Right? When things are constantly changing, we can't afford to do what is the non adual way, which would be what's called waterfall, where we're starting with will come up with all the requirements up front, then basically build to those requirements, then basically test based on those requirements and then deploy it. That works wen your building aircraft carriers and airplanes, because you don't want to be adulely building an airplane, not a good idea, but when you're building soft ware that is responding to continuous changes, it is an approach. The problem is s that we can't just take agile and use it for data projects. Because even agile starts with the idea of function, which is, what does the system do? And in the case of a i and data projects, all of that is highly dependent on data. The data determines what the model does. The data determines how well it works. You can have two of the exact same functionality. I build the chapbot, you build the chapot. We could even use the exact same technology. Pick your tech logy choice, amazon, google, microsop, whatever. We may agree to the same technology. We may even use the same original source of data. The problem is, if you train your conversational model one way and id train my conversational model another way, they will have completely different success. And of course, what we're looking for is not function lity. We're looking for success. So a the key to the c p m i methodology, which is a best practices methodology for running aian data projects, is itself. Just like agile, doesn't come out of nowhere. It is itself based on a methodology that bin around for two, two plus s called crisp dm, the cross industry standard process for data mining, which was not really built for a i, but it was built for data centric project. So it is data centric. That's a good plus. The problem was, it's not really built for agile sprints. It doesn't, it's not that centric to the specific needs of ing a i and advanced data project. So we have to do something to it. And that thing that you have to do something to it, it's called c p m a i, which is the cognitive project management for ai methodology, which basically enhances crisp d m with a series of things. And what we're ging to do here in this pod casses will provide an introduction to what cp i is and talk about its phases and how we use those phases across agile sprints to get to a successful a projectanasron mentioned there are six phases, two s p m i. So phase one is business understanding. It's really important start here. Some people say, we'll doo start with, you know, your business needs are your data needs? And we say, always start with your business needs, because if you're not solving a problem. Don't do it. Don't do it. Unke and i. People skip this stip if you're not following a methodology. And we have seen this. This is part of why projects fail. Don't skip this step. Don't

skip it. It's so important.