Discover AI Foundations for Business Leaders

0:00My name is Simona. That's me here.

0:02And welcome to this content for the Microsoft AB731 AI Transformational Leader exam.

0:10Transformational leader, huh? It sounds so grand, doesn't it?

0:14Now, our first task in this video is to understand what generative AI is

0:19and how it fits into the broader AI landscape.

0:23And with that in mind, let's start with just the AI bit.

0:27And then we'll see where the generative AI part fits in.

0:31So you need to think of AI, artificial intelligence, as a collection of capabilities.

0:37So AI isn't just one thing, but what they do all have in common

0:41is that these capabilities allow computers to perform tasks

0:45which previously an actual person would have to have done

0:49with their human judgment and their human intelligence

0:51and their human experience and human learnings and whatnot.

0:54Hence, the idea of artificial intelligence as opposed to real-life human intelligence.

1:00And the kind of tasks that we're talking about here will include things like

1:05spotting patterns and recognizing images and understanding language

1:10and even understanding the sentiment of language

1:12and things like creating summaries and, I don't know,

1:15providing recommendations and more. We could go on.

1:19But of course, because it's a computer doing that rather than a mere mortal,

1:23one of the most obvious advantages we see is that we get faster results

1:28because, well, computers can process more information

1:31and they can do it faster than, you know,

1:34my brain or your brain could ever do without exploding.

1:37And for you, as a business leader, that's the interesting bit.

1:42We'll start to see as we progress through this course

1:44that these capabilities do deliver real business outcomes.

1:48And that, for sure, is going to be more interesting to you, I think,

1:52than the, you know, the technology and the algorithms and whatnot that make it all happen.

1:58So, therefore, keeping in mind that AI is this bigger picture of a collection of capabilities,

2:04well, generative AI is a subset of that bigger picture of AI.

2:09And you can see a good definition of it here.

2:10So, in a nutshell, we can describe generative AI as a subdomain

2:16or a branch of artificial intelligence,

2:19which is designed to produce or to generate.

2:22Did you see what I did there?

2:23It's designed to produce brand new content such as text or images

2:28or even video and audio from just a simple description.

2:32So, no programming required.

2:35Well, by that, I mean that there's no programming required

2:38by the mere mortal human being using the generative AI-powered tool.

2:44But, of course, there is an awful lot going on behind the scenes.

2:47And, in fact, under the hood, we actually are programming when we use these tools.

2:51It's just that we're programming in plain language,

2:53you know, natural language as if you were talking to another human.

2:57And we will pick up on that idea again another time when we learn how and why

3:02we need to put some thought into that simple description that we type

3:06rather than just type a thought and hope for the best.

3:08Oh, and I should say that this simple description would be known as a prompt.

3:13But as I said, more on that another time.

3:15So, I think this is probably a good definition for us to hold on to

3:19as to what generative AI is.

3:21And it is quite possible that you're already using a generative AI-powered tool

3:26like Copilot to produce summaries of, I don't know,

3:30Teams meetings or to give you insights on your Excel spreadsheet.

3:34And so, as a practical example here, in your Excel spreadsheet,

3:38you might launch, for example, Copilot Chat and then type in something like this.

3:43So, this would be the prompt.

3:45So, something like, analyze this workbook and give me interesting data insights.

3:49So, that's the prompt.

3:50That's our natural language request.

3:53And Copilot is going to review your big, horrible data set

3:56or whatever it is you've got in Excel.

3:58And then, as you can see here, it'll give you an overview

4:02of what your data actually looks like.

4:04And then it starts to identify some trends

4:07or whatever insights and correlations, et cetera, it can find in that data.

4:12Now, we will learn more about using Copilot and its capabilities later on.

4:17But I just wanted to show you a practical example of what I mean by generative AI.

4:21So, let me just go back to that previous slide.

4:24Yep, there it is.

4:26And because generative AI tends to be that branch of artificial intelligence

4:30which suddenly puts the power of AI into the hands of a regular end user,

4:35you know, looking at their Excel spreadsheet or whatever,

4:37and they neither know nor care what technical wizardry is going on behind the scenes,

4:42then it is tempting to think that generative AI is the sum total of what AI can do.

4:48But I really do need you to understand that, well, yes,

4:51generative AI generates new content from a natural language request.

4:55Marvelous.

4:56But also, I need you to appreciate that it's just a small part of the wider field of AI.

