AI assessment and readiness

- [Instructor] Now it's important for organizations to take a reality check. Where are they in the AI journey? I've divided this AI journey into four stages. So there's the descriptive AI phase, the diagnostic AI phase, predictive AI and finally, a prescriptive AI phase. And as you move to the right, the more mature your AI practice. So what does this mean in terms of the time that organizations spend on different tasks? As you can see, the ratio of the time varies across the continuum. The more mature your AI practice, the smaller the proportion of time you spend capturing data and the greater the proportion of your time that is spent analyzing the data and acting on it. Descriptive analytics is the foundation. Essential business data is captured, but it can only tell you what happened. Using an analogy of an online retailer, this is the retailer capturing all of your online transactions. So you've captured the data, and perhaps you know what are the best-selling products. So how do you know if you're still in the descriptive AI phase? Well there are a couple of telltale signs. Business stakeholders rely upon business intelligence that primarily focuses on reporting what has occurred in the past. The vast majority of your time is capturing the data, but no one is using it until a report needs to be made for the management team. So you're providing periodic reporting of historical performance to business leaders. You're at the descriptive AI stage if you're still trying to acquire systems and processes to help make data-driven decisions. Your organization often rely on the talent, instincts and experience of leaders to make decisions, rather than data. And you know you're at the descriptive AI stage when insights can sometimes take months and they're not shared across the organization. Data remains in silos and is generally not accessible and there's no centralized data management across the organization. And the only reporting might be some shared network drive with several Microsoft Excel and PowerPoint presentations that were made to the management team. You're probably at the descriptive AI phase if executive sponsorship and investment is lacking. Now if, based on your assessment, you are currently in the descriptive AI phase, then the focus needs to be getting to diagnostic AI. How do you do that? So you want to initiate a project with your platform or IT team where you're moving from where data is generally not easily accessible, to data becoming more available. You move to more advanced data processing methodologies, although this might still exist in silos and you're moving from data in silos to connectivity existing between various data entities. So how do you continue to move from the descriptive AI phase to diagnostic AI? You need to start thinking about the three Vs. So that's volume, variety and velocity. So you're starting to think about getting an understanding of the available data, so that's the volume. It's structure, that's the variety. And frequency, which is the velocity. And how data can drive business outcomes. So you're identifying business rules and ensuring that compliance policies are documented. You're visualizing data outside of Excel and PowerPoint using Dashboarding. This might still not be centralized, but you're working towards that. And how do you continue to move from descriptive AI to diagnostic AI? Well you're working towards having reusable code in some business units. You might not have had any data scientists or machine-learning engineers, but you're starting to invest in these resources and crucially, for them to be working very closely with the business stakeholders. Now this might be difficult for some of the seasoned leaders to get used to, but as an organization, you are starting to make a transition from depending on the talent and experience of leaders to having a data-driven culture. You aren't replacing the experienced leaders. Rather, you're increasingly using data to inform decisions. Now if in your assessment, your organization is in the diagnostic AI phase, so basically you did something in the business and something else happened, you are able to analyze your data. So these are some signs that you are in the diagnostic AI phase. Business stakeholders recognize the value of more advanced analytics and are developing a strategy and a plan to invest in technology. So you've invested in data science and AI resources and personnel and the change in management necessary to take advantage of advanced analytics capabilities. The data scientist and the business stakeholders are working closely together to identify business use cases and the organization is aware of new data applicable to business needs and is able to harness them with some effort. So if you're firmly in the diagnostic AI stage, then you're probably in a position to seriously consider creating a business case for AI in the organization. You need to start thinking about how you can embed AI insights into applications and systems and business processes. This stage allows you to take action. So if you're an online retailer, out of the hundreds of thousands of items that you can sell, using a recommendation engine, you put on your shopper's screen the five items that you think they're likely to be interested in. These are the items you think they are most likely to be interested in buying based on what is currently in their shopping cart, their buying history with you and what similar buyers to their profile would opt for. And at the predictive AI stage, you as an organization are also thinking about the ethical implications. You know what you can do with AI. The question sometimes is should you? And finally, there's the prescriptive AI. As an organization, you are running agile experiments and iterating between analyzing and acting on the data. This is where the true business transformation and disruption takes place. At the prescriptive AI stage, there could be several things that are taking place. So think back to the online retailer. They now know some of their customers so well and let's say you're one of them and they are so confident in their predictions that they have set up an informal subscription model with you. So before you've even ordered anything, they have shipped it to you. So you ordered dishwashing powder and dog biscuits every couple of weeks, it's at your door even before you've ordered it. Now the online retailer goes one step further. In addition to the stuff that you normally buy, they include a couple of items that they're certain you would be interested in. If you don't buy them, just leave them in the box and they'll pick it up later in the week. See what they've done? You do most of your online shopping with them now. You don't go anywhere else. You see, all along it's been about you shopping and then they ship those items across to you. They've turned shopping on its head. Now they ship first and you shop later. This is where you get true disruption. You think that a retailer truly understands you. Now some of you might get freaked out and some of you might think this is great and one less thing to worry about. Prescriptive AI, this is true business transformation and disruption. This is where most organizations want to be.