The Intelligence Age

AI, context, and New Zealand’s data center future

Full Show Notes:
https://www.theintelligenceagepodcast.com/843

Mark Smith speaks with Oliver Hartwich about how AI has changed since January, why agentic tools have become the real shift, and what that means for workflows, context, and cross-model collaboration. They also dig into New Zealand’s position in the global AI and infrastructure landscape, especially around data centers, energy, solar, batteries, and water use.

Key topics

• In this episode: Oliver says the biggest change in 2026 so far has not just been better models, but the rise of agentic AI tools like Claude Code, ChatGPT work modes, and Perplexity Computer.
• He explains why AI now feels less like a novelty and more like a practical, increasingly capable working layer, especially when multiple models are used together and checked against each other.
• Mark raises the concern that New Zealand may be vulnerable because AI compute and data may not stay in-country across the major hyperscalers operating locally.
• Oliver discusses the post Fable moment, when access uncertainty highlighted how dependent users have become on frontier tools and how quickly expectations shift across providers.
• The conversation turns to context management across platforms, including when to preserve long-term memory and when to use a fresh AI for independent review.
• Oliver describes building a personal Claude Skill from his own writing archive, compressing books, reports, articles, and newsletters into markdown so the system can reflect his style and past thinking.
• They discuss using one AI to peer review another AI’s work, including ping ponging a skill between Claude and ChatGPT to improve quality and completeness.
• Oliver shares how Codex can be used for computer control and debugging by going directly into system settings and config files instead of relying on manual UI hunting.
• The discussion moves to MCP-style integrations, including using Site CITE AI to connect large academic literature databases into AI workflows for faster research and self peer review.
• Mark and Oliver compare the old PhD research process with the newer AI-assisted version, where initial literature collation can be compressed from roughly a year to a few weeks.
• They reflect on the tradeoff between speed and thinking time, arguing that the slower, manual research process also created space for reflection, sleep, and deeper synthesis.
• The final section focuses on New Zealand data centers, geothermal energy, grid stability, solar adoption, battery storage, and the challenge of building infrastructure with social license.

Resources:
Oliver Hartwich's essay on AI in education: https://oliverhartwich.com/2026/06/25/bildung-and-the-machine/

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Data sovereignty, governance debt, and AI for Māori startups

Full Show Notes
https://www.theintelligenceagepodcast.com/842 

Amber Taylor joins Mark Smith to unpack Māori data sovereignty at the point of AI system design, with a sharp focus on what startups need to think about before they build. The conversation gets practical fast, moving from governance debt and investor pressure to the risks of using third party AI platforms with sensitive indigenous knowledge.

We discuss why governance has to be designed in upfront, how sovereignty changes when data touches external AI systems, and what an indigenous layer for AI could look like in practice.

Key topics

• Amber shares why the paper started with a startup lens, not just a sovereignty lens, because small teams do not have the same governance or architecture resources as large organisations.
• She introduces the idea of governance debt, arguing that AI governance added after a system is built creates the same kind of accumulated risk as technical debt.
• Amber explains why many AI governance frameworks miss the realities of founders and small teams who need workable, early-stage decision support.
• The discussion covers the limits of simply keeping data on Aotearoa soil if it still passes through external AI systems that the organisation does not control.
• Amber and Mark dig into the misconception that platforms are not training on user data, and why ticking a consent box is not the same as governance.
• They explore how investors can pressure startups to monetise data in ways that conflict with the original agreements made with communities and data holders.
• Amber shares how her own work with indigenous storytelling set clear protocols upfront, including ownership, attribution, and what happens if a company is sold.
• The conversation uses 23andMe as a cautionary example of what happens when DNA data becomes an acquired asset under new ownership.
• Amber outlines a future direction for an indigenous AI layer that could sit on top of any model, govern knowledge use, improve accuracy, and return reciprocity to source communities.
• The episode closes on the idea that New Zealand often builds first and regulation catches up later, creating opportunity for smaller AI companies to move quickly and responsibly.

Resources:

1. Native Sentient Website: https://www.nativesentient.com/
2. Link to paper we talked about: https://zenodo.org/records/21219022
3. Marae TV interview; https://www.tvnz.co.nz/tvseries/marae

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

AI, Shipping Faster, and the Future of Distribution

Full Show Notes
https://www.theintelligenceagepodcast.com/841

Mark Smith talks with Steve Mordue about how AI has changed the pace of building, shipping, and distributing software over the last six months. Steve shares what it looks like to move from being a seller and product visionary to building almost everything himself, why the last 5 percent is often the hardest part, and why distribution is now as important as the product itself.