5:03It's just a branch or a subdomain of artificial intelligence.

5:07And in fact, let me show you this slide here,

5:09which I have stolen from my esteemed colleague, Jonathan Barrios,

5:13who has created loads of AI training here at CBT Nuggets.

5:16And I always say this, but he really is quite possibly

5:19one of the most interesting people you could ever hope to meet.

5:22But on his slide here, you can see that he says,

5:24if we think of AI as a box of goodies just here,

5:28well, yes, generative AI is in the box, but it is not the whole box.

5:34So AI, artificial intelligence, is the broad field,

5:38including things like machine learning and deep learning.

5:41And then, yep, there we go, generative AI.

5:45And this is where those generative AI powered tools like chatbots sit,

5:50which for many of us is our first experience of seeing generative AI in action with,

5:55I think it must have been chat GPT being one of the first on the scene.

5:59And then, of course, since then, we've seen others pop up with Gemini being

6:02Google's generative AI tool and, of course, Copilot being Microsoft's.

6:07And there are others, too.

6:09And leading us on to the next video in this skill,

6:12there is a term that you might hear in the context of AI and generative AI,

6:17and that's LLM, standing for large language model.

6:22And LLMs are key players in generative AI because they provide the foundation

6:27for your generative AI powered tools to be able to understand your natural language requests,

6:33you know, that prompt that you type in in natural language rather than programming.

6:38Now, under this video in the skill, I have given you a link to one of Jonathan's AI courses,

6:43the first skill of which does a really fantastic job of exploring AI and its different types

6:49and therefore how generative AI fits into the bigger picture.

6:52And I urge you to take a look if you would like to, but this is optional.

6:57You do not need that extra background for the AB731 exam.

7:01And if you really do want to get more technical,

7:03then Jonathan's got loads of other AI courses as well.

7:06And I urge you to take a look if that is the direction you want to go in.

7:09But anyway, having got our heads around what we mean by generative AI in the next video,

7:15I'd like us to jump in and learn about some of the different generative AI models.

0:00This is one of the objectives from the skills measured document for the AB731 exam.

0:07Describe the differences between AI models.

0:10And I think a good place to start is, well, what do we mean by an AI model?

0:14And we can kind of think of it as sort of like the sum total of all of the background maths

0:19or math as my American co-workers would say.

0:22So all of the background maths and probabilities and statistics and whatnot,

0:26which have been formed from learned patterns that have come out of vast amounts of data.

0:31We're talking internet scale data here.

0:33And these patterns and algorithms and whatnot are used to generate new content.

0:39And so at the simplest level,

0:41the model is using a kind of like a sophisticated network, if you like, of predictions

0:47to guide it as to what is the most likely thing it should produce based on the context.

0:52And the context would be, for example, the prompt that you provide it with.

0:57But another term I need you to be aware of in the context of generative AI models is training.

1:03And training is how these learned patterns in the model emerge.

1:08So, yes, a model is trained on huge data sets, massive internet scale data.

1:14And this idea of training means that its parameters or its own internal

1:19thinking, if you like, is adjusted in order to produce more accurate results.

1:24And the actual process of training is a fascinating and complex subject,

1:28the details of which we don't need to concern ourselves with right now.

1:31But what happens is it involves the model making a prediction,

1:36comparing what it comes up with, with the correct answer.

1:38And then it changes itself in order to be more accurate next time.

1:42So it's making changes to its parameters.

1:45And it's making mathematical corrections to all of this background wizardry in order to improve.

1:51Now, I realize this idea, if it's new to you, will take a little bit of processing.

1:55So go easy on yourself.

1:57All of this will start to feel more comfortable as we learn more.

2:01But with this idea of training in mind, be aware that traditional

2:06artificial intelligence solutions focused on building custom models for predictions.

2:13And I'm going to remind you of this when we come back to talk about

2:15machine learning later on in this skill.

2:18And so, yes, artificial intelligence solutions of a few years ago

2:21would all be based on custom models.

2:24But generative AI is the next layer of evolution and usability, if you like, of AI solutions,

2:32because it uses these models.

2:34And for that reason, we often describe generative AI as using pre-trained models.

2:41And that means that we can implement a generative AI solution

2:45without having to build our own custom model as per those traditional AI solutions.

2:52And that's great.

2:53But in addition, a pre-trained model could be further refined and customized

2:59for improved accuracy and relevance for your industry or your organization

3:04by a process of fine tuning.

