I recently read a case study looking at a US$18B wealth management firm that used AI to identify clients at risk of leaving, then intervened earlier.
The result: a 15% reduction in churn.
The interesting part isn’t that AI predicted churn. Many models can do that.
What mattered is who owned the AI project.
In this case, it wasn’t the data team dictating models from the side. The business users (the people accountable for retention, revenue, client lifetime value) owned the KPI, drove the iteration, and dictated when and how the model connected into operations. The AI system became a business lever, not a technical experiment.
That is why it worked.
AI became meaningful because it was deployed where decisions actually happen: the engagement moments that determine whether a client stays or leaves.
What this means for New Zealand’s wealth industry
Some NZ wealth managers treat AI as a “tool” or a “pilot” owned by data or innovation teams. The value in this case study shows a different path:
- the business must own the outcome
- the KPI must be directly tied to a commercial lever
- AI must sit inside a live workflow, not a demo deck
- Don’t treat churn prevention as a one-off review, but as continuous intervention that protects value long before churn happens.
The real lesson
The core driver of success here was ownership rather than the model.
When the business owns AI – not just data or innovation teams – AI becomes a true commercial lever.
That is the takeaway New Zealand Wealth Managers should pay attention to.