In our first major financial services engagement, we walked in convinced we had it figured out. Salesforce expertise, a strong delivery model, an engaged client team. We were sure. We had not cracked anything.

The compliance overhead never showed up in the Statement of Work. It showed up in week six, when one audit-readiness conversation reshaped half the architecture. The data that looked fine for a CRM was nowhere near fit for grounding a regulated decision. And the people we built it for did not change how they worked just because the product was better. They changed when they trusted that the regulator was comfortable with them changing.

It took a lot of unglamorous work for financial services to become one of our strongest areas. That experience taught me the thing this whole repositioning rests on: most business transformations fail not because of the technology, but because companies scale faster than their operating model can handle. I know that because I have lived it.

When people hear AI-native, they assume we have fallen for the hype. The opposite is true. For us it does not mean selling AI. It means rebuilding how we advise and how we build so we own the outcome a client actually wanted, instead of billing them for the hours it took to get there. The technology does the heavy lifting underneath. The point on top is the same one we have always cared about: did anything real change for the people who have to live with the result.

Most agentic AI pilots in Australian financial services will not reach production. The model is rarely the reason. The reason is the integration layer nobody scoped, the data foundation that was fit for a CRM and not for grounding an agent, the context work everyone files as a backlog item that turns out to be the entire job. None of it is glamorous. None of it makes the keynote. All of it is the difference between a clever demo and a production system a CIO is comfortable signing off.

A while back I ran my first half-marathon. I told my partner on my birthday that I wanted to do it, my friends promptly called my bluff and set up a fund so the whole world knew, and I trained for months from barely being able to run 10 kilometres. Then I went to Bali on holiday and panicked. What if I had lost my fitness. What if I could not keep up when I got back.

My accountability buddy said one thing that stuck: trust the process, trust your training. The lesson generalises. At execution time, the work is preparation, not control. Doubt is a feeling, not evidence. The same holds for a decision like this one. The case for repositioning is not in how confident I feel on any given morning. It is in the engagements that already taught us the hard lessons. At some point you stop optimising and you back the training you have already done.

Here is the part of going AI-native I am proudest of, and it is the most human part. I want to tell you about Carlos, our Director of Data and Integration. While the AI hype cycle has been doing what hype cycles do, Carlos and his team have quietly built the data foundations and integration architectures that turn a clever demo into a regulated, audit-ready system that actually returns something. The AI work gets the keynote. His work gets the production rollout.

The hardest part of my job is restraining myself from claiming what people like Carlos build. I sell what he and the team make possible. The future of consulting is not the founder on the stage. It is the people doing the work the keynote forgets to mention. That is what human-led has to mean if it is going to mean anything, not a line in a brand deck.

Years ago, before any of this existed, Jannis and I sat near Martin Place with a coffee and asked what our non-negotiables would be if we ever built something, especially when pressure hit. We even imagined what 100 people might look like. It felt absurd then.

We are past 100 now, and I think about that conversation more, not less, because the decisions you make when nobody is watching become the culture everyone experiences later. I still wrestle with the same tensions: the fear-driven yes when the pressure rises, protecting standards while scaling, staying human when the stakes are high. Repositioning the firm has not removed any of that. It has just raised what is at stake.

We did not build this to chase a cycle. We rebuilt it so the work would still be worth doing on the other side of one.

So here is my question for any leader feeling the same pull. When the familiar version is right there and the harder version is the one you actually believe in, what does it take for you to back your own training and choose it anyway?