How much was AI in the development of Speedo

IQ? So yeah, we really tap into the machine learning aspect of it.

But in the context how we train, we effectively spend a lot of time making sure that the data that we capture to train the algo is as pure as possible.

So that, in a really simple form, is we would agree like a protocol.

So let's say you're doing 50 free, then 50 backstroke, 50 fly.

We have cameras, we have additional sensors outside of the goggles that we do.

And literally us at the end of the lanes with old fashioned clipboards, really making sure that if someone goes down that lane and then just stops for whatever reason, or the hand goes up in a different directional, you know, we've had people just just accidentally get distracted by something over on the deck.

Luke's been on deck, so he's the best person to answer that question.