3:06So that's a term I need you to be aware of.

3:08So a fine tuned model has undergone further training in order to, as I said,

3:13refine and customize the model so it performs better for a particular scenario.

3:18And sure, that will give you better results, you know, more accurate results.

3:22But the more fine tuning that you do, unsurprisingly,

3:25the more cost and complexity you add to the deployment of your AI solution.

3:31Now, you might also hear the term instruction tuned,

3:34which means the model has gone through another process of enhancement

3:38to learn how best to follow human commands.

3:41So this means they're trained to better interpret and execute human instructions

3:45and to respond in a conversational and helpful way that humans would expect.

3:50Now, some of the Microsoft learning content that you might have come across

3:54in your journey for preparing for this exam,

3:56well, some of that content refers to the term pre-tuned,

4:00which I'm not going to lie, isn't, I don't think,

4:03a general term that we see that often in the context of AI models.

4:07We normally just see the idea of various types of training,

4:10such as pre-training and fine tuning and instruction tuning.

4:14But if you do happen to see the term pre-tuned in the AB731 exam,

4:19then assume that the idea is that a pre-tuned AI model is ready to use

4:25and therefore offers a less expensive solution compared to a fine-tuned model

4:30that you've taken the trouble to further refine for your organisation.

4:34OK, so my goal on this slide was to explain the idea of pre-trained and fine-tuned.

4:42But in addition to all of this, there are some specific model types

4:46that I really do need you to be aware of.

4:48And I have laid them out in this rather boring table here

4:50because I do need you to be clear on these,

4:53since you are likely to need to be able to distinguish between them for the AB731 exam.

4:59And the first one on the list, well, this is the one that,

5:01if any of these that you've heard of, it's going to be this first one,

5:04an LLM, a large language model.

5:06And I did vaguely mention this, didn't I, in the previous video?

5:10Because as I said, that tends to be the one,

5:12if anybody's heard of an AI model, that's the one they've heard of.

5:15So this is the model that generates and understands natural language text.

5:20It's what we first think about when we talk about generative AI.

5:23This is what understands our plain language instructions.

5:27And it's this type of model that's going to need to undergo instruction training

5:30or instruction tuning, I should say,

5:32to better be able to interact in a natural conversational way.

5:36And an example of a large language model would be GPT-5,

5:41which at the time of recording is the model behind the chat GPT chatbot

5:46and also behind Copilot.

5:48So this is a screenshot here of Microsoft 365 Copilot chat.

5:53And up here, when you're starting a new conversation,

5:56if you were to click the dropdown,

5:58that's where you can specify the model that you want to use.

6:01And yeah, it's using GPT-5 in the background there as the underlying model.

6:06But anyway, let me just go back to the table.

6:08There we go.

6:09And next on our list here, small language models.

6:12You won't hear about these as much as you will hear about their bigger brother,

6:15the large language models.

6:16But a small language model is similar to an LLM in that it is a language model.

6:22So yes, it generates and understands natural language,

6:25but it's characterized by having a much smaller size and scope

6:29with fewer parameters than a large language model.

6:32And that means that they require less memory and less computational power.

6:36And that might be useful for mobile apps or even for offline use,

6:40or maybe for very task-specific or very closely defined

6:44or targeted domain-specific applications.

6:47You know, something with a small, carefully designed scope.

6:51So yeah, small is the operative word there in that type of model.

6:56Next on my list, code models.

6:58Well, these are models which have been optimized for use

7:01in generative AI-powered tools designed to help developers.

7:05So it's going to be used in solutions which will assist with tasks

7:09such as, I don't know, completing code or debugging or automating tasks

7:14or running tests or proposing changes to code, you know, that kind of thing.

7:18You know, it's specialized to the world of coding.

7:21So therefore, I think that's probably an easy one to remember

7:24because it does do what it says on the tin.

7:26Diffusion models here.

7:28Now, these types of models are primarily for image creation and image processing tasks.

7:34But the applications of these types of models has been extended to other domains

7:39such as the creation of audio and video from, you know, user text-based prompts.

7:45And it is quite fascinating and mind-blowing how these models work.

7:48But hey, I will leave you to research that if you're interested.

7:53Next, you can see I've got a multimodal model,

7:56which is a bit of a tongue twister, I'm not going to lie.

7:59But these models can handle multiple input types

8:02such as it says here, text, image and audio.

8:05So they are designed to process and integrate information from multiple data types or modalities.

