AI skills versus AI experience is a question that sits under a lot of the data and AI leadership work we’re doing right now. A lot of our current client work is shifting through the noise of who has actually done the work and who has just rebranded themselves, because there’s a lot of that. Rebranding doesn’t have to be dishonest, but hiring leaders still need to test whether the underlying work changed or only the framing.
The chronology is where the rebrand shows
I’ve been recruiting data in the local market for over eight years, since before AI was at the forefront, which it now is.
A recent title change is a fair trigger for questions, but it’s not a verdict. A title change usually doesn’t come from nowhere: either the work changed, or the market moved and the person moved with it. The question is whether the work behind the words changed around the same time. The easy version is the person who put AI in the title in the last six months and can’t point to a project that predates it. The harder version is the person who can, and the rest of their record doesn’t back it up.
This is why a job ad now returns hundreds and hundreds of applications that all look pretty similar. They list the same activities, but the chronology exposes which ones were built towards AI and which ones had AI added to them later.
Ask what sits underneath the AI layer
The second thing to test is how the person talks about the layer underneath the AI. When you push past the model to the data underneath, a real data and AI leader can talk about quality, lineage and governance without losing the thread. I do a lot on governance, especially AI governance, because it’s becoming relevant for regulated organisations.
Ask a candidate to walk through how a model got trained on a past project, and then ask what happened before the training. Where did the data come from, who owned it and how was quality checked. The answer tells you whether the person was accountable for the data layer or only for the model.
Ask what they have stopped or killed
The third question is what they have stopped or killed. Ask what they shut down, what they decided not to build, or what they killed after it didn’t work. A person who can answer that is describing an operating history, and a person who can’t is describing a list of activities; this is the hardest one to answer without rehearsing because it’s where the accountability sits.
Frequently asked questions
The hiring board’s job is to test whether the underlying work changed or only the framing. If you’re hiring data or AI leadership in Australia and want to talk through how these get tested, start a conversation.
