a market small enough that a single post can reach a meaningful fraction of your specific audience, that attribution is worth capturing.
Direct message volume from content is often the most underreported metric in professional services. The path from social content to a commercial conversation frequently runs through a DM rather than a link click. If content is producing private conversations with qualified people, that is a more meaningful signal than any public engagement metric.
The LinkedIn outreach and social content functions compound when they are coordinated: outreach opens the door, and consistent content presence in the prospect’s feed in the days after connection reinforces why they should pay attention. The combination works in ways that either function alone does not.
Opinion content is produced least and performs best. A direct, clearly held view on something happening in your industry, stated without excessive qualification, generates more discussion, more profile visits, and better-quality conversations than informational content in the same category.
NZ business culture tends toward understatement. That serves professional relationships well and social media poorly. A post that says “fixed-fee models in NZ professional services are overdue, and here is why they serve clients better than most providers are willing to admit” will outperform a post explaining what fixed-fee models are, every week, on every metric that matters. The first creates a position. The second creates content.
The pattern is consistent enough to be worth stating plainly: businesses that publish specific opinions from a clear point of view build recognisable presence in small markets faster than businesses that publish accurate but neutral information. AI drafts opinion content faster, structures the argument, and sharpens the hook. The view has to come from somewhere real. No tool can provide that.
For any NZ business taking social seriously in 2026: identifying two or three genuinely held opinions about your industry that most competitors are too cautious to publish, and starting to post them consistently with AI handling the production, is the highest-leverage starting point available. The alternative, waiting until there is time to do it properly without assistance, is the reason most NZ business social profiles look the way they do.
From automated guest comms to multilingual booking content, here's where the tools are earning their place.
The tools getting traction in real estate and property management look nothing like the early hype.
After a year of hybrid AI video production for NZ clients, here's what works, what doesn't, and where the economics change.