Do you sort of have a priority where you place your bets in terms of your AI investments? Is there maybe a benchmark by which you judge where you should be spending? Presumably you can't be spending everywhere all at once.

Absolutely, and I think where to spend your AI investment is probably the hardest challenge that any business has, right? So for us, I guess we're looking at it really in two different places.

The first one is, can I serve a customer for a cheaper internal price? So that's going to be around operation efficiency, delivering payments for us more quickly.

Can I use smart routing to make sure the payment gets there more quickly for a customer? All those pieces that mean internally we're delivering our core business more quickly to customers.

And then the other side is, well, do I build AI-facing features for customers? Do our customers want to interface with us through just typing into the platform now? Is that the way we want to work? Or do they want to have extra information brought into their journeys, but it's pulling from the real outside? All those kind of use cases.

And also, do we just build new features for our customers, right? Like what features before where it was marginal gain for us to build, but actually we could build those much more quickly now.

And trying to balance those two things is quite interesting.

Are there other implications, perhaps, into the cybersecurity side there around autonomy?