They also get into Microsoft’s role in AI, the rise of open weight models, model routing, enterprise trust, and how sovereign data concerns are reshaping what companies can safely adopt.

Key topics

• In this episode, Steve Mordue describes the last six months in AI as a mix of speed, exhaustion, and concentration
• He says work is moving so fast that things built just months ago already feel outdated
• AI has not reduced the workload so much as multiplied the number of possible ideas
• Steve reflects on building Rapid Start CRM and what that taught him about scale
• He says the platform reached about 170,000 users
• Looking back, he feels it should have reached far more if marketing had been stronger
• We discuss the “last 5 percent” problem
• Mark raises the challenge of turning a near-finished product into something shippable
• Steve admits he is deeply affected by perfectionism and often obsesses over details others do not care about
• Steve explains why he now launches before perfection
• He says he recently forced himself to ship an AI app at 95 percent complete
• His view is that waiting for 100 percent means never launching
• We talk about how AI has changed his role from seller to builder
• Steve says his core strength has always been selling
• He now uses AI to write code, prototype, and build without depending on a large dev team
• He says the old dependency on teams and contractors has largely disappeared
• Every future Rapid Start product, he says, will be built by him alone
• His main limitation is no longer engineering access, but time
• They explore the value of domain expertise in the AI era
• Steve argues that people with deep knowledge in a narrow business area are best positioned to apply AI well
• He believes AI works best when paired with years of real industry experience
• Distribution becomes the central strategic problem
• Mark and Steve discuss how social platforms, email, and newsletters are losing effectiveness
• Steve says LinkedIn followers are usually peers, not buyers, which limits direct revenue impact
• They compare Microsoft’s strengths and weaknesses in AI
• Steve acknowledges Microsoft’s global distribution, infrastructure, and governance strengths

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Build AI That Works

Full Show Notes
https://www.theintelligenceagepodcast.com/840

Mark Smth talks with James Diekman about how Accelerate Tech has scaled, where AI is actually creating value in government workflows, and why most successful projects are targeted workflow changes rather than broad rollout tools. They also dig into Australia’s growing caution around AI governance, data residency, and the practical push toward local models and sovereign infrastructure.

We discuss how AI is being used inside engineering and marketing, why many Copilot-style deployments lose momentum, and how project management itself may evolve as agents take on governance, reporting, and coordination tasks. James shares what’s working in production, what keeps failing before launch, and where the next 6 to 12 months of AI adoption is heading.

Key topics

• James Diekman shares that Accelerate Tech has grown to about 32 staff in the last six months, driven by demand in government projects and AI-enabled solutions.
• The conversation contrasts AI use in engineering and marketing versus more passive consumption in functions like HR and finance.
• James explains that the strongest results come from embedding AI into specific workflows, rather than treating it as a standalone chatbot or a broad deployment.
• They discuss why many Copilot rollouts see mixed adoption and usage drop-off when the tool sits outside the daily workflow.
• James describes the most successful approach as identifying a business process, pulling apart a sub-workflow, and then applying AI, automation, or judgment-based reasoning to that narrow area first.
• He notes that many AI projects never reach production, and says that out of 30-plus AI projects delivered, only five or six have made it fully into production.
• The discussion turns to local government systems, with James outlining how his team often acts as the integration layer, or “the plumbers,” between older council platforms and newer software.
• They cover Australia’s increasing AI governance maturity, including state frameworks, federal requirements, and the rise of dedicated AI roles inside agencies.
• Mark and James debate public concerns around data centers, water use, and energy, with James emphasizing the need for better policy and the practical constraints of local compute.
• They explore the move toward local models, onshore hosting, and reserving compute capacity as organizations seek more control, lower risk, and better throughput.
• James makes the case that teams do not always need frontier models like Opus for every task and that model choice should match the job, cost, and risk.
• The episode closes on project management, where James outlines a “project brain” concept using agents, shared knowledge, registers, ticketing, and workflow automation to support or partially replace manual PM effort.

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Full Show Notes
https://www.theintelligenceagepodcast.com/839

Mark Smth speaks with Ashish Bhatia about how his move from Microsoft to Audible changed the way he thinks about customers, AI, and product design. The conversation centers on personal AI, running OpenClaw locally, and why owning your memory, workflows, and data matters more as AI systems become more capable.

They also explore Audible’s role in learning, the future of interactive audio, and how agents are beginning to replace clunky app-based workflows with more direct, personalized experiences.

