 So today I'm going to talk about the UX of data security intelligence. Let's understand what we're doing at Informatica. I lead the product team for Informatica USA on for the data security group, okay? So what is happening? There is like a lot of things happening in industry, a lot of transformation happening in the industry. So earlier we used to have previous dialer phones to mobile phones to now everything is digital, right? Even like everything we are storing, our fridges are talking, our thermostats are talking, our TVs are talking, everything is getting digital. Not only that side, if you look at the other side of the spectrum, like you know, industrial internet revolution, our Boeing engines are talking and all, what is this happening across? So there are from a consumer side, there is a lot of these digitalization happening. And same from the industrial revolution side, it's happening. The machines are talking a lot of these happening on the technology front. So what is digital transformation again? So the digital transformation happened. Digital transformation is all about data, right? The transformation happening what you're doing is you're doing with data pack. Let's understand a little bit of data, how data is stored, right? In the initially, in the data 1.0 era, the data is used for businesses in terms of like you know used to reservation systems, booking rail with airplane tickets or salary automation. So that happened at the 1.0 era. This is I'm talking about more of the 30 years spectrum kind of like 1960s, 1970s and so on. And then 2.0, which is basically 20 years old, which we are talking about is all about the business automation kind of thing. So where like you know, you get this CRM system for the business processes and stuff, you're actually in HRM systems and so on and stuff like that. The data is used for, that's all about data 2.0. And now and the future, it's all about data 3.0. It's all about the transformation which you are looking at. The data is the fundamental base of digital transformation, okay? What's happening here? So you have data, data is all about like the credit card information or your personal information, data birth information, everything. Everyone want to steal data because data is fundamentally a new oil. Data is so valuable, everyone want to steal it. That's the reason if you look at that, right? Number of cyber attacks you keep hearing, whether it's going to be Athena or basically like, you know, the key facts, like, you know, target, all these breaches happen, someone is stealing data, right? How do you predict data is one of the bigger things. And also, across the countries, every country have a lot of regulation related to that predicting data. For example, GDPR in Europe, it's mandated by the government that you have to predict your data. So what is in it? The security is one of the key factor about it. So you can't leave your data just like that. So 10 years back, it's more about like a perimeter defined thing. It's like, you know, it's perimeter, you are actually, your data is locked into your data centers or control access you're managing like. And five years back, it's more of isolated data security where you are actually doing at a broader level, organization level, let's say, this is the department level or integrated file, email, disk, encryption kind of thing. Future, with these level of disruption in the data, future, it's all going to be the data centric. Like, you know, you can't like, you know, protect at this level. If you have data level, if you have a credit card numbers data, you should able to actually protect at that level, right? So that's called data centric security. So when we look at this problem, when we start looking at this security problem, there are fundamentally four challenges. One is lack of visibility. You don't know like, you know, what is protected, what is not protected. And second one is detection part. Detection takes a lot of time. So it takes a lot of time to inaccurate, slow, inaccurate detection. Once you know it, right? Third one is all about the inability to prioritize. How do you prioritize stuff when you have a lot of stuff to do? So that's one problem. And inadequate protection techniques. So this is one of the challenge in terms of how do you want to protect there are a lot of methods. But it's not like, you know, realistic data centric way. You can protect it. When we start looking out this problem from a solution standpoint, okay, we center it like, you know, let's do it a data centric way. It's called data centric way and put intelligence in terms of, on the top of it and see how intelligence will gonna help this data security problem. So this is the way we did. We look at the whole data pack. You get 360 degree insights. So it's, we call it as intelligent detection. And once you detect it, intelligent production. Whether it's going to be inside that direction, inside that means like, you know, malicious user or someone like that or multi-risk factor analytics, which is both of using analytics to find risk, some measurement to that or orchestrate production. So what happens is when you look at a problem, when it, so when you look at a problem, it has a scale, right? So when you have a lot of complexity by axis, you're gonna, you haven't other axis, you have a time. So what happens is when you're solving a problem from complexity over a period of time, your complexity comes down and your design opportunity keeps growing because you're dealing with it tons of, tons of data and you're dealing with various interesting opportunities around. So let me share you some challenges which, when we are creating this product, what are the five challenges which we realized and how did we overcome? So this is a little bit of a case study. I'm going to show you some design, how we solved it. So first problem, one is new market category. So for example, what is new market category? For example, let's say when iPhone first launched, there is no user using it, right? So there is two aspects of it. One is basically what user knows because normally no one use touch, like user know nothing about it. And second thing is what system can offer? There are two angles to it. If you look at two, right? So it's a pretty symmetrical way. There is a traditional security systems. When you start looking from this angle, what we did is when you're creating something in new market category, you don't have any references. You don't know how it looks like. You are just experimenting. You're based on your design, right? What we tried is like, you know, you need a lot of these design iteration, expedited iteration and design thinking, ideation of these concepts. So you don't have customers, you don't have users. How