Skip to content

Ideas

The smartest AI investment for Australian business isn’t the models

Anyone can bolt on a busted chatbot. The single most important AI investment is the tooling to measure whether any of it actually works.

As models get cheaper and better, the real upside is in being able to move quickly, argues Dan Nolan. Shutterstock.

If you’re an Australian business and you’ve actually rolled out AI to customers in some meaningful way, congrats. You’re in the minority. The vast majority of Australian businesses had someone tick the ‘enable Copilot’ box in Microsoft and went to the pub. Voilà. Just wait six more months and the productivity fairy will surely visit.

This is the issue with Australian business. Almost all of it is incredibly risk-averse and takes forever to make decisions. We are the land of the proud second, third and fourth mover. AI seems interesting, but I won’t touch it until it lands in the magic quadrant and I can point to half a dozen competitors who’ve shipped a busted chatbot on their website. Now we can act.

But even the companies that have moved face a problem that has nothing to do with which model they’re using or how they’re rolling it out.

It’s about how they’re measuring whether any of it works. Getting that measurement right might be the single most important AI investment an Australian business can make, because it’s the thing that stops you getting locked into a single provider and paying over market indefinitely.

Ideas is where we publish opinion and analysis from external contributors on the most important topics in the new economy.