8:12So modalities being the posh term for different data types.

8:16And GPT-5 is described as a multimodal LLM.

8:22It's a multimodal large language model.

8:24So a model can be both an LLM and multimodal.

8:30And then lastly, again, hopefully an easy one to understand.

8:33A domain-specific model, as you might guess, is a model that has been fine-tuned.

8:39Aha, we know about fine-tuning.

8:40It has been fine-tuned for specific industries or tasks.

8:44So hopefully that's an easy one to remember.

8:47Now, as a business leader, it is unlikely that you need to understand

8:50the technical detail behind these models.

8:53But you do, I think, need to be able to hold your own in technical discussions

8:58where different models are being discussed and being assessed

9:01for use in the AI solution that you're deploying.

9:03And I really do expect you to be asking these types of questions

9:07when your AI solution is being developed.

9:09So, you know, how much cost and complexity will fine-tuning of the model add?

9:14That is a very relevant question to be asking.

9:18And what about something like this?

9:20Is there a model available which has been pre-tuned for our domain?

9:23So you'll see I've used that term pre-tuned there to mean

9:27something that's been further refined specific to our particular area of domain,

9:31whether it's healthcare or legal or something.

9:33Or maybe something like, is the model multimodal?

9:37So perhaps your solution is going to be used by marketing people

9:40who might be requiring something that can handle different data inputs,

9:44such as text and images and whatnot.

9:46But you get the idea.

9:48And later on, you are going to get the opportunity

9:51to match up the right AI model to a business objective.

9:55And that is absolutely going to be the kind of discussion you'll need to be a part of.

0:00Taking a look at the list of skills measured for the AB731 exam,

0:05there are two which are all about machine learning. Here they are. So number one, identify

0:11scenarios where machine learning adds value. And then number two, describe the life cycle of a

0:16machine learning solution. And surprise, surprise, that's what we're going to tackle in this video.

0:22I think we should probably start by making sure that we can define the term machine learning.

0:27And let's start though with what you already know. So you will remember that generative AI

0:34is that subdomain of artificial intelligence that generates stuff.

0:38So we saw this slide, didn't we, in the first video in this skill. And you might also remember

0:43this slide, which I stole from Jonathan Barrios, one of the other CBT Nuggets trainers.

0:48And you can see that, yes, aha, here we have machine learning as one of the other subdomains

0:55of AI. It's one of the other goodies inside the box. So let's get ourselves a definition

1:01of machine learning. And taking Microsoft's definition here seems as good a one as any.

1:06So what is machine learning? Well, you can see here it's a branch of AI that enables computers

1:12to learn from data and improve their performance over time without being explicitly programmed.

1:20And if this is starting to sound slightly familiar from the last video when we learned

1:25about generative AI models, then yes, generative AI is powered by models which have been trained

1:33using machine learning or should I say pre-trained. Remember, we talked about this

1:37idea of a model being pre-trained that we can plug into our generative AI solution.

1:43And you might remember that these models adjust themselves during training in order to become

1:49more accurate. They make changes themselves to their own parameters and the mathematical

1:55wizardry and whatnot in order to provide more accurate responses. And that's what we're saying

2:00here with this definition. They improve their performance over time without being explicitly

2:05programmed. That's the beauty of the idea of training a model. We're not hard coding.

2:12If this happens, then do that. We are training it to interpret the data correctly.

2:18So if that's our definition of machine learning that we will go with, now let's tackle our first

2:24objective for this video. And that is to identify scenarios where machine learning adds value.

2:30And I have to say, I am kind of stepping on the toes a little bit of what I want to say

2:34when we talk more later on about the business value of AI. But what we need to appreciate

2:39at this moment is that the point of using machine learning as part of your AI solution

2:44lies in scenarios where you are not explicitly focused on generating new content like we do

2:51with the generative AI subdomain of artificial intelligence. But with machine learning,

2:56you're doing these kinds of things here. And so yes, generative AI is great for generating new

3:02content. But machine learning is better suited for tasks like this, such as numerical prediction

3:08or classification and clustering. And what you see here would be some common scenarios.