Key topics

• Ashish explains the shift from B2B at Microsoft to B2C at Audible, and how that changes the way you learn from customers
• He shares why personal AI matters to him more than ever, especially when it involves health data, finances, and memory
• Mark describes building a large OpenClaw setup with 26 agents, including a nine-agent DevOps team and automated bug fixing
• They discuss why running AI locally teaches real systems thinking through failure, debugging, and repeated iteration
• Ashish talks about using OpenClaw to manage daily life, including lunch ordering through Grubhub as a mission-critical workflow
• The conversation highlights the importance of owning your memory and being able to move your AI system across machines and platforms
• They explore the idea of the death of UI, where agents increasingly bypass app interfaces and handle tasks directly
• Mark raises concerns about AI companies ingesting books and content, leading to a discussion of copyright, hypocrisy, and cultural attitudes toward books
• Ashish shares how Audible supports different learning styles, especially for commuters, slow readers, and people who learn better through audio plus text
• He outlines Audible’s direction toward interactive, multilingual audiobooks with AI-powered recall, discovery, and personalization

Resource Recommendations:

1. The Infinity Machine - Demis Hassabis - Amazon.com: The Infinity Machine: Demis Hassabis, DeepMind, and the Quest for Superintelligence: 9780593831847: Mallaby, Sebastian: Books
2. The Thinking Machine - Jensen Huang - The Thinking Machine: Jensen Huang, Nvidia, and the World's Most Coveted Microchip: Witt, Stephen: 9780593832691: Amazon.com: Books
3. Sapiens - Yual Noal Harari - Sapiens: A Brief History of Humankind eBook : Harari, Yuval Noah: Amazon.ca: Kindle Store
4. The Apple Podcast that I have curated - https://podcast.ashish-bhatia.com/feed.xml
5. Ashish Bhatia's website - ashish-bhatia.com

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Full Show Notes
https://www.theintelligenceagepodcast.com/838

AI is about to replace bloated SaaS, and most businesses are still only playing with chatbots

Sean G Muller says the next wave of AI is not about better prompts or prettier copilots. It is about rebuilding business around context, agents, and what actually creates value - before your software stack becomes the expensive middleman.
Mark Smth and Sean unpack why the last six months have been a genuine shift: agentic loops are now good enough to handle real business work, not just experiments. Sean explains how he moved from traditional technical architecture into building full application pipelines, MCP servers, and background agents that review email, track social signals, draft responses, and keep business moving without adding more human overhead.

You'll discover why context is the missing ingredient in almost every failed AI project, how Sean uses a simple meal-planning example to explain it, and why companies that scatter knowledge across laptops, SharePoint, Google Cloud, and people's heads are sitting on hidden risk. Sean also breaks down the difference between AI as a feature and AI as a business transformation engine, including the mistake many firms make when they bolt chat onto old workflows and call it progress.

We also get into the coming SaaS pocalypse - the idea that tools like HubSpot, Salesforce, Xero, Slack, and Atlassian may face a serious reckoning as businesses realize they can build leaner, custom, agent-first systems for less than the cost of endless licenses and modules. Sean shares how he built a headless, agent-driven CRM and why he thinks greenfield builds will replace expensive transformation projects much sooner than most executives expect. This conversation matters if you lead a business, run operations, own a small or mid-sized company, or simply suspect your current software is forcing you to work the wrong way. If you want to understand where AI is actually delivering leverage right now - and how to avoid wasting money on shallow pilots - this episode is essential listening.

Mark Smth hosts the conversation and brings the enterprise and product lens, pushing Sean to get specific about what success looks like for real businesses in New Zealand.

Sean G Muller is an AI and enterprise architecture specialist based in New Zealand, known for helping organizations build practical AI systems, implement agentic workflows, and rethink business process from the ground up.

Resources
1. The Cuckoo's Egg: Tracking a Spy Through the Maze of Computer Espionage - https://www.amazon.com.au/dp/0385249462?ref_=mr_referred_us_au_nz

2. Gemini: A Family of Highly Capable Multimodal Models — 2312.11805.pdf ⁠https://arxiv.org/abs/2312.11805

3. On the Measure of Intelligence — 1911.01547.pdf - https://arxiv.org/pdf/1911.01547

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Get featured on the show by leaving us a Voice Mail: https://bit.ly/MIPVM

AI adoption fails when it stops at access to tools. This episode introduces a practical model that bridges the gap between AI tools like Copilot and measurable business outcomes. The focus shifts from technology to people, addressing key blockers such as time, confidence, and relevance. By grounding AI in real workflows, tailoring role-specific use cases, and measuring outcomes, organisations can move from low usage to meaningful impact fast.

