We spent a week running AI workshops with ten Australian commerce businesses, back-to-back, Melbourne to Sydney. In the weeks' prior, the same conversations with nine New Zealand companies. Same frameworks, similar demos, much the same questions. Completely different conversations.
The pattern was simpler than expected: each market thinks it needs something different from AI, and both are half right.
Each market misjudges where it stands
NZ leaders consistently downplay what they've done. An agriculture company described their efforts as "gut feel and spreadsheets", then walked through a sensitivity analysis most organisations would kill for. A fitness brand said they were "looking to be inspired", having already completed an AI audit, run product feed enrichment, and applied for an AI-enabled checkout pilot. A retailer described himself as not having deep AI knowledge, while experimenting with open-source agent frameworks and pushing his board for increased investment. This wasn't false modesty. It was a genuine inability to benchmark their own progress, because the reference points don't exist in a small market. When nobody around them is talking about what they've done, they assume they haven't done much.
AU was the mirror image. Confident language, uneven reality. One distributor described active AI conversations while casuals were copy-pasting product data into their ecomm platform. A boutique retailer listed every AI tool the team used, then admitted their customer service bot lasted weeks before they shut it off. A major retailer had individual experimenters across multiple teams but no shared architecture, no measurement, and no credits allocated for bulk work. One beauty retailer had named their internal AI, built it sub-agents, and deployed it across the business in months. A fashion brand was planning to create organisational dependency on AI tools as a deliberate adoption strategy. The range in AU was wider than NZ, and the spread was wider than the language suggested.
The language tells the story
NZ spoke in constrained, careful terms throughout. "Runs on the board". "Don't boil the ocean". "Constrain this to something manageable". Even the most technically advanced NZ clients opened with deference: "tell us what good looks like" before revealing they'd already started building it. The language wasn't performative. It reflected a genuine caution about over-promising, rooted in smaller teams, tighter budgets, and less room to absorb a failed bet.
AU spoke in expansive terms. "Infinite opportunities, finite resource". When a live demo landed well, the immediate response was "how fast can we build this" rather than "what should we learn first". The vocabulary matched the posture: AU tends to overestimate its readiness, NZ tends to underestimate its own. That mismatch between language and reality showed up in several AU conversations once the questions got specific.
Each market buys something different
NZ buys confidence. The consistent ask tends to be education before action, strategy before build, permission before investment. A fitness brand wanted a roadmap before any engineering. A distributor wanted process mapping, not a technology pitch. The underlying need was validation that the direction was right before committing resource to it, which makes sense in a market where a wrong bet is harder to absorb.
AU buys proof. A sporting goods team wanted to quantify casual staff savings to the dollar. A fashion brand's head of technology described her goal as demonstrating enough value to unlock further funding for more sophisticated tools. The most technically advanced client on the trip didn't want more ideas. He wanted a repeatable operating model, because two previous vendor engagements had failed to deliver. The AU ask was frameworks, proof, and pace. They'd already committed internally, and now needed evidence to justify the commitment to the board.
"Show me" lands differently
NZ is comfortable with strategic conversation first. The workshops that resonated most in New Zealand were the ones that started with context, education, and possibility before any live build. NZ clients wanted to understand the landscape before seeing what a tool could do. Jumping straight to a demo felt premature. The best NZ sessions spent the first half on where the market was heading and only the second half on what could be built, and that sequencing built genuine buy-in.
AU wants the demo before the strategy. AU clients leaned forward when a prototype appeared on screen and leaned back during strategic framing. The live build was the credibility moment. Strategy without a working example felt theoretical. In several AU workshops, the energy in the room shifted entirely once something was built live, and only then did the strategic conversation become productive. The demo unlocked permission to think bigger.
Failure experience divides them
NZ clients, broadly, haven't shipped enough AI into production to have failure stories. In New Zealand, the opening ten minutes are spent explaining what's possible. Most NZ organisations are still imagining what good looks like, which means the conversation starts from a place of curiosity rather than caution. NZ needs to be convinced the investment is worth starting, not that this attempt will be different from the last one.
AU clients have scar tissue from AI that didn't work. A chatbot recommended the wrong product category, generating warranty returns. A beauty retailer built an impressive internal AI that hallucinates frequently enough to require analyst review before anything reaches the board. Multiple vendor projects with real business requirements came to nothing. The first ten minutes in Australia are spent acknowledging what went wrong before. AU knows what bad looks like because they've lived it, and that shapes the entire starting posture: they need to be convinced this time will be different.
Competitive urgency registers differently
NZ clients almost never mentioned competitors by name. Their urgency came from board pressure to explain revenue trends, government mandates to collaborate and find efficiencies across an industry, or a vague insurance-policy concern about being "the last ones in the race". One fitness brand captured the NZ posture perfectly: not wanting to invest so far ahead that returns don't materialise, but also not wanting the board knocking asking what they're doing. The competitive pressure in NZ is abstract rather than specific, which means it motivates differently.