3:15So predictive analysis would help us, for example, to forecast sales or understand customer demand

3:21or predict risk or predict customer churn, that kind of thing. While personalization

3:28could be more about classification. So perhaps grouping customers together who share similar

3:32traits or behaviors, grouping them into different market segments to help us understand our customer

3:37base better. And then automation might be things like detecting anomalies in our data. And that

3:44could be in the context of financial transactions in the case of detecting fraud, for example,

3:49or maybe even looking for faults and probable causes in a manufacturing process as another

3:54example. So do you see that these kind of solutions aren't about generating new and

4:00original content from a simple text-based description, which is what we see with generative

4:05AI. These are about predictions. And I really do think that predictions is the key word here.

4:11And in fact, you might also hear the term inferencing, which simply describes the scenario

4:16where a trained machine learning model is used to make, you guessed it, predictions. And we will be

4:23embedding this predictive functionality, if you like, into applications or services or whatever

4:28solution we're building. And I do hope that you can read sideways there. So that does say that

4:32we embed this type of stuff into an app or a service. Anyway, I'm going to be testing you on

4:38some of this shortly when we start to bring all of this together by looking at some of the business

4:42value of AI solutions. But for now, just know that machine learning is very much focused on

4:47predictions. But the other objective for this video is, as you might remember, to describe the

4:54life cycle of a machine learning solution. So that's what we'll do now. And you can see here

5:00that I've got the bare bones of a flowchart. And we are going to fill in the gaps together.

5:06So the first shape on the left hand side here, that is the data. Now, this could be your data,

5:13it could be just general internet data. But the point is, there's some kind of data source that

5:17you're going to be building your machine learning model with. And then you train the model. Well,

5:22I say you train the model, your technical team will get involved in training the model. But the

5:27point is that the model gets trained on that data, we then deploy the model, we will then

5:33be monitoring the model in terms of determining how it's getting on with its training and its

5:38learning. And then we can analyse its performance and retrain it. That will lead us to updating the

5:43data and the cycle starts again, you know, the process continues. Now, this cycle describes

5:50what we would call MLOps. That's machine learning operations. It's kind of like DevOps. But for

5:56machine learning. So it's the process for developing machine learning models for production.

6:00It's kind of the structure of processes for reliably building and monitoring and maintaining

6:06machine learning models. So it's a, you know, it's a set of practices for repeatable processes

6:10for the development of machine learning models. And while it will be your technical teams who

6:15are concerned with the detail of MLOps, that's why I kind of glossed over it quite briefly.

6:21It is really important for business leaders to have this overview and your input as a business

6:26leader will make sure that all of this MLOps work stays aligned with business goals. So for example,

6:33in the red here, you will be involved in exploring and preparing the data, especially if this is

6:40going to be data based from your organisation, then it's up to you as a business leader to be

6:45clarifying that this is appropriate and current and the right data to be using. And you will also

6:51be part of the process when it comes to evaluating the outcomes. Because when we say analyse the

6:55performance, well, what are we measuring it by? How are we determining whether performance is good

7:00or bad? What are we analysing for? And the answer is, we are checking for alignment with our business

7:06goals, aren't we? So if we take a big step back and look at the entire machine learning project

7:11as a whole, as outlined in purple here, then there's one more thing we need to add to complete

7:16the picture. And that's defining the goals of the solution that we're building this model for.

7:22What are we trying to achieve and why? You know, what are the tasks we want this solution to

7:26accomplish? So this is where it all starts. And that requires business leader input.

7:33So what I'd like you to take away from this slide is that the blue bits in the middle

7:37is a simple representation of an MLOps machine learning model lifecycle, the detail of which

7:44will be of concern to your technical teams. The red bits is where we require your input as a

7:49business leader. And then the overall diagram outlined in purple might be one way of looking

7:54at the lifecycle of a machine learning project.

0:00As a business leader, you might be thinking, sure, some of this AI stuff looks interesting

0:05and maybe you've got people working with you or for you who are already excited about it.

0:10But I'm hoping that you're thinking, where is the business value?

0:13And yes, you really should be asking this question.

0:16We can't just deploy stuff because we can.

0:18We have to understand where the potential returns on investment might be.

0:22So where is the business value of the AI solution?

0:26And what we're finding is that, yes, AI does have the potential to offer

0:31every kind of organisation new ways to streamline their internal processes

0:35and their wider operations and new ways to innovate,

0:39whether that's new ways to connect with customers or

0:41new insights into customer behaviours or, I don't know, new ways to design products.

0:45All of this means that we've got new ways to stay competitive.

0:49And therefore, if we aren't using AI and our competitors are,

0:54then, you know, maybe we're in danger of being left behind.

