 I'm Saleem, so currently I'm working with PegaSystems as a user experience designer. So today I'm basically going to talk about how we can design for vice user interfaces. So I mean, if you notice, last past 20 years, every few years there is a digital disruption. So we started in 1994, there is a web 2002 search, 2006 social and 2010 mobile. So 2016 is going to be AI and vice. So I mean, the requirement is more for vice user interface designing. But there are across, I mean, industry leaders, even with the industry leaders like Amazon, Google and Microsoft, there is very less documentation available how to compile design and execute vice enabled products. So I mean, at Pega, we haven't started designing anything on vice yet, but I mean, this presentation is purely about how, I mean, how we can design vice user interfaces that is purely based on the research and my interest towards the vice user interfaces. So if you notice, I mean, there are around 500 plus million people in India who are yet to come online. So they haven't seen the internet. They haven't seen any mobile. They haven't used laptops or computers. So do you think vice, I mean, so vice is very accessible. It comes natural to us. I mean, there are so many benefits of vice. I mean, we start, I mean, as a child, the first thing we do is talk. I mean, it is very natural to us and vice is available everywhere. I mean, it is accessible for everyone. But we just starting to use the technology. I mean, vice enabled interfaces now. I think it is very, there are so many benefits of vice enable you guys going forward. So accessibility is one of the major, major thing. So that brings, I mean, we're going to discuss few use cases where exactly in Indian context we use vice enabled interfaces. So that brings my first use case, which is language support. So currently we, I mean, India is a very culturally diversified country. We have around 17 official languages and every 500 kilometers you travel from point A to point B, there are different dialects, people speak different languages. But we don't have a system where I can use my native language and get results out of it. So if we can, if you can get that, I think a lot of people, I mean, it is differently hard to bring that, but if you want to reach that 500 plus million people market, you need to have applications catering to them. I mean, they are definitely not as literate as they haven't used any applications so far. Only thing they can do is talk in their native languages and get, I mean, if we could bring results from their native languages, I think it will be a huge, huge market. I mean, if you could identify that one product, it will be a big success. So that is one use case. Second use case is agricultural, I mean, India is again, Indian economy is mostly dependent on agriculture. We are doing agriculture for a long time, but we are not using any technology in agricultural sector. If you notice, I mean, even US and China, these are the countries who use most of the technology in agriculture and they produce more yield than India. I mean, traditionally we are saying we are agricultural country, but we are not able to produce that yield. Our other countries which use more technology, they are doing well in that country. If we could utilize AI and vice-enabled interfaces into that sector, I think that will be, again, big success. So we have data. I mean, we have a lot of data in terms of, I mean, our weather and we have, I mean, if you can provide what is the best thing, if we produce, I mean, how it will get better yields. If we could do that information, if we could provide that information to the agricultural sector, I mean, farmers, I think to be, I mean, again, to be used in their own language, I think that that is, again, combined with use case one and use case two, it will be a huge hit again. My third use case is healthcare. I mean, again, healthcare industry, we, there are a few applications which can be used for our common, common issues like everyday scars and things. I mean, I can ask some applications to provide me some tips on if there is something happened to me, but we don't have a bigger conversation. I mean, we cannot converse lengthy. I mean, we, I mean, if we have, I mean, we have very dangerous diseases like cancer, diabetic. Using AI, we could identify those diseases very early, but we are not able to give that information using wise-enabled interfaces like, I mean, if you notice, even in Bombay, Tata Health Cancer Center, there are people, I mean, around 70 to 80 percent of people who are coming there from the rural areas, I mean, they're not part of Maharashtra. They're not part of Bangalore. Only 15 to 20 percent of people are local people, but a lot of people are coming from the outside. They are, they don't know the technology. I mean, if we, if we could provide some insights to them, early insights to them with respect to these diseases, I think that is again a big success. I mean, we could do a lot of things for these users. Design process, I mean, it is definitely, I mean, we have, we have identified these use cases. What is the design process we can use to, it is definitely, so traditionally we are always doing graphical user interfaces. We, we haven't, we are just starting to do voice user interfaces. So, what is, I mean, the design process is typically different from what GUIs and VUIs. So, what is the design process we need to follow for the voice user interfaces? One is