AI Instinct: The Most Important Professional Muscle To Build This Decade

If your AI use only looks like this...

You’re still at the starting line. That’s useful, but it’s surface-level productivity. You’re querying AI, not leveraging it. But you're missing an AI Instinct.

New Zealand loves talking about AI. We quote the $76 billion opportunity like it’s guaranteed. Yet most of the 38% of working Kiwis “using AI” are stuck right here. They think they’re ahead, when in-fact there's a large gap to close.

AI instinct is becoming the new professional reflex, the difference between using technology and using it to win. It will define who stays competitive, from individuals to entire economies.

Microsoft and Accenture estimate AI could add $76 billion to New Zealand’s economy by 2038. But 68% of SMEs have no plans to invest. Their reasons: “no understanding” and “no clear value”. Meanwhile, 80% of large firms are already scaling. New Zealand ranks 21st globally for adoption.

To capture that potential, New Zealand needs more than awareness. We need AI instinct across three levels of competitiveness.

1. Global Competitiveness

While other nations train AI-fluent workforces and build compounding business models, we’re still learning to prompt. Each month of hesitation widens the gap. Fast followers still have to move, yet we're sitting in the bucket of slow followers.

2. Local Competitiveness

Large firms are scaling AI. Most SMEs, which employ most Kiwis, aren’t. The biggest productivity gains sit in the industries adopting slowest. If SMEs don’t move soon, they risk being locked out of the next decade of growth.

3. Personal Competitiveness

By 2027, “I use ChatGPT” won’t matter. The question will be, “What have you built with it?”, AI doesn’t replace top performers; it amplifies them. Those without AI instinct risk being automated or replaced. Those who build it become exponentially more valuable to their companies and to the country.

AI instinct is tactile. It’s the muscle memory that comes from building, breaking, and fixing with AI. People with it see the world differently. They don’t just use AI tools; they focus on architecting them into their work.

Their skill spans four interconnected layers that build on one another:

Each layer compounds. Together, they form instinct: the reflex that lets people move faster, solve smarter, and build systems that run without them.

The Pattern Recognition Layer

1. Spotting automation opportunities in daily workflows: Where most people see routine tasks, those with instinct see possibilities. Watching someone copy data between systems isn’t “just how it’s done”: it’s a 10-minute Make.com workflow waiting to happen.

2. Knowing which AI approach to apply: Is this a chatbot, a copilot, or a simple automation? People with instinct understand which tool fits which problem. They know when a quick Zapier connection is perfect and when something more advanced is worth the effort.

3. Abstract problem-solving with AI: Beyond content generation or email writing, they use AI for mapping processes, designing decision trees, or architecting workflows: things most wouldn’t think to delegate to AI.

The Tool Fluency Layer

4. Tool literacy: They instinctively know what to reach for: Make or n8n for API connections, Autohive for agents, and Claude Projects or Custom GPTs for persistent context. This fluency only comes from hands-on experience.

5. Rapid tool assessment: New tools drop every week. Some are transformative, others all promise and no delivery. Those with instinct develop a sense for what’s real and what’s hype. They can spot the difference between genuine progress and “agent washing”. Moving beyond Copilot or Gemini to test and experiment is critical to keep the tools sharp.

6. Failure literacy: AI fails often, models hallucinate, workflows break, integrations crash. People with instinct have learned from those failures. They understand why things go wrong and how to work around the limitations.

The Execution Layer

7. Confidence to experiment without perfect knowledge: They lean toward experimentation and iteration rather than waiting for perfect certainty. They don’t need to understand the internals of large language models to build something useful, they just start, test, and refine.

8. Muscle memory of working with AI: AI becomes quiet infrastructure. It’s reflexive. They hit a roadblock and their mind automatically asks, “Could AI handle this?”.

The Judgement Layer

9. Distinguishing real value from marketing hype: They can see through grand claims about “autonomous agents”. Their question is simple: Who creates the flowchart, and when? They’ve seen enough cycles to separate capability from theatre.

10. Responsible AI practice: They understand privacy, bias, and hallucination risk, not because of compliance seminars, but because they’ve built systems and seen where things can go wrong.

The critical point: this instinct develops only through hands-on work.

You can’t learn it in theory. You build it by rolling up your sleeves: automating, testing, debugging, and shipping small systems that actually run.

MIT economists David Autor and Neil Thompson recently studied how automation changes the value of work. Their findings show a clear pattern that explains why some people become more valuable with technology, while others become replaceable.

They found two main effects:

1. The Bookkeeper Effect: Expertise Concentration. When automation removes the simple tasks, what’s left demands more expertise. Bookkeepers are the perfect example. From 1980 to 2018, computers took over most of their repetitive data entry. Employment dropped by a third, but those who remained earned 40% higher wages. This was because the work that stayed required judgement, problem-solving, and financial insight: skills that automation can’t easily copy.

2. The Taxi Driver Effect: Expertise Commoditisation. When automation replaces the expert tasks, the job becomes easier to do, and more people can do it. Think of taxi drivers before GPS. Local knowledge of every shortcut was valuable, it separated experts from amateurs. Then GPS came along, and that expertise became irrelevant. Suddenly anyone could drive and earn a living from it. Employment doubled, but real wages fell by 13%. Uber exacerbated this challenge. The value of the taxi medallion was effectively democratised at a very low cost.

The question for every knowledge worker is which path are you on?

If you can't demonstrate AI instinct, you're competing against people who can do 3x your output. And "I use ChatGPT sometimes" won't cut it as a resume line. The gap between "has instinct" and "doesn't have instinct" is widening every month, and it's going to be brutal on a large portion of the workforce.

There's no playbook apart from getting in the arena and using the tools. You don’t learn instinct in theory. You build it by doing. By automating small things, breaking systems, fixing them, and building again. The goal isn’t perfection (at least half of of what I build after hours fails first time around!), it’s momentum. And learning what to do about it is where an AI Instinct gets built. Here are 8 ways to build an AI Instinct:

AI instinct doesn’t emerge from reading about transformation. It’s forged in these small experiments, the ones that move you from “I use AI” to “I build with AI”.

AI instinct isn’t an individual skill or a company project. It’s both. It’s a muscle that only grows through consistent use, and New Zealand can’t afford to skip the workout.

Those building now will define the next decade of work, productivity, and competitiveness. The rest will still be talking about it.