🎙 Full Show Notes
https://www.microsoftinnovationpodcast.com/837

👉 What you’ll learn     

• Identify the real blockers stopping AI adoption in teams 
• Shift from tool access to measurable business outcomes 
• Design role-specific AI use cases that drive daily value 
• Build AI training that fits real work, not theory 
• Measure AI adoption with practical KPIs and usage data 

✅ Highlights     

• "These two things are very, very different" 
• "Some things really need to happen in between those two places" 
• "This is not about the tech. This is about the people" 
• "They didn't know how to use the tool" 
• "They didn't know what they were allowed to do" 
• "When would they sit and just learn a link" 
• "To actually tailor use cases that are relevant" 
• "We have to measure how they are moving" 
• "We moved a department from 15% to 83%" 
• "It shifts how you are working" 

🧰 Mentioned     

• Microsoft Copilot (https://copilot.microsoft.com/)  
• Copilot Notebooks (https://support.microsoft.com/en-us/topic/get-started-with-microsoft-365-copilot-notebooks-0775e693-11c6-4d80-8aba-fcc81a737a06)  
• ChatGPT (https://chatgpt.com/)  
• AI Boost Academy (https://hannabergstrom.com/) [ 
• Vattenfall (https://group.vattenfall.com/) 
• Microsoft MVP YouTube Series - How to Become a Microsoft MVP - https://www.youtube.com/playlist?list=PLzf0yupPbVkqdRJDPVE4PtTlm6quDhiu7 

✅ Keywords      
ai adoption, copilot, generative ai, ai training, business transformation, ai use cases, digital productivity, ai skills, organisational change, ai strategy, workplace efficiency, ai literacy 

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Get featured on the show by leaving us a Voice Mail: https://bit.ly/MIPVM 
 
This episode dives into why large CRM and ERP projects keep failing and how AI is reshaping consulting, software delivery, and platform decisions. The core insight is simple. Organisations fail when they buy software instead of outcomes. With AI, experienced teams can move faster, strip away legacy complexity, and build only what the business actually needs. The conversation explores outcome-based thinking, flawed RFP processes, and why AI is accelerating the gap between great and average practitioners. 

👉 Full Show Notes
https://www.microsoftinnovationpodcast.com/836   

🎙️ What you’ll learn   

• How to frame technology decisions around outcomes, not vendors or features 
• Why traditional RFP processes drive cost blowouts and failed projects 
• The four core business problems most systems are really trying to solve 
• How AI tools are changing the speed and shape of consulting work 
• When building custom solutions now beats buying off-the-shelf software 

✅ Highlights 

• “So many people are still buying software like it’s a software decision rather than a business decision.” 
• “You’re really only trying to solve one of four business problems.” 
• “Most organizations aren’t even thinking in terms of what is the root of what I’m trying to solve.” 
• “Your highest qualified people are not even in the game because you’ve already excluded quality.” 
• “AI is now to the developer what Visual Studio was in the 90s.” 
• “It made the good developers really good and the bad developers a lot faster.” 
• “Low code was developed because we couldn’t get enough developers.” 
• “There’s a third option now, and that is vibe code your own add-on.” 
• “I think we’re less than 18 months away from this changing enterprise software.” 

🧰 Mentioned 

• Microsoft Power Platform: https://learn.microsoft.com/power-platform  
• Microsoft Dynamics CRM (Dynamics 365 CRM):
  https://www.microsoft.com/dynamics-365/solutions/crm  
• Microsoft Business Central:
  https://www.microsoft.com/dynamics-365/products/business-central  
• Microsoft Finance and Operations (Dynamics 365 Finance):
  https://www.microsoft.com/dynamics-365/products/finance 
• Notebook LM: https://notebooklm.google  
• LinkedIn Sales Navigator: https://business.linkedin.com/sales-solutions/sales-navigator 
• DynamicConsultantsGroup.com: https://dynamicconsultantsgroup.com 

✅Keywords   
ai in consulting, outcome-based software, crm failures, erp projects, power platform, low code vs pro code, vibe coding, rfp process, enterprise software, applied ai skills, digital transformation, consulting strategy

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Get featured on the show by leaving us a Voice Mail: https://bit.ly/MIPVM 
 
AI is shifting from a tool to an operating system, enabling individuals to move faster, build bespoke solutions, and dramatically increase productivity. The biggest advantage now comes from curiosity, not technical depth. Those who experiment, learn a few tools deeply, and treat AI as a collaborator can compress hours into minutes and unlock new business models. Meanwhile, large organisations risk falling behind due to governance, inertia, and slow decision-making. 

