In this section, we're going to summarize the five phases in the Path to ML. In the last section, we talked about each of these phases in more detail. But now, what I want to do is zoom out, kind of rotate the issue on its side and give you a different view of all that same content. So, we talked about these five phases and it was Contributor, Delegation, Digitization, Big Data, and Machine Learning. One of the axes of change between each of these is who or what is executing the process. So, as we look at this diagram, which was made for you visual learners out there, we see that in the individual contributor phase, we have one person, when we get to delegation, we have many people. But the other three phases, it's always the same entity executing the process. It's always a computer or some server rack. But there's another axis of differentiation, and that's how do we choose the operational parameters? So, from the individual contributor, all the way out to digitization, we're doing some kind of market research or some kind of less empirical, more hypothetical exercise about, well, this is what I think users want, this is what I think we should try to accomplish. And then we get to Big Data, we actually, for the first time, have the opportunity to produce an incredible amount of very deep rich market insight or user insight or operational insight. And this gives us a new method to choose these operational parameters. And then in machine learning, you could imagine we're just going to start to regress or to make some, we're going to fit the data with a line or obviously maybe more complicated curve, where we start to actually perform these fitting parameters and extract to cases we may not have seen yet. And then this final axis of differentiation is given that we have these better operational parameters, how are they fed back into the execution of the core process? So, as long as we have humans accomplishing the task in the first row, then the way we feed back the execution, is we have some kind of like HR training or reference manual or something that the humans refer back to as needed. Of course, if you have a truly individual contributor, the worst case scenario is nothing's written down. So, we always hope that there is some kind of employee handbook. But then, as we move from digitization and Big Data, now, when we have these new operational parameters, we know what we should be offering to our users. We're going to rely on our software engineer with the headphones to go ahead and put that back into our code so that it may be acted on and presented to our users. And then, when we get to machine learning, the machine learning is actually going to be deriving these new operational parameters and feeding them back in automatically. So, I want to zoom even further back into some final reminders about this lecture in general. It's tempting to jump from nothing to a fully machine-learned solution, but it's a risky move. Success at Google and likely in your organization too, typically follows a more structured approach that steadily increases that upfront investment as the business uncertainties decrease with more experience and research into your users and your product niche. And we've talked a lot about this business process this core feedback loop here. And the main goal of the Path to ML is of course, to automate all the blue boxes, and that's what we're trying to do and that's the Path ML. The good news is Google can help, Google has a variety of services that can help you in across this path, starting at an indvidual contributor, but of course, because we're a cloud offering, we get very strong once you get to the digitization and beyond. And I've given you a link here where you can find more training opportunities from Google Cloud. So, what are the next steps here? So, in this module you got an overview of the path to ML-ify your business. You're taking a Coursera course, this is your opportunity to think big. How many different ways can you think of changing your industry? Changing your business? Changing your group? Doesn't matter how small you start, but what are we going to do that's new? Finally, welcome to machine learning specialization, into a career in machine learning, you're going to love the ride. Thank you so much.