Kia ora! Welcome to New Zealand’s weekly roundup of AI news, events, jobs and education.
Some people have asked about the link between my work and The AI Corner, so here’s the clarity.
Alongside producing The AI Corner, I’ve incubated an AI services company inside Overdose Digital called Allexive, focusing on AI strategy, workflow and agent integration, and adoption across both internal operations and agentic commerce.
For transparency, The AI Corner stays independent and is not an Allexive sales channel. But if you do want to explore what delivering real outcomes with AI could look like in your business, you can reach me here.
Separately, thank you to everyone who signed up for The AI Corner’s ‘Go Beyond the Basics of ChatGPT & Copilot’ course.
We filled the cohort in just a few days. If you do want to join the next cohort, reply to this email to let us know of your interest!
This week’s highlights:
Why Copilot rollout isn't an AI strategy
Asa Cox: profitable AI growth for SMBs
Kiwi retailers deploy AI but won't trust it
My AI stack shifts again: ChatGPT at risk
88% of Kiwi firms will slow entry-level hiring
The $3 trillion agentic commerce shift forming
30-Minute AI Tool Challenge - Product Launch
Google's Gemini 3 vs OpenAI: two different bets
ACC uses AI to exit claimants; advocates alarmed
Happy reading and listening ✌️
Did someone forward you this? Sign up!
HEADLINE STORY
📊 KPMG: Kiwi CEOs back AI for growth but capability lags confidence
📣 Word On The Street: 40% of New Zealand CEOs plan to allocate 20% or more of their budgets to AI spending compared to 14% globally.
🔎 Zooming In:
70% expect agentic AI transformation vs 57% globally but 60% cite technical capability concerns.
Only 50% feel equipped to educate staff on AI vs 78% globally.
67% of New Zealand CEOs plan to increase headcount over the next three years (down on 92% globally), even as they modernise processes and operating models.
🏘️ Our Take: KPMG shows 90% of NZ CEOs feel their leadership has a clear view on how AI will disrupt business models. That confidence doesn't match conversations I'm having from enterprise through to SMBs. We're deep in the Dunning-Kruger valley. Most organisations think they have a strategy because they've embedded Copilot into email workflows. A third of CEOs admit their boards can't navigate AI adoption while allocating record budgets to it. That's institutional gambling, not transformation.
The headcount hedge exposes this. Global CEOs expect 92% growth. We're at 67%. We're planning workforce redeployment without questioning whether the work itself should exist. We're running processes faster rather than asking which deserve to survive at all.
We understand threats at the same rate as global peers but execute tactically, solving for noise instead of value. Meanwhile, competitors aren't just AI-native startups. They're international players, cashed-up PE-backed entrants, and local insurgents all asking "what becomes possible when intelligence is ambient and workflows are autonomous". Incumbents are asking "how do we make existing business faster". The question is whether we're willing to fundamentally reimagine how value gets created before someone else does it for us.
💼 Uncut: 30-Minute AI Tool Challenge - Product Launch
→ AI gives me a random task. I get 30 minutes. Only AI tools.
Challenge #2: Launch a new adult beach version of the classic Kiwi game Four Square (“Square Up”) on the Slideshow podcast with Dave Hayward at Europa Creative Partners.
Tools: ChatGPT, Comet, Reve, Nano Banana, Veo 3, Gamma, Wispr Flow, Claude, Gemini.
These aren’t polished walkthroughs. Lots of failures at pace to show how fast you can get from zero to a working concept when you stack AI tools.
💼 Business & Industry
Kiwi retailers fight Temu with AI, but won't let it talk to customers. New Zealand retailers are deploying AI to compete with overseas platforms like Temu, Amazon and Shein, but most refuse to let algorithms handle customer service unsupervised, revealing deep trust issues even as adoption accelerates.
3-min read.
Our take: Kiwi retailers understand something the overseas players forget forgot: in a market our size, losing the human touch means losing the only advantage we have against algorithmic giants who'll always win on price and speed.
Why 84% of enterprises see telcos as their primary AI partners. Accenture identifies Spark and One New Zealand as regional AI leaders, investing in autonomous networks and agentic AI to transform from infrastructure providers into AI aggregators, with 84% of enterprises viewing telcos as primary AI partners.
