 One of the transformations we've made along the way in our DevOps journey is changing how we do deployment. We have what we call a pre-production environment where every single change that is destined for the service has to go first. Using release management, we're able to have the individual feature teams and developers are able to trigger their own deployments and deploy hot fixes, be able to put their changes directly onto the service. Automation is a critical part of running a service. It's important across the totality of the development lifecycle from automating code check-ins and testing to make sure the code that you just checked in is actually going to work. And that enables us to have a much faster pace. You go from one change every week to two changes every week to hundreds of changes every week. I want to know about problems before my customers do. So, being able to understand what's going on on my service, being able to diagnose those issues and get it done quickly is incredibly important. Telemetry is absolutely key. We're starting to begin developing skills for understanding data and becoming data scientists, really, to be able to pick through and see the trends necessary to help us make decisions moving forward. Having the ability now to actually see what customers are doing, to instantly know if that capability that you just created and put into production is being used, that's fundamentally different. And it's really, I think, lit up the organization. Power BI is a service that's designed specifically to allow you to collect a lot of data, to sift through it, to visualize it, to create dashboards, to ask questions using natural language. Our developers are excited by the fact that they now have a direct connection with the people who are benefiting from their work. Now, having said that, because our engineers also operate the service, when something goes wrong at two o'clock in the morning, they're getting called. When we have an incident, the operations team will coordinate that incident, bring in our partner teams, do communications. But Dev's really the best suited to understand and troubleshoot their code. They're very well suited to understand those issues, and so you want to enable that access. If you don't open it up, you're going to miss out on the critical piece of that accountability and responsibility that the engineering team had. We became very metrics-driven, very scorecard-driven, but the whole breaking down the ops and the Dev responsibilities helped reduce the live site incidents more than any scorecard ever could and help just drive the accountability problem. We're seeing an improvement in terms of how people are feeling about their jobs and the impact that they're making on their customers and on the world.