0:57And while adopting AI might seem complex and costly,

1:00the potential return on investment really is significant.

1:04And that's where you come in.

1:05So if we think about AI as being more than just a technological upgrade,

1:10the responsibility of which will fall to your technical teams,

1:14think of AI instead as a strategic asset.

1:18And then it becomes easier to see that the success of it depends not only on the tools,

1:23but also on strong leadership and a clear roadmap and a clear shared understanding

1:28on what we're trying to achieve and what problems we're solving.

1:31And as I said, that's where you come in.

1:33So a successful AI project relies on you

1:36clearly understanding what the business value is going to be,

1:40making sure that that vision is shared by the technical teams who will be implementing it,

1:44and then leading the organisation through the,

1:47you know, through the transition to get to the point where business value is realised.

1:51And truly, the possibilities are endless because AI powers everything from chatbots

1:56to automation tools to advanced analytics.

1:59So your role is going to be to identify the potential that AI offers

2:04for solving real world problems in your organisation.

2:08So let's start with some scenarios where we can apply AI to business

2:12to start giving you some ideas.

2:14And I have lifted these from the Microsoft documentation,

2:17just to be sure that we stay on track with the lines of thought

2:19that the exam will be testing us on.

2:21And these are just four examples to help us bring the possibilities to life.

2:25So firstly, you can see here generating marketing content.

2:29And what we're talking about here is using AI to create personalised or localised

2:34campaign copy or promotional emails or product descriptions

2:38or social media content, whatever it is, you know,

2:40give AI the task of creating this kind of content,

2:43which is customised for different market segments,

2:46or different audiences or different geographies or whatever.

2:49That legwork of creating this personalised copy can take minutes using AI

2:54rather than literally weeks using humans.

2:57So the marketing folks will still be reviewing and refining it, of course,

3:01but the time savings are tremendous.

3:04And the people who don't have to write all of that content from scratch

3:07now have the opportunity to apply their talents to other value add tasks instead.

3:12So in terms of business value, then yes, sure,

3:16using AI in this marketing context is going to accelerate launches,

3:19it's going to accelerate, you know, campaigns and product launches and whatnot.

3:23It will, of course, reduce production costs in terms of the hours spent creating it.

3:27And it is going to increase engagement by easily

3:30being able to have customised messaging for our different customer segments.

3:35Next on the list, though, in fact, let me change my pen colour to purple here

3:38because I can already see that this slide is going to get rather busy.

3:41But in a customer support scenario,

3:43we can use AI to automate responses to common enquiries,

3:48probably in the form of a chatbot,

3:50or by integrating AI into the processes of our customer service teams.

3:54I mean, maybe you could use AI to summarise previous interactions with that customer

3:59and then use it to draft emails in the tone of voice that you specify.

4:03And maybe the solution is even configured

4:05to suggest what steps the customer service representative should take next.

4:10And in fact, we would call that prescriptive AI

4:13when it's configured to help with recommendations and optimisation,

4:16you know, when it helps with decisions and recommends specific actions.

4:20But again, where's the value?

4:22Well, the value is going to be in reducing resolution times for customer queries

4:27and reducing support costs and improving customer satisfaction, of course.

4:31And ultimately, being able to provide better

4:34and more complete customer service without having to increase your headcount.

4:39Next on the list here, document drafts and summaries.

4:42Now that might sound a bit lame.

4:44But look, if you've got weighty documentation to deal with,

4:47such as regulatory compliance documentation or legal documentation,

4:52or very heavy, detailed financial reports,

4:55then we can use AI to extract the essence of that documentation

5:00in the form of text-based summaries or even audio summaries,

5:03or to generate FAQs for the team,

5:06or for your customers, or whatever the audience is,

5:09or to pull out what's changed from the last version of that document.

5:12It's saving many, many hours of human review

5:15when the size of the documentation is significant and the content is complex.

5:20And the business value?

5:21Well, aside from the time saved by automating these tasks,

5:25we are reducing the risk of non-compliance

5:27in the context of regulations and whatnot

5:30by making sure that the content is easily available

5:32and consumable to the people who need to know it.

5:35And we're potentially reducing legal review time.

5:37And there's certainly going to be a cost saving there, isn't there?

5:40And then last on my list here in the green scribble,

5:44AI is going to help us accelerate product design and innovation too

5:48by helping to simulate new product designs,

5:50or helping product teams to brainstorm,

5:53you know, helping to explore creative options,

5:55or optimizing supply chains.

