Charting Our AI Course: TMNZ's Journey
By Eric Troebner
As Chief Technology Officer at TMNZ, I’ve had the privilege of being in the cockpit of our exciting artificial intelligence (AI) transformation. When we first started talking about AI around the time that ChatGPT was released, the conversations were filled with equal parts excitement and uncertainty. Today, I’m thrilled to share how we’ve turned that initial curiosity into some real, measurable impact across our entire business.
Building Our Launch Pad: Getting the Fundamentals Right
The temptation with any new technology is to dive headfirst into the flashiest applications. But our approach has been different from day one. We knew that sustainable AI adoption required rock-solid foundations, so we spent our first year building what I like to call our “AI launch pad.”
This wasn’t glamorous work, but it was essential. We completely overhauled our security framework, implementing stronger role-based access controls that ensure every team member has exactly the right level of system access. We also improved data classification and governance protocols. Every piece of company and client data remains securely within our infrastructure, and no AI model ever touches our sensitive information for training purposes.
The technical infrastructure we built gives our teams access to a carefully curated ecosystem of AI models, from OpenAI and Azure OpenAI to Anthropic Claude, Gemini, and Mistral. But we took it one step further: we created an internal AI portal where every single employee can safely experiment with these powerful tools. Whether it’s document analysis, image generation, or building custom agents, our people can explore AI’s potential within secure, compliant boundaries.
Perhaps most importantly, we implemented modern agent orchestration using platforms like n8n. This means both our technical and non-technical staff can create sophisticated automations without writing code. This launch pad has democratised AI development across our organisation.
Pre-Flight Checks: Education and Experimentation
Once our technical foundation was solid, we moved into what I call our “pre-flight” phase. This was all about preparation, education, and focused experimentation.
At a recent strategy all hands meeting we re-iterated an important truth for everyone: None of us are AI experts, and that is perfectly fine.
This humility is one of our secret weapons. Instead of pretending to know all the answers, we are working to create an environment where curiosity thrives. We run hands-on workshops and proof of concepts for every team, from operations to engineering to sales and marketing.
Each session focused on one key question: what AI application would make the biggest difference to your daily work? We actively encouraged teams to focus solely on what would make the largest impact in their daily workflows – no blue-sky thinking unless it solves their biggest bottleneck.
The results are remarkable. Teams didn’t just brainstorm abstract possibilities – they mapped out specific AI agents that could solve their biggest bottlenecks. Within a month of these workshops, over half our teams had working proof-of-concepts deployed. In just the first four weeks, we tracked over 200 hours of productivity gains across the company, having invested only about 20 hours to build these solutions.
Achieving Liftoff: Real Impact Across the Business
The transition from experimentation to production happened faster than we anticipated, and the results speak for themselves. Today, every business unit at TMNZ has secure AI tooling available to build into their daily workflows. Our product and engineering teams have accelerated discovery, analysis, development and testing, while our sales, marketing, finance, and customer support teams benefit from the first iteration of AI agents that handle everything from content creation to data analysis.
The numbers tell an impressive story. We’ve achieved an estimated 2 percent business-wide productivity uplift from our initial proof-of-concepts alone – equivalent to gaining a full-time employee. These aren’t just theoretical gains; they’re measurable improvements that allow us to do more with less and grow TMNZ.
To be able to speed up the delivery of practical results, proofs of concept at TMNZ are treated as if they are already in production. This allows us to quickly assess effectiveness, adjust on the fly, and confidently decide where to invest further resources and scale solutions.
Landing Safely: Governance and Responsibility
Success with AI isn’t just about speed and efficiency, it’s about doing things responsibly. Our “landing” phase focuses on sustainable scaling while maintaining the highest standards of ethics and governance.
We’ve embedded human oversight into every human-facing AI output – it’s part of our policy to protect both our clients and our reputation. Our practical and clear-cut AI policy ensures that only approved tools run on approved infrastructure, with clear guidelines around confidentiality and responsible usage.
Continuous education remains a cornerstone of this approach. Every new team member receives foundational training on AI terminology, data privacy, and risk management. We’re also building partnerships across organisations that are on the same journey as well as technology partners to enhance our governance frameworks and maintain our leadership in responsible Fintech AI.
Education and training is a true team effort and I am lucky to have highly capable and motivated members of my team and across TMNZ who deliver lunch & learn sessions, present to everyone the emerging AI phishing risks and champion innovative projects with partners.
The Journey Continues
What really excites me about our AI journey is that we’re just getting started. As part of our next phase, we are committed to sharing our learnings with the broader New Zealand business community because we believe AI’s greatest potential lies in collaborative exploration with our partners and customers.
For any business considering their own AI journey, my advice is simple: invest time in getting your fundamentals right, create space for genuine team-driven experimentation, and stay flexible as you learn. The destination matters less than building the capability to navigate whatever comes next.
Here is an outline of the topics we are going to share over the coming months:
- Practical AI Foundations: Policies, Security, Tech Stack
- Managing your AI Agents like staff members
- Shifting to agentic AI
- Avoiding the Number 8 Wire trap
- Practical low-cost AI setup for NZ SMEs
- AI for Accountants and Tax Professionals in NZ
To learn more about TMNZ’s technology innovation for clients, go here.