 So today I'm going to take you on a journey of how I transition from a designer to a designer. So today I'll be taking you on a journey of how I transition from a designer into a designer. So this talk is about incorporating generative AI into product as a management to bring down the code. So just like any other traditional dedicated US designer, I'm also passionate about designing users and experiences, bringing down cognitive load and delivering some great products. It means a lot of strategy, a lot of purpose, a lot of that professional understanding that I have. But then there was one question that we always had. Am I doing it right? You see that uncertainty always existed. I was hearing AI like this, AI like that, AI was everywhere. So I thought, why don't you use AI? Why not take AI's help in itself? So that's when I discovered the AI revolution. The AI revolution was here. Well, quite frankly, it's not new. AI is not new. It's new to us for a while now. But then generative AI in specific is something that is different right now. A lot of companies are making a run for it. In fact, we as a designer have been using generative AI tools in our design work, design process to make things easy for us. Maybe for, you know, like in this story, you see the person who has information architecture, the site maps, even if you do that. But then half you talk about, are we making the users like this? How do we ensure that the user experiences that they have? How can we benefit out of it? How will users benefit out of it? I've been thinking, you know, I found that interesting. So that is when I decided to incorporate generative AI in my app. But before that, I wanted to know one single question and I asked myself what value does AI bring into your product? Maybe say money for users or maybe reduce corporate portfolio or improve efficiency or maybe some of the value that you find possible. But be very clear about your goal. Because this can make you create your design for your product. All right. Now, then heading to the implementation. Generative AI in as a product. I have three simple steps. I'll list it down all the primary workflows for my product. As a maintenance task, health monitoring, inventory management or post-production. Because I wanted to perform or bring down the corporate workload, I performed the corporate workload analysis. I identified the areas where there was complex decision making, data interpretation, management tasks and information overload. So by this time, I was clear, you see, I identified the units where generative AI could sit in. So this can actually help improve the product. Let me give you an example. I'm taking asset maintenance as one. So asset maintenance, the user, as I said, manager, usually maintains the process of inspecting, servicing, repairing of the company assets for their longevity, save time, save costs, all of that. How is it traditionally done? The user must check through the equipment, equipment, conditions, and all of this data from their dashboards, record pages, and all of the data. And then they also have to analyze historical data. In the data you're bringing back years ago, we were talking about hundreds and thousands of records. Just imagine the mental effort the user has to go through. The corporate AI actually, you know, clear role in this is by analyzing the equipment data. Now, you can predict the maintenance needs and these are predictions for expert time. In general, it's optimized convenience schedules. So imagine you have a lower maintenance schedules. How do you know something is effective and efficient amongst them? Genea can do a good job with that. You know, award or, you know, you can do a good job with that. Okay, so setting back, setting back on how to do AI has actually got down on the company. By analyzing the equipment data, we have brought down the company data by making the data identification numbers. Predicting maybe the same speed from efficiency by acting down the management data. So generally, optimized maintenance schedules we again got down on the company data by making the decision making easy for the user. I'm sure you might be curious to look at the screen because it's management where a generative AI is applied. But then, it's going to be a bummer. I started to disappoint you. You're going to be able to show those screens or pinch out the issues. But then, let me give you one better. I got some generative AI implementation tips. You know, the tips that I've got. Tip one, don't force AI on your users. You see, provide them an option that's flexibility. Maybe by using SSD functional element, an element that can stand out on the screen and can be accessible and returnizable by the user whenever necessary. Because we, you know, in the time when the user feels helpless, they were treating that button to, you know, use the way AI exists out of our system. Next is GVD chat box. The most common way of interacting with AI in the system, GVD chat box. A lot of companies and, you know, products are already using their support. How do you make it more effective? You see, by providing some question solutions, don't just give out a black mask to the user based on the text box over the button. But then, you know, analyze the user behavior that the generative AI studies with user partners. And let me give them some kind of generative questions that can help out the user in, you know, asking about your questions which are much more effective and they can perceive the generations in a more effective fashion. Tip three, use a control of freedom. Once there is a generation happening, and once you have the answer to it, provide the option for the user to implement it. When you regenerate it, and in fact, even if you vote it once in a while, that way, you can control. And there is one rule. They will have the flexibility and the freedom that they want. That we vote about it. But quite frankly, all of this can happen but then it's definitely not the same thing. As in, you know, there are always challenges around. Always need to look out for bias. So what's bias? AI runs on algorithms that have a lot of data. You see, this data gives up the generations that we have. So, you know, there are a fair amount of chances for you to see data that is biased. So be sure about it. Let me just go about it. He was not talking about it. Transparency. So, let me show you some generations to allow the time period in case there's something that's limited to that period. For example, a chat review is a great job calling out that it's database is limited to 2021. Give me the same for user. Can we help? Explain your data. User shouldn't blindly follow what AI generates. You know, they have to understand how to understand the context of bias they've been generating in the past. So, let me give you an example. When the user allows for a content summarization from a page, that's a page that might have many different sections. Calling that out may be a sectionized manner. You know, explains it better and makes it more comfortable for the summarizations. Lastly, ethical implications. When you talk about data, there's always sensitive data. You know, there are data that, you know, financial information using these passwords. So, these are the kind of screens where you want to feed the convention. You don't want the AI to read over it. So, be sure to draw boundaries and show that you have clear for a lot of what the AI should be and what the AI should be. So, with that said, I'd like to conclude by commenting on the user generating the AI in your applications. And, you know, bring down the quality of the user experiences. In fact, we use this now for the exact best philosophy. So, I hope the conventional methods may also enhance the quality of the AI that it will bring products to value partners. Okay. Yeah. So, yeah, well, we can talk. Thank you very much for being a great audience. I hope this talk has been insightful and considerate to viewers and ideas in your applications. Thank you very much. Questions on the stage. Thanks.