5:57The possibilities are endless.

5:59And the business value?

6:00Well, we're going to be able to have faster innovation times

6:03and faster times to market.

6:06Note, though, that these are all examples of scenarios

6:10where generative AI would be used.

6:12And I think the clue was in the first one, wasn't it?

6:14That we generate marketing content.

6:16So these are all examples of generative AI.

6:19And if we turn our minds to examples of where machine learning

6:22can add business value,

6:24well, happily, this is what we mentioned in the last video.

6:28So thinking of those machine learning models

6:30embedded as a solution into an app or a service,

6:33we did cover these same examples, in fact, in the last video.

6:36So yes, the first two on my list here,

6:39predicting customer churn or assessing risk forecasts, for example,

6:43that's all to do with the predictive analysis

6:45that we know machine learning models are so good at.

6:48Anything to do with predictions is going to be an indicator

6:51that we're talking about some kind of machine learning model.

6:54And then personalizing product suggestions here.

6:56So that's going to be all about personalization and classification,

7:00you know, grouping customers who share traits,

7:02grouping them together,

7:03and then uncovering the likely next product of interest.

7:07If, for example, this type of customer buys that type of product,

7:10that kind of thing.

7:11And we also mentioned, didn't we,

7:12how machine learning models are very useful

7:14for automatically detecting anomalies and outliers in data

7:18and detecting fraud in financial data,

7:21being a great example here,

7:22and maybe spotting problems in a manufacturing process

7:25by analyzing quality inspection results, that kind of thing.

7:28So yes, we already mentioned those, didn't we, in the last video

7:32as examples of where machine learning adds value.

7:35So having got some ideas as to where AI could be applied in business,

7:39now you need to start thinking about

7:41where it could be applied to your organization.

7:45And Microsoft recommends that we, and I quote,

7:48identify and elevate business value,

7:52which sounds rather grander than it is.

7:55But it is basically these really sensible and practical three steps here,

8:01which are kind of obvious when you think about them.

8:03But the point is, somebody somewhere has to think about them.

8:05So firstly, when we're talking about implementing some kind of AI solution,

8:10we have to ask ourselves, what problem is this solving?

8:15What are the pain points?

8:16And again, I put pain points in quotes there.

8:18That's a term that Microsoft loves to use

8:21to describe any areas of inefficiency or slow processes or high costs

8:26or, I don't know, bottlenecks in your systems

8:29or scenarios where you lose customers to a competitor

8:32or you've got too many customer complaints or returns, whatever it is,

8:36identify the pain points in your organization.

8:39Because if we don't know what problem we're trying to solve with our AI solution,

8:43then how do we know if we've been successful?

8:46We're not going to know if it's delivered any business value

8:48if we haven't properly identified what problem is it designed to solve.

8:55And then number two on the list here, we've got to make it measurable.

8:58We've got to tie all of our AI initiatives to clear data,

9:03you know, identifying those KPIs, key performance indicators,

9:07such as time to value or accuracy or speed or, I don't know, time to close a ticket

9:13or the number of returns or cost savings or the customer satisfaction ratings.

9:18Whatever it is, make sure it's measurable.

9:20Put a number on it.

9:22Let the data do the talking.

9:23Because otherwise, how can we determine whether you've got a return

9:27on the investment that you made in your AI solution?

9:30And then thirdly here, does it align with your organization's priorities?

9:35Does it fit in with your organization's bigger picture and mission?

9:39You know, does it support your compliance regulations?

9:41That kind of thing.

9:42Because otherwise, I mean, if what you're trying to do

9:45doesn't align with your organization's strategic goals,

9:48then you're not going to get funding or support for it.

9:51So why bother?

9:52And at the end of the day, you have to show that your AI initiative

9:56does have a part to play in moving the organization

10:00forwards in the direction that senior management want it to go.

10:04So I think these three steps here is a great place to start

10:08when it comes to identifying where AI could add business value to your organization.

10:13But having done that, we need to think about how we,

10:16and again, I quote here, how we turn capability into outcomes.

10:21And I think these steps here would be performed

10:24in conjunction with your technical teams.

10:27But the point is, for business leaders, the question we need answering is

10:32how do we translate the AI capabilities into business value?

10:36That's the outcomes here.

10:38And we need some kind of end result that's a working AI solution

10:43that is repeatable and reliable and doesn't expose us to additional risk.

