By Australian Financial Review
Publication Date: 2026-05-17 22:55:00
“Your data reflects your business processes and how your company actually works,” says O’Donnell. “If the data is bad, the decisions are bad. And fixing those mistakes later is expensive.”
As AI scales, data problems explode
Many organizations have been living with duplicate records, missing fields, and outdated information for years. These issues often remain hidden when data is reviewed manually.
AI changes that.
The automation works quickly. It’s not just a question of whether the data looks right.
“AI proof-of-concepts often look fantastic,” says O’Donnell. “The data used is usually a small, clean subset. Everything works perfectly.”
The problem comes later.
“When the same AI is deployed across the entire organization, it suddenly sees all the messy data. That’s when problems arise.”
The scaling trap
Take shopping as an example.
An AI system can find four supplier records that look similar. Two could be old. One may no longer exist. Someone else might never have done it…