do you deal with it? How do you design a great old-class product? So one what we did is we said, okay, let's visit customers and talk to them and see, like now, we did a lot of visit multiple customer X or multiple customer streams and started looking at in terms of what are they doing? How are they organized and stuff like that? And we create customer advisory boards. And we created designs and started validating with the advisory boards. And we started, so even like, you know, being a designer, we used to show up in the trade shows, the company trade shows. So we used to be shared of a salesman and start listening to why someone is buying your product. Or basically, who is the buyer? Who is the seller? Who is the user? There are three dimensions of it. It's not just user work. If you're looking from a business angle perspective, you should also understand who is your buyer persona. Because initially what we did is we totally, when we did all this research, we realized, like, you know, we are targeting wrong user. Our buyer persona is very different than whom we are actually targeting. Same with, I'm gonna show you one more couple of things when we did this, like, you know, personal research project, what we realized. And hands-on labs and internal audits and basically expert reviews is something which you see. This is my other team of colleagues who don't know anything about my product. So they were doing a complete expert review and doing completely documenting of what is the feedback. Because in other way, we were also looked about is proxy users. Proxy users are users who is basically like, you know, assumed users kind of thing. So pulled someone from other departments and started validation and stuff. And third other thing which you see is, look at that picture which you see, is basically we said alpha, beta testing and invited all the sales people to come onto the board and test the product and give feedback. So we got piles of, like, you know, half of the queue was done by them. That's why we collected lots of feedback. So when you map all this feedback from various way, you can make some sensible information about, like, you know, how do you want to ship the product. Second problem, no limited access of users, users place data. When you're building a new product, you'll not have, obviously, don't have any data, basically, because you don't have any usage data. And users also, because you're selling something new, based on your assumptions, based on your, whatever you believe in, what are your business model, who you're targeting, you're actually, you're learning at that particular term. So a few things we have done. One is we recruited proxy users. Basically we have some agencies helping in San Francisco for us to hire proxy users for testing our designs. And second interesting thing which we have done is basically this streamline exercise with some of these customers. We went to customer sites, and said, ask them to draw on a board. How is your hierarchy looks like? If you look at this, one of the streamline from various customers. So these are the various steps used. Yellow ones, what you see, is various steps involved. What do they do? And red ones is all of the pain points. And they're hand-offing, handing over to different, different people. It's not one person doing it. So that's how we can play. Third is basically big data. You all know about big data, right? So big data is like, you know, piles of data. So how do we get meaningful information on this data? And how do we compute it? There are two parts of it. One, how do you analyze that particular data with lot of information? And how do you make meaningful and sensible information and bring your user experience to that? That's one. And second thing is how technology allows to that. That's also, I think, the designer's jobs to think about it, right? How do you compute it? Because if you are creating a system, it's gonna take like, you know, two days to load, it's useless, right? So how do you smartly design those things? For that, what we tried it to do is basically use of power of data visualization. So by data visualization, you can actually do pretty interesting stuff. These are some of the designs which we designed. So we try to use all the data visualization techniques, how you can actually make meaningful of that particular data. It's actually computing a lot of information from various systems across various departments and stuff. It's all in cool dashboards, it's actually showing up. All these things can be configurable and it actually shows up based on user what information is required. And also, there is a concept of like, you know, how much you can need to show. There is a big balance between when you are designing dashboards or the UX part. It's also like, you know, how much is required? Is it information overload? Or are you giving only information which is required based on the user persona perspective? That's how these were designed. These are a couple of screenshots. And what you see here is, I was talking about power of data visualization. So there are multiple ways you can look at data. So in terms of like, say, you are looking at a larger volume, where is your risk? In this case, we're looking at where is your risk? So the top one is telling about where is your data moving? Let's say your credit card number. So where is your credit card numbers moving from where? And is it protected or not? Is there any risk involved in that? And here, in the tree map, we are showing in terms of like, you know, which department is risky? Whom likely, like, you know, so that you can directly go and act on that and do that. And some of these jobs or some of these actions in terms of users, what they are actually doing or they're downloading something a lot of valuable information. Or the way data visualization really helps you to think optimized data. And also like, as a designer, we should be smart enough how we want to, we should not overuse it, but we should make it meaningful in terms of how we want to use it. For example, let's say, when you have computational data of like a lot of records, most of the user segment, like, you know, they like to work on the black screens and they want to work on the like, you know, tables and grids. So that's also balanced in between, as a designer, we should take a call in terms of the filtering or the how you actually want to slice and dice that particular data. For example, let's say, if they want to look at the table view, how can you actually make, make that particular tables interesting? That's an example. These are a couple of other data vigilations which I created. So it's computing from various systems. So I did a lot of design patents on data visualization. Actually, this