10:47But anyway, as I said, having identified the business problem,

10:50as per the last slide, we now need to be sure to look at this,

10:54match the right capability with the need.

10:56So this is all about matching the right AI capability with the need.

11:00Is it a generative AI solution or a predictive AI solution

11:04as per the machine learning scenarios we talked about?

11:07How much customization or fine tuning do we need to do?

11:10So matching the right capability with the need.

11:12And then number two here, investing in data quality.

11:15And again, this is going to be something that will be performed

11:18in conjunction with your technical teams.

11:20So having clean, consistent, and well-labeled and well-classified data

11:25is going to be the foundation of good AI, because without it,

11:28even the best models aren't going to be able to deliver what you need it to deliver.

11:33And then planning for operations.

11:35So you need to be defining how you're going to be measuring performance

11:38and monitoring the output and how you're going to be detecting

11:41any drift in the results.

11:43Or when do we need to retrain models as conditions change

11:46in order to keep the outcomes that we need for our business

11:49to get the value that we're expecting?

11:51How do we keep that output stable as things change and time passes?

11:57And we must always maintain human oversight.

12:00So the goal is that we're using AI to add expertise,

12:04to augment and enhance what our humans are doing,

12:07not to replace humans, especially for high-stake decisions

12:10where oversight and context and human judgment really do matter.

12:14And then, of course, we need to, well, measure what matters.

12:18And that's just a catchy way of saying,

12:20sure, having, as per the last slide,

12:22identified the metrics that we're trying to achieve here,

12:25we need to make sure we are measuring those outcomes

12:28that the solution delivers so that we can see

12:31whether we have achieved that business value,

12:34have we achieved a return on our investment?

12:36The overall message being, if we don't have the data,

12:39if we can't measure the outcomes that we're after,

12:42then how can we determine business value that the AI solution has brought?

12:46We can't.

0:00OK, it's time to review what we've learned with a few extra quiz questions.

0:04And the first one was one of those matching quiz questions.

0:07So match up the AI model with the business need.

0:10So if we have a business need of needing to create visual content,

0:14then that's going to be a diffusion model type.

0:18That's what we call those types of AI models.

0:21Interaction with different input types requires a multimodal model.

0:26Programming assistance. Well, hopefully that was easy.

0:29I was trying to avoid using the word coding here because that would have been too easy.

0:32But yeah, obviously programming assistance is going to be a code model.

0:35A customer support chat bot is probably a good example of an LLM,

0:40a large language model for natural language input and responses.

0:45And then look here, a compact language model for a closely defined scope.

0:49So that was the large language model's little brother, the small language model.

0:54OK, on to the next question. It's another matching question.

0:58And quite a few matches to make here.

1:00So I know this might have been quite tricky, but I really do want you to think about this.

1:04But this is all about the machine learning project lifecycle.

1:08So these are the different stages in the correct order here.

1:11Let's match it up with the description. So firstly, we need to define the problem,

1:15which would be described as identify the business question the model should answer.

1:20Then we collect the data. So there we go. Ensure access to the right data sources.

1:26And then we prepare the data.

1:29Yeah, there we go. So the technical people will clean it up

1:31and then leaders are going to ensure that it's accurate and complete and relevant

1:34and compliant and whatnot. And then train the model.

1:37Oh, yeah, there we go. That's where the model learns from the data.

1:40We deploy the model. So the model gets integrated into the business process

1:45or the app, whatever the solution is. And then number six here, monitoring performance.

1:49That's got to be the only option left. We track how the model performs

1:52and ensure alignment with business needs. Good.

1:56Okay, the next question then, which three of the following are use cases

2:00for a generative AI solution?

2:03So yes, summarizing documents, creating marketing content.

2:08So we're looking for three and drafting reports.

2:10There we go. That'll do. And you might have spotted that the next question is the same set

2:16of possible answers, but this time it's three use cases for a machine learning solution.

2:20So this time we're looking for personalization of product suggestions

2:24and I've just spotted a typing error there.

2:27So I'll hopefully try and remember to correct that before you get to see this.

2:30But also detecting fraud. Yes. And predicting customer churn.

2:33Good. And then finally, what should you do to convince senior management

2:39that the AI solution is worth the investment?

2:42So hopefully you disregarded many of these as being complete and utter nonsense.

2:47It's this one we're after.

2:48We need to show them how the solution will align with their strategic goals.

2:53Well done.