I mean, we've heard a lot about, I feel like today and for the past couple of years about how we're using AI to generate content and build our workflows. I really, if I take a step back, I just kind of want to take you through a journey on something we went through in the pharma space and we'll overlap with what Paul was talking about a little bit as far as what it takes in pharma to get something from an idea to live. And really something that we landed on of just, there's no value no matter what we're doing in content until there's actually eyeballs on it, right? So if I just take a step back and kind of tell you our journey, when I personally first started working in pharma, it was part of a big project, super aggressive timelines, five, 600 websites. We wanted to migrate to Magnolia. And the goal was like, let's get that done in a year. And so we were like, yeah, let's do it. So we got aggressive and building it really wasn't the problem we ran into. We found a way to get everything built in Magnolia we needed. We found a way to get everything set up the way we wanted to. And then we were showing our timelines to people and they were kind of laughing at us. They're like, you're not getting that done in a year. And we were like, why? And they're like, well, it has to go through medical, legal, and regulatory review. And I was like, well, they just have to look at it, right? Just a review process. It can't be that hard. What can we do to speed it up? So that was naturally our first target. We were like, how do we make the review process faster? It took a little bit of time, but we went and dug in and said, just tell me about this. What are you guys doing? What is this whole process? Why is it so slow? The insight here wasn't that reviews took long. It was preparing for review. We learned in this world, it wasn't my idea, but we helped make it better. They're not just looking at the website. They need to take a PDF of every way that website can look. Every carousel on the page. Every accordion open and closed. Every form error state across different devices. So they're making these manual PDFs, putting them all together, and then, believe it or not, they have to go back in and add comments. What does every link do? What does every button do? On and on and on. So we broke it down and said, okay, there's a few problems here we need to solve. We're going to focus on making the prep work fast. The review stuff already is flying just fine. So we broke it down to a few pieces. We need to make the PDFs. We need to get the annotations or the comments and all the functionality in there. And then once we have that platform built and it's already talking to Viva Promomats where all the review work happens, we can start taking advantage of more features, like making sure that content is unpublished when it's expired or that we're syncing the metadata as it needs to go through. So, I don't know, are we playing here? Click again. All right, great. So what this turned into is it used to be multiple teams. There was the team who built the CMS. There was the team who put the content into the CMS. The agency who set that all up. Then they would get to another team who created PDFs, who sent it to another team, who added all these annotations in. When we're done here, and as you can see, we made something that integrated right into Magnolia where the content authors are working so agencies can make their content. And when they're ready, send it straight into Promomats with all of the functional annotations in place. So this is something that in real time was taking easily weeks, just from this team to that team, and then the back and forth. We really condensed that all into, hey, let's just push a button, right? So we thought this was awesome. Everyone who used it was like, this is great. We're saving so much time. And when we took a step back from it, and something I think we can apply anywhere is when we started this whole process, we were like, hey, how do we make the reviews faster? We really zeroed in on there. This is a lightning talk, so I'm not going to go through all of what we did there. But we could have easily gone down the wrong road, solved the wrong problem. And I think especially where we're at now with the AI tools available to us, it's very easy to be like, idea, let me go program something out. Let my AI agent help me do something. And it's very easy to go down the wrong workflow. For us, what we really needed to do was just kind of take a step back and understand who's doing what, who owns what step of the process. And what's funny to me, maybe, is as important as this is in the entire process of getting content live, there's no one single owner. There's no one person to say, this is your job. It goes across teams, across processes, shared ownership, which meant there just wasn't one place to go and say, hey, I need you to fix this. I need you to do better. It was really about taking all of those processes, getting everyone's key parts that they need to do, and then putting it into one pipeline. And so now we have something that's automated, button push, but also everyone's still included in the process, seeing what they need, abling everyone to take, again, what was days and days in roadmaps, really condensing that down to go faster. So my takeaway is don't always just jump at the first problem you see. Back up and understand who owns this problem, who are they working with, what's the overall pipeline from content creation to content publishing. And in this case, include legal up front so you can figure out what they need to get it right. Wow. It was almost like a voice from above, like the lightning just hit right after your point. So good timing there. I was wondering, I mean, it seems like what you were talking about, it's a bit about like hidden things in a project. In your case, it was this like preparatory step for the topic. I mean, do you have any hints or tips on how to identify those hidden things up front? Yeah. I mean, I guess two things. One is, you know, sometimes things feel like they go into a black box. I don't really think there are black boxes. I think it's going somewhere. Someone owns that. And I think also in any process we're doing, right, there's usually like some, an owner of it. Who owns this thing? If you can't find that, you probably have a broken process. Great. Thanks so much, Jeremy.