Yes. We talked a little bit about this before, I think, in last quarter that I think there's a number of companies that created AI departments, et cetera. And I think they probably spend a lot of money with not a lot of gain. We've been very disciplined around how we approach this, and we very much look at use case solutions that actually drive an ROI. We've got a number of activities inside MTS that are funded by government, as you would expect, as we look at that from an R&D perspective, and that hangs off the back of our digital engineering labs that we press released recently and talked a little bit about in the prepared remarks, which are gaining good traction because of the speed to market of R&D projects, et cetera, as you would expect. More in the STS world, again, looking very strongly at use cases around accelerating engineering, making sure there's checks and balances within that engineering that avoid human errors, speed up progress. But at the same time, looking at how we operate facilities across the world or our customers operate facilities and using particularly digital twins and applying AI and machine learning to really sort of draw data and get trending over time to know what good looks like and make sure that operators can intercede at the appropriate time or the maintenance crews can intercede at the appropriate time. So it's a multifaceted approach, but ultimately, it's driven by use case ROI, and we put quite a bit of front-end effort into that, Andy, rather than just saying AI is good and just running at it. We've been quite disciplined. And that's on the front of office. I think in the back of office, increasing use of bots to drive efficiency and decrease human error, keep our SG&A in check or reduce it in fact, over time. We're rolling out Microsoft Dynamics across the STS portfolio, which is really the forefront of a digitalized ERP because our project controls, which gives us all the project data hangs off that, and we can look at things real time and start to make real-time decisions on commercial execution. And we've also got digital procurement hanging off the back of that and with a similar upside. And so I think that whole digital project execution philosophy underpinned by really a very modern and digitally enabled ERP is going to stand us in really good stead as we come out of the spin. So I think it's multifaceted this front of office driven by use case, the back of office, again, driven by use case, but obviously with different drivers.