I didn’t really get AI for a long time. Not because I wasn’t interested, but because everything I tried either didn’t work very well or didn’t actually make my work better.
When ChatGPT was launched in early 2022 I was excited about what it might mean for our business. I started by asking it to create images for my blog posts and newsletters, which is always the hardest part of writing. The results were terrible. They were stereotyped and clumsy. When I asked for a non-European New Zealander, I was given images that looked like Māori from the 1800s. When I asked for that person to be in a city, I got the Sky Tower every time. It was so far from what I needed that I gave up.
I kept experimenting, though. I signed up for Otter.ai to help transcribe meetings. It struggled badly with New Zealand accents and couldn’t reliably tell the difference between female voices. In committee meetings it often decided that every woman speaking was me, which made it unusable. I stepped away again.
By this point, every time I tried to use ChatGPT it felt no better than Google. I did start using Perplexity, which gave more depth, but it still didn’t change the way I worked. So I stopped chasing AI altogether and went back to what we do well at Moneyworks, making Millie, our robotic process automation system, even more efficient.
Then in July 2025 something finally clicked.
We started using a transcription system called Marlow, which had been designed specifically for financial advisers. The difference was obvious straight away. The team behind it had actually spent time understanding how advisers work before building the product. It didn’t just save time. It produced better information and fitted naturally into our processes. We became raving fans very quickly.
That experience told me something important. AI could be genuinely useful, but only when it was built for the real world, not as a generic tool.
Over the following months I spoke with a couple of people who were using AI in a much more sophisticated way. What I learned was simple but transformative. The way you interact with AI fundamentally shapes what you get out of it.
Around the same time our industry body offered a twelve-week course on AI in financial services. Much of it focused on marketing funnels and social media, which isn’t really our focus, but the early sessions were excellent. They covered how to customise AI tools properly, how to give them enough context to work well, and how important it is to be explicit about the role you want the system to play.
That changed everything.
My ChatGPT is now called Susie. I work on multiple projects at once, and Susie helps me think through issues in a depth that simply wasn’t possible before. Not because she replaces my thinking, but because she sharpens it. I still use other tools when needed, and I use design platforms like EasyPeasy and Canva for images, but Susie has become my primary thinking partner.
I’ve also learned that discipline matters. Susie has very clear rules when she creates documents for me, particularly internal ones. She knows our formatting standards, our different working modes such as compliance, research, marketing and writing, and what matters to me when I’m making decisions. That means I don’t have to start from scratch every time. It’s like having a private operating system rather than a blank screen.
There are limits, though, and they matter.
If you don’t really understand what you’re asking about, AI can still get things wrong. I’ve corrected Susie more than once when we were researching our KiwiSaver white papers. AI can still hallucinate, and it doesn’t always follow instructions perfectly, especially when generating downloadable outputs. You need to know enough to challenge it and check its work. That human judgement doesn’t go away.
I now use AI across all my devices and pick up conversations wherever I left off. It has become part of how I work every day. Seeing the productivity gains first-hand has made it very clear to me that AI represents a fundamental shift in how businesses operate.
But it has also reinforced something else. The real advantage won’t belong to those who simply adopt AI the fastest. It will belong to businesses that combine it with deep knowledge, experience, and sound judgement. That distinction matters, not just for how businesses use AI, but for how we think about AI as an investment theme as well.
