How We Do Social Media Management in New Zealand (and Why It Sounds Like You)

Social media management is broken for most small businesses in New Zealand. You either do it yourself (and run out of time by week three) or you hire someone who writes posts that sound nothing like you. The founder’s voice disappears. The content becomes generic. The audience can tell.

We built a different approach. Over the past month I published a four-part series explaining the AI systems we actually use: how to build prompts that sound like a person, how to give AI persistent memory, how to create reusable writing tools, and how to extract a business owner’s expertise into structured documents. That series wasn’t theory. It was the operating manual for how we run social media management for our clients.

This article connects the dots: what the system looks like when it’s running, and why the content it produces sounds like the person who built the business.

The standard social media management workflow is: someone asks you for a list of topics, writes posts based on your website, and schedules them. The posts are competent. They’re also interchangeable with any competitor in your space.

AI made it worse, not better. Now the same workflow runs faster, but the output is even more generic. Give ChatGPT a topic and a tone and you get content that reads like the average of everything on the internet. No opinions. No stories. No vocabulary that belongs to a specific person.

The problem isn’t the AI. The problem is that nobody gave the AI anything real to work with. A one-paragraph brand voice guide is not enough. “Professional but approachable” describes every LinkedIn profile on the platform.

If you want social media management that actually represents you, the AI needs to know who you are before it writes a single word.

Our social media management runs on five layers. Each one exists because generic content failed without it.

At the bottom is the prompting method. Every prompt we write embeds expert instructions and follows a structured pattern: context gathering, execution, coaching. This is what separates “write me a LinkedIn post” from a prompt that knows how to interview you first.

On top of that sits a persistent workspace for each client. It holds everything the AI needs: who you are, how you think, what your market looks like, what’s happening in your business right now.

Inside that workspace lives your knowledge base: six documents extracted from real conversations with you. Voice and positioning. Service delivery. Customer journey. Pricing logic. FAQ and objections. Brand story. Not what a copywriter guesses about your business. What you actually said, structured so the AI can use it.

The writing skills are reusable prompts for specific tasks: LinkedIn post writing, hook generation, email drafting, content repurposing. Each one is versioned, tested, and refined across clients.

The output is the social media content itself. Posts, emails, website copy. All generated against your knowledge base, in your voice, reviewed by a human before it goes live.

Most social media management starts with a questionnaire. Ours starts with a conversation.

We sit down with you and ask you to talk about your work. Not your strategy. Not your five-year plan. What happened yesterday. Who called. What annoyed you. What you’re proud of. Within 30 minutes, the business starts to reveal itself. Not the website version. The real version.

From those conversations, we build a voice profile. Not a paragraph that says “professional but approachable.” A structured document that captures:

• Sentence patterns: Do you open with a question or a statement? Do you use fragments? How long do your paragraphs run?
• Vocabulary: The words you reach for. The metaphors you use without thinking. The industry terms you use and the ones you avoid.
• Opinions: What you believe strongly. What you push back on. The hills you’ll die on.
• Stories: The anecdotes you tell repeatedly. The client wins you reference. The failures you’ve internalised.
• Emotional range: When you get fired up. When you get quiet. When you use humour.

Two conversations typically produce everything we need. You don’t fill out a brief. You don’t write anything. You talk, we extract, we structure.

Once onboarding is done, your knowledge base is built, and the project is set up, social media management runs on a weekly rhythm:

AI handles the parts that are slow for humans and fast for machines: pulling relevant context from the knowledge base, generating multiple angles on the same topic, maintaining consistent voice across dozens of posts. Jessie handles the parts that are slow for machines and fast for humans: knowing what feels right, catching tone shifts, adding the timing and nuance that turn a good post into one that gets shared.

Every week the system gets sharper. Posts that perform well feed back into the knowledge base. New topics surface from client conversations. The voice profile refines as we learn which phrases land and which don’t.

The same system powers content for NZ businesses across different industries. For NICH Cycling, an Auckland-based custom bike brand, it means social content that speaks to two completely different audiences: competitive triathletes who want specs and retirees who want a story. Same founder voice. Different angles. Both authentic.

For Manawa, a wellness and events business, it means content that captures the energy of live experiences without sounding like every other wellness brand on Instagram.

The pattern is the same every time: extract the real voice, build the knowledge base, produce content that the founder looks at and says “that sounds like me.” You can see how the founder relationship drives the output on my own LinkedIn, where we use the same system to produce content every week.

Publishing the infrastructure might seem like giving away the advantage. It isn’t. The system is transferable. The judgment from doing it dozens of times is not: knowing which questions to ask in the first interview, when a voice profile needs a third conversation, that the pricing logic document is where most businesses stall.

And the kind of NZ business owner who reads this and thinks “I should build this myself” usually realises they want to understand the system, not run it every week. They want to know the social media management is rigorous. Then they want someone else to do it. That’s the same pattern from our cold outreach case study: teaching beats pitching.

Onboarding is a conversation, not a form. We sit down with you and talk about your work. Two conversations produce the knowledge base. You don’t write a brief. You don’t fill out a questionnaire.

The content sounds like what you’d write if you had time. Not AI content you have to heavily edit. Not brand-voice content from someone who spent an afternoon on your website. Content that sounds like you on a good day.

It compounds. The knowledge base grows. The voice profile sharpens. The writing skills improve. Month three is measurably better than month one, because the system has learned from every post, every reaction, every conversation.

Technical founders in New Zealand who have deep domain expertise but no content pipeline. Your competitors with less depth are getting all the visibility because they post and you don’t.

B2B companies with smart people who should be visible. Teams with expertise that never reaches LinkedIn.

CEOs and VPs who know they should be building a public presence but don’t have three hours a week to figure out LinkedIn.

Two models: done-for-you (monthly retainer, we handle everything) or done-with-you (we build the system and train your team to run it). We’re based in Auckland and work with businesses across New Zealand.