4-min read.
Our take: The irony of telcos becoming AI kingmakers, after spending decades as "dumb pipes" losing value to the platforms they carried, shows how infrastructure advantage still matters when compute moves to the edge and every enterprise needs low-latency AI that understands network topology, something Spark and One NZ have figured out how to monetise.
Why 88% of Kiwi firms will slow entry-level hiring in three years. Research shows 34% of New Zealand companies have already slowed entry-level hiring due to AI, with 88% expecting to continue, creating a talent pipeline crisis as pathways to senior roles disappear.
3-min read.
Our take: The economics here are backwards: AI-native graduates represent arguably the highest ROI hire for many NZ SMEs: unencumbered by legacy processes, fluent in new tools, delivering senior-level output at junior salaries. Instead, we're opting to not hire for these roles in-place of holding a stack of experienced campaigners, yet the narrative from the economy is to complain about skills shortages? New Zealand's small economy can't import enough senior talent to replace the career ladder we're demolishing, and the sinking lid effect means business’s will slowly erode their own future capability and create a generation of graduates whose only option is Australia.
Award-winning Kiwi photographer, Yvonne Todd, ditches camera entirely for AI. One of New Zealand's most distinguished photographers, Yvonne Todd, has abandoned cameras after three decades to work exclusively in AI-generated imager, calling it a "lament" for photography as we know it.
19-min listen.
Our take: Todd's honest ambivalence (calling AI "slop" while using it daily, describing it as "bingeing on junk food”) captures exactly where creative industries are right now: economically compelled to adopt technology that makes them professionally queasy.
Your performance review is measuring the wrong things in the AI era. Christian Yao's research shows performance reviews still evaluate individual productivity when capability now resides across human-algorithm partnerships, creating strategic blind spots as three-quarters of knowledge workers adopt AI.
3-min read.
Our take: New Zealand's productivity problem won't be solved by workers using AI faster, it'll be solved when we figure out how to measure, reward and develop the uniquely human judgement that makes AI useful not dangerous.
⚖️ Government & Legal
ACC is using AI to decide who goes back to work; advocates call it "shocking". ACC has deployed AI to identify long-term claimants who should return to work, aiming to exit 11,675 people by mid-2026, but advocates warn the technology can't distinguish between simple sprains and complex misdiagnoses.
2-min read.
Our take: When a Crown entity with explicit cost-reduction targets deploys AI to flag people for exit, calling it "helping injured people return to independence" isn't care improvement, it's automation giving bureaucratic cover to a political mandate to cut the claims pool.
⚖️ Society & Education
Christchurch Tech used for AI skin scanning catches melanomas before you can see them. Skin Institute's SkinTracker uses AI-powered full-body imaging to detect skin changes invisible to annual naked-eye checks, targeting New Zealand's world-leading melanoma rates with algorithmic surveillance.
3-min read.
Our take: In a country where our outdoor lifestyle creates a skin cancer epidemic, AI that industrialises what was previously artisanal specialist observation is exactly the preventive healthcare infrastructure we need.
How Nelson's library is saving non-profits $20,000 a year with free AI training. The Nelson AI Sandbox's free training programme has taught 300 people from 150 regional non-profits to use AI effectively, with participants reporting up to $20,000 annual savings from bringing outsourced work in-house.
3-min read.
Our take: Big fan of the genuinely community-led capability programme approach that understands technology adoption fails without addressing the resource and knowledge barriers that exclude everyone outside the big city innovation postcodes.
📚️ AI for Business
Helping leaders and teams adapt, learn, and scale with AI.
1️⃣ Google's Gemini 3 vs OpenAI: Two different bets on AI's future. OpenAI builds an exceptional assistant you talk to. Google and Anthropic build infrastructure you work through. One optimises for conversation, the other for orchestration across your existing tools.
1-min read.
2️⃣ My personal AI stack update in November: AI platform monogamy is over and my ChatGPT subscription is under threat. ChatGPT, Claude & Gemini are drifting apart, and my user behaviour shifts again.
4-min read.