your conversational anticipations. So, as a user, I mean, as a designer, you should be able to identify what are those conversational anticipations, what are the keywords user is going to use. We have to anticipate that first and provide your inputs to the machine. And second one is adaptability. Design adaptability is user can, I mean, there could be multiple things user can say in different types. For example, one example I can see is, I can say my vacation starts tomorrow, I am out of office tomorrow. So, both are basically the same things, but different ways user can interact with your machine. And we should be, I mean, as a designer, we should be able to anticipate happy path as well as non-happy path to the user. As a, I mean, graphic user interface design, you have navigation where user, I mean, he can see the navigation structure, even if he lost somewhere, he can come back and find his way. But in graphic, I mean, in voice user interface, there is no visual queue for the user where he can, if we are not able to provide him with the proper navigation structure, there will be a lot of error messages will be on for him. I mean, he cannot come back to what exactly he is looking for. So, how we can, I mean, what is the design process we can, so what I, I mean, there are different type of design process we can approach. What I feel is the majorly three things. One is user, one is user-centered research. How user can do the research for coming up with these applications? Second one is iterating, sketching out your interaction flows. So, basically, before doing anything, I mean, first thing is you need to do user-centered research similar to how we do for graphic user interfaces. For VUIs also, we should do user-centered research first, then sketch out the interaction flows. How the dialogue, I mean, we should be able to sketch out the dialogue flows as well as interaction flows. Third one is testing those interaction flows. You have to test those interaction flows with your colleagues or there are some applications where you can test them as well. So, as part of a user-centered research, what I mean, these are the four things as a designer we should be able to answer before moving ahead with the interaction flows. One is what is the purpose of the system? So, identify how, why exactly we are doing this application and how user can invoke. So, for example, for Alexa, you say Alexa or computer, I mean, how users can invoke your application, that is one we have to identify. And what information does system require? So, that is also one major thing we need to understand. So, mapping out the interaction flows or sketching out the interaction flows is the second part of a design process. So, as a designer, we need to identify what are the keywords user can interact with our applications and what are the, I mean, welcome, messages. All those things we should be able to provide as part of that interaction flows. And entry and exit points of conversation and example dialogues. As I said, we should be able to write the example dialogues starting from the, I mean, you take the use case, start the, I mean, writing those dialogues and you should be able to complete that complete flow before doing this. Next part is testing those interaction flows. So, there are few tools from Google where you can test, otherwise you can test these interaction flows with your colleagues as well. So, as part of these testing, you should be able to test conversational dead ends. What are the conversational dead ends you need to test? And interaction where flow is chopped. I mean, if you test this with your friend or with your colleague, you should be able to understand how this conversation is taking apart. I mean, while doing these testing, you will come to know if, as part of that use case, if you are missing something, you should be able to understand that as well. And the system feedbacks, I mean, those are important for user, I mean, if you are not able to, we should be able to guide users through these system feedbacks as well. So, we have taken, I mean, we have seen the use cases, we have seen the three design process we can follow to design voice user interfaces. Now, we are looking at the design, I mean, actual designing and testing, testing these VUIs. So, where exactly will start? I mean, first thing is, we need to identify focus groups. How do we, I mean, we can do this within our, I mean, within our organization, we can actually employ couple of people from different age groups and different departments. And we should be able to, I mean, tell our problem statement and we should able to take their inputs for this one. And second part is, as I tell, flow design and sketching out the flow design. This is, I mean, example of how we can design, so we can use tools like dry-o and visual usage charts. I mean, there are other tools also we can use, but majorly our focus should be how user interaction with the system is happening. That is how we should be able to do it. I mean, again, as part of designing, role play of usability testing is again very important. How we do it, I mean, the goal is to identify, again, what is the feedbacks and all. Based on the feedback we have received from the user testing, we should iterate the interaction and based on the feedback. So key insights of, I mean, key insights are the, even we are getting started, but there is very less documentation for a designer to get started with the vice user interfaces.