is also this problem. When you have complexity, it's fun, guys. It's basically you will not have any references to refer for, but it's really fun because as a designer that challenges us to really think about cool ideas which don't exist. So these are some of these. So this one basically we made almost 90 or 70 claims in one design patent. So this is purely a design patent in terms of how you actually slice and dice and filter out all this data and looking for various dimensions of that. Another thing, trust, data accuracy and high cost of value. This is one of the key thing which always like more important in terms of like trust. How do you build customer trust? And how do you actually, for example, let me give an example. So you all know about the Tesla accident, right? What happened is basically the system was fine. It was in an autonomous car. When it's driving, suddenly a big white trailer came in between the road and system could not able to recognize that particular white color, a flat big surface beyond the white sky. Okay, that was the middle of the day. And it had, unfortunately we lost Mr. Brown lost his life. So what it teaches, for example, machine critical system, think about this. If a system happens something like for a company like this, company will go bankruptcy or like no company will go get into loss or data breaches happens, it's going to be hard. So whatever we display, this should be accurate and the user should know that's very important even though it's not basically in our control but we should push our engineering teams to get more accurate data. If you don't have, so even design is great but your data is incomplete, it's not again a right design because it's design and development and design and it goes everything hand in hand. So whatever numbers you show, you should really make sense and also the way you compute that information. For example, let's say you can't actually get the right information in one go. How do you do it? So one smart way we did it is basically we said like, first we go and scan easy things, go and scan it up and slowly in the background slowly it actually increases. That's the reason what you see that dotted line it's scanning behind this picture. So at least you know that high level information where is that you need to focus on. Fifth, designing intelligent design. So we keep hearing about intelligence, intelligence. So what does that intelligence mean? Intelligence mean about like, it's not that just basically system is intelligent. It basically helps us use it to take actions automatically. The system predicts and do. For example, let's say take a couple of examples, right? So in machine critical systems like, let's say aviation, right? For example, let's say, if completely like flight is completely automated, the pilot loses his skill, right? So basically like if there's some, because like he's doing nothing there, just basically sitting and doing it. Or second thing, think for this robotic missions, healthcare, in the healthcare thing. So if system is doing everything basically like, you know, sometimes if something goes wrong, doctor, there is a risk of these killing skills. So here in this, if you look at this, right? There's a good opportunity. This is a UX problem, serious UX problem in terms of how do you actually give control to the doctor back whenever required or maybe the system intelligently give ask doctor, like, you know, okay. So this is the situation what should I do? Or some interaction where doctor knows what's happening or say with the pilot knows what's happening. It's a combination. So it's called, that's a concept called mixed initiative. Mixed initiative is basically AI plus human. Fully automated systems also sometimes have risk. For example, something goes wrong. It happens like, you know, so the company gets into risk, like, you know, we can't afford to take those kinds of risk, right? So in that case, as a designers, that's a great opportunity for us. So we keep trying out various concepts in this. We are innovating pretty fast in this particular space in terms of how systems can be smarter than us and tell us before a friend before doing it. So, and in fact, like, you know, the way we're pushing ourselves is basically we are thinking about what patent we can do. Can we make a design patent which don't exist? So that's how I'm pushing my team and as a team, including developers, we are trying out and see, thinking about how system can be smarter than us. And but still have machine and AI, we can use these mixed initiatives. So in case, for example, let's say, how do you predict, so it can be a predictive analysis or it can be basically using deep learning. If like, for example, let's say, someone is downloading large information of data in your, from HR system or like enough system, system can tell like, this guy, how like, you know, most possibility that, you know, he's gonna live in next week or whatever. So we can predict all kinds of things. So a few things based on my experience in terms of like designing this system. So one is basically start simple, land and expand. What that means is basically, you start something simple, don't complicate stuff even though it's basically a bigger one. Prior to this information, what is the key things? It starts simple. Once you launch it, right? Don't build all features for the sake of building features. Okay, you don't need like all features at version one. You need to learn or you need to understand that particular over period of time, you get to know what exactly it is. And second thing, most important thing is, no one based someone shouldn't, we should drive innovation. We should take a lead to drive innovation. Like no one will come and tell us like, you know, okay, you are a designer, go and innovate. As a designer, we should lead it and say, okay, guys, I have a cool idea. So let's explore this and let's do that. And if you show idea it more and come up with a lot of design ideas and do a lot of validation, you are really, really, really respected. And as a designer, we should drive that. And third one is basically, once you launch it, improve it. Basically, I can learn for like, how user is using it and learn based on, then slowly you can also upgrade your algorithms and stuff like that based on your usage data, which is also very important. So we did this and I think it's been three years, this product is in the market. So we are doing three releases. And so it's done. So we won 23 awards for this product. And this talks about like, you know, how rapidly we are doing and how rapidly doing. And what I would say in the last is, design plus data powers digital transformation. Thank you guys. Thank you. Thank you. Thank you. Thank you.