AU clients name competitors, reference investor presentations, and benchmark AI search visibility against specific rivals. AU boards ask "how are we getting efficiencies into the business?", driven by investor and analyst pressure tied to market valuation. AU boards want receipts. The competitive urgency is personal: specific rivals, specific metrics, specific timelines. That creates momentum but also risks prioritising competitive response over strategic fit.
Board pressure serves different masters
NZ boards are governance-driven. NZ board pressure was about risk management, compliance, and responsible adoption. Two NZ clients are building formal AI governance frameworks before deploying any tools. Government mandates and industry body requirements have shaped the conversation in ways that never came up in Australia. NZ boards want reassurance that the organisation won't be caught out, not that it's winning a race.
AU boards are investor-driven. The questions AU leaders prepared for were about efficiency gains, margin improvement, and competitive positioning. Several clients described board decks where AI appeared as a line item tied to headcount optimisation or revenue growth. The pressure was external: analysts, investors, market valuation. That external pressure creates urgency, but it also means the AI conversation gets framed as a cost story rather than a capability story, which limits what gets funded.
Trust and data readiness split the two markets
NZ governs first, then decides whether to go. A head of technology explicitly pushed back on encouraging safe experimentation outside sanctioned tools, concerned about IP leaking into third-party platforms. NZ also admits messy data upfront. A wholesaler has no market share model across the industry. One company has assets dated 1900 by default because the actual date was never captured. The honesty about data quality was disarming, and it meant the conversation could start from reality rather than optimism.
AU governs first, then goes. Most clients moved past trust within minutes to talk timelines and ROI, then discovered their data problems mid-conversation. A distributor realised product IDs were wrong when talking through an automation use case. A startup revealed a +100-column spreadsheet for demand planning. AU treats governance as a gate to pass through quickly rather than a reason to pause, which means data quality issues surface later when they're more expensive to fix. The AU instinct to move fast often means the gaps get discovered at build time rather than planning time.
The vocabulary gap for what comes next
NZ doesn't have the vocabulary for the gap between productivity AI and engineered AI yet. Most NZ clients hadn't distinguished between using ChatGPT for individual tasks and building AI into their systems and workflows. The concept landed when explained, but it wasn't part of their existing frame. That vocabulary gap matters because it shapes what organisations ask for, what they think is possible, and how ROI is quantified. Without the language, the ask stays at "help us use AI better" rather than "help us build AI into our operations".
AU knows the difference, even if they haven't closed it. AU clients could articulate the gap between productivity tools and engineered solutions. They had language for it, even when they hadn't bridged it. The conversation was about how to cross from one to the other, which meant the workshops could start at a more specific point. Understanding the vocabulary means the conversation can move faster.
Talent scarcity shows up differently
NZ frames it as an existential constraint. NZ clients described a market where the talent simply doesn't exist locally. They can't hire their way out of the gap. That changes the conversation from "how do we scale the team" to "how do we make the team we have dramatically more capable", which is where AI amplification becomes existential rather than incremental. For NZ, the AI conversation is inseparable from the talent conversation.
AU frames it as a scaling problem. AU clients talked about needing more engineers, more data capability, and more specialist hires to move faster. The constraint was volume: not enough hands for the work they could see needed doing. AU has a deeper talent pool to draw from, which means the conversation stays operational rather than existential, but the scaling pressure still drives urgency around AI-enabled productivity.
Four lessons that showed up in both markets
The amplification gap is compounding. Across 19 conversations, it was obvious who had been learning and applying over the past twelve months and who was still at the starting line. Every month of delay isn't a month behind. It's a month where the people who started are building on top of what they've already learned. The organisations that started early aren't just ahead on knowledge. They're ahead on the patterns, the prompts, the failures, and the instinct for what works. That compounds in ways that are difficult to shortcut.
Integration is the bottleneck, not intelligence. The organisations struggling most aren't the ones without AI ideas. They're the ones without the plumbing to make those ideas operational. For AI to work in the real world, systems need to connect quickly and securely with repeatable patterns. Both markets had clients with strong AI ambitions stalled by legacy systems, fragmented data, and manual handoffs between platforms. The intelligence is available. The integration isn't.
Programmes beat pilots. The organisations making real progress treat AI as a programme with sequenced education, governance, tooling, and use cases with clear owners and milestones. Collecting pilots without a programme generates activity reports, not outcomes. This showed up identically in both markets: the clients with a structured approach were moving, the ones collecting experiments were spinning. A pilot proves something is possible. A programme makes it operational.
Speed without governance creates organisational drift. Both markets are moving faster than their oversight structures can support. The rule hasn't changed: a human in the loop, senior enough to understand what's being built, reviewing what the AI produces. NZ is more naturally cautious here, which slows them down but protects them. AU moves faster but accumulates governance debt. Neither posture is wrong, but both need adjusting before the gap between speed and oversight becomes a liability.
After 19 conversations across both markets in a few weeks, the pattern is evident: the gap between organisations that are moving and organisations that are planning to move is becoming structural. The time to start was six months ago. The second-best time is now.
Passionate about all things AI, emerging tech and start-ups, Mike is the Founder of The AI Corner.
Subscribe to The AI Corner
