Ethical AI practice maturity model - Salesforce Tutorial

- [Instructor] Now let's jump into the ethical AI practice maturity model. So this is a super important piece for the exam, and it is a cyclical model that helps you go from deciding what you need for your AI tools to then having that tool grow up and grow with your company to then I guess going back to the beginning of is this working for us? And then how do we keep growing with it? It can be used with new products, and help increase your previous AI products as well. So the first one is going to be ad hoc. So starts when companies realize that they need to use AI tools. During this stage, you're going to identify the biases and consequences, as well as a lot of the benefits that you're going to need or the benefits that you're going to derive from using AI. So while you want to start using maybe a lead tool that has AI baked into it, like the lead scoring tool that Salesforce has, you will also realize that your information has a lot of bias. So you'll need to do a lot of data cleanup in order to effectively use this AI tool. Also, a big part of this stage is going to be getting buy-in from your stakeholders to implement these tools and these features. Not only do they have ethical and legal challenges, but they'll also be fairly expensive to implement and to get up and running. So it requires a lot of preparation and planning from stakeholders. So the next stage is going to be organized and repeatable. This is when you have decided that you have need and want for AI tools, and where you are developing your AI tools and models. So in this stage, you'll want to use the trusted AI principles while you are creating these models, and make sure that you are organized in your data. And a mnemonic to help you remember this is that you'll want to be organized in your data and then make sure that your outcomes are repeatable. If you have a lot of outliers or biases within your data, then it can become fairly difficult to, one, be organized. And two, the outcomes are typically not going to be repeatable and they'll vary widely. So just make sure that you have clean data and you're using trusted principles. Next is going to be managed and sustainable. So if we're thinking about this AI model in stages of ages, you can think of ad hoc as like the very infancy stage, and then maybe your organized and repeatable is going to be the toddler and kid stage. Now, managed and sustainable is going to be like your young adult, maybe teen stage. This is where you're already deep into your practice of using AI. You've already done a lot of the heavy lifting as far as keeping your data clean and keeping it organized, and you've come up with a bunch of different strategies. Now, this is where you are going to maintain those strategies. So using maybe third party tools to keep your data on track. Maybe you're using validation roles to help keep your data together and that there aren't too many outliers. Maybe you are using different meetings that you have to frequently go over your data, make sure that it's looking good, and that it is not biased in any way. This also helps you make sure that you have the right size solution. So frequently coming back and understanding the size of solution that your company needs, and making sure that your current solution meets that. And if it doesn't, how can you adjust your solution to be more of a right sized for your company? So finally, it's going to be optimized and innovative. We're going back to using the different ages for the different stages of this maturity model. This is going to be your middle aged to senior aged stage in this model. You're going to be taking out the fluff when you're using your AIs and models, and you're making the solution work as best as possible for you, taking out all the things that you don't need, and that it is working as best as possible for the scenarios and the solutions. So if you're using lead scoring, making sure that you have systems and processes in place to put your data into Salesforce so then the tools can output a score for you, and taking out any of the fluff that you don't need. So that's optimized and innovative, is going to be more of testing new products as they show useful for your company. Maybe you're finding new ways to use the products that you already have and to leverage them for greater growth.