3️⃣ Rolling out Copilot or Gemini isn't an AI strategy. Most businesses tick the box but six months later nothing has changed. Why: no commercial objective, no capability development, no process integration, no plan for freed-up capacity.
4-min read.
4️⃣ NEW SEGMENT: Mike’s Musings | The $3 trillion agentic commerce divide is forming now. AI agents read APIs, not websites and most businesses aren’t prepared for the shift in how they get found by and engaged with by AI.
🎥 Watch & 🎧 Listen
5️⃣ You're not losing your job to someone who uses AI. You're losing it to someone who thinks about outcomes with AI. The threat: over-attachment to process (the how) over results (the why) in your work.
4-min read.
🎙️ The AI Corner Podcast
This week's guest is Asa Cox, Founder and CEO of Arcanum AI. Hear:
Why data is no longer the blocker and how a single PDF now delivers more value than gigabytes used to
How Arcanum built Numa to serve the forgotten SMBs in manufacturing, construction, and engineering that legacy tech overlooked
What changed when AI moved from "you need massive datasets" to "bring your own single cloud instance with pre-built apps"
🌍 Tech Updates From Global
The selected top headlines from each major AI tech company.
OpenAI
GPT-5 mathematical breakthrough claim exposed as false after model simply identified existing solutions rather than solving new problems.
Paid Microsoft $866 million in revenue share during first three quarters of 2025, reflecting 20% agreement on total revenues.
Unveiled Gemini 3 achieving state-of-the-art reasoning with 1501 Elo score on LMArena, featuring breakthrough multimodal understanding capabilities.
Launched Antigravity agentic coding platform combining context from browser, terminal, and editor for autonomous multi-step task execution.
Introduced Canvas tools that generate interactive interfaces on-the-fly, trading cliché responses for genuine insight and deeper logic.
Integrated Gemini 3 across products including Gemini app (650M monthly users), AI Overviews (2B monthly users), AI Studio, and Vertex AI.
Anthropic
Claude now available in Microsoft Foundry and Microsoft 365 Copilot.
Secured up to USD $15b from Microsoft and NVIDIA to diversify cloud computing away from AWS dependency.
Rolled out Claude's memory feature to Max users, allowing assistant to remember previous conversations for personalised interactions.
Amazon
Major AWS outage disrupted Amazon and major platforms for hours following internal DNS failure in US-EAST-1 region.
Launched AI translation feature for self-published Kindle books, enabling authors to reach broader international audiences.
Plans to automate 75% of operations, avoiding hiring 600,000 US workers by 2033 through robotic automation.
Introduced Creative Studio AI tool for mid-market brands to produce ads across Prime Video and media properties.
Signed $38 billion partnership with OpenAI to provide extensive AWS computing infrastructure through 2032.
Microsoft
Nvidia and Microsoft pour USD $15b into Claude maker Anthropic amid AI bubble talk
Ownership stake in OpenAI dropped from 32.5% to 27%, now valued at approximately $135 billion following recent funding.
xAI
Announced delay in Grok 5 launch to early 2026 for model scaling and hardware infrastructure improvements.
Released Grok 4.1 emphasising emotional intelligence and creative writing enhancements over raw reasoning advancements.
NVIDIA
Invested $250 million in generative AI music startup Suno, signalling growing interest in entertainment sector applications.
CEO Jensen Huang warned China is 'nanoseconds' behind US in AI development during competitive landscape interview.
📅 AI Events in New Zealand
17 AI events happening across the country this week.
This week’s featured events:
Autohive beginner AI Agent Cooking Class - this Friday in Auckland. A hands-on 90 minute workshop for ChatGPT users who want to build their first custom AI agents. A good chance to get practical, build something real, and leave with working skills you can apply straight away.
📅 Promote your event with us. Reply to let us know.
💼 AI Roles Around Aotearoa
Picklist of 🌶️ HOT 🌶️ new roles in AI this week
AI Consultant, Deloitte: Auckland
Senior Data & AI Scientist, Lloyds Banking Group: Wellington
💼 Promote your job with us. Reply to let us know.
🤦 ️ AI Funny Of The Week
We all love AI, but it’s certainly far from perfect 🤔 …
👋 Mike & Erin