👉 Full Show Notes
https://www.microsoftinnovationpodcast.com/835    

🎙️ What you’ll learn 

• Use AI as an operating system to replace and create software 
• Move from idea to working solution in minutes, not hours 
• Build capability by mastering a few tools, not chasing every new model 
• Apply curiosity-driven workflows to unlock better AI outcomes 
• Avoid common enterprise pitfalls around governance and tool sprawl 

✅ Highlights  

• "This software rewards the patient and the curious." 
• "Learn two or three of these tools really well." 
• "Are you using these tools to improve your quality of life or aren’t you?" 
• "This is the last software that I’ll ever train." 
• "We can create bespoke applications through any of these tools." 
• "It’s eating other software." 
• "People who choose to be left behind." 
• "From idea to iteration in 30 minutes." 
• "I feel like I’m becoming a bionic person." 
• "The barriers to entry are just gone." 

🧰 Mentioned 

• ChatGPT – https://chatgpt.com/ 
• Bing (Copilot Search) – https://www.microsoft.com/en-us/bing/copilot-search 
• Google – https://www.google.com 
• Gemini – https://gemini.google.com/ 
• Claude – https://claude.ai/ 
• Copilot – https://copilot.microsoft.com/ 
• Perplexity – https://www.perplexity.ai/ 
• Stability AI (Stable Diffusion) – https://stability.ai/ 
• Ethan Mollick (Wharton profile) – https://mgmt.wharton.upenn.edu/profile/emollick/ 
• Co-Intelligence (book) – https://www.penguinrandomhouse.com/books/741805/co-intelligence-by-ethan-mollick/ 
• Suno (AI music generator) – https://suno.com/ 

✅Keywords 
ai tools, generative ai, chatgpt, ai productivity, solopreneur ai, ai governance, llm workflows, ai operating system, automation, business ai, ai adoption, digital transformation 

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith

Get featured on the show by leaving us a Voice Mail: https://bit.ly/MIPVM
 
This episode explores why data governance must come before enabling Microsoft 365 Copilot, with insights from Khurram Hafeez. It breaks down how sensitivity labels, data loss prevention, and Microsoft Purview reduce the risk of unintended data exposure. You will hear practical guidance on preparing your environment, protecting sensitive information, and managing AI use across Microsoft tools and third‑party AI sites. The focus is on real‑world decisions organisations must make to safely adopt Copilot at scale.

🎙 Full Show Notes
https://www.microsoftinnovationpodcast.com/834

👉 What you’ll learn     

• Why Copilot increases the impact of poor data governance 
• How sensitivity labels affect what Copilot can access and summarise 
• When to use DLP versus inline DLP for AI risk management 
• How organisations can limit data sharing with third‑party AI tools 
• What licensing choices matter for Copilot and Purview governance 

✅ Highlights     

• “With the intro of the AI, the importance of the data governance compliance has much increased.” 
• “The first defense layer that you should be apply, that is the sensitivity label.” 
• “If a user doesn’t have permission, Copilot can’t actually extract the data.” 
• “Standard policy actually is not sufficient these days.” 
• “Users copy and paste or upload a sensitive file to third‑party AI sites.” 
• “Inline DLP policy is one step ahead of endpoint DLP policy.” 
• “Microsoft focus not to block something, to protect something.” 
• “Before enabling the copilot, they ensure that their environment is ready.” 
• “Copilot certainly honors the permissions the user have.” 

 🧰 Mentioned     

• Microsoft 365 Copilot: https://learn.microsoft.com/microsoft-365/copilot/ 
• Microsoft Entra Suite: https://www.microsoft.com/security/business/microsoft-entra-suite  
• E5 licence: https://www.microsoft.com/microsoft-365/enterprise/e5  
• Pay‑as‑you‑go Purview licensing: https://learn.microsoft.com/purview/purview-billing-models 
• ChatGPT: https://chatgpt.com/ 
• Gemini: https://gemini.google.com/ 
• DeepSeek: https://www.deepseek.com/  
• Microsoft MVP YouTube Series - How to Become a Microsoft MVP: https://www.youtube.com/playlist?list=PLzf0yupPbVkqdRJDPVE4PtTlm6quDhiu7 

✅ Keywords 
microsoft copilot, microsoft purview, data governance, sensitivity labels, data loss prevention, inline dlp, endpoint dlp, ai security, m365 e5, third party ai, information protection, enterprise ai 

If you want to get in touch with me, you can message me here on Linkedin.

Thanks for listening 🚀 - Mark Smith