What is UX for conversational UIs?

hi everyone and thanks for watching this presentation about conversational interfaces but it is probably the password of the moment I hurriedly created this deck as an internal training for my department team at AKQA London where I work on where around 20 to 23 UX designers share our passion for user experience design whatever that means nowadays so the main idea behind this talk was ok here you have a summary of all those books and articles you didn’t have time to read on all the conference’s you could have returned Samara is for you in less than one hour so I choose books like best interface is no interface by Colton Krishna don’t make me tap the conversational interface by micro McTeer and many many other good books and articles on this topic so I thought that after having spent so many hours summarizing them it was a nice idea all releasing this for you also so before I start and just in case you don’t know me yet when I’m not starting my workmate with a camera I mainly work try vintage user experience of one of our clients or it in Atlantic airlines my yet emotional background is bitter science and engineering and before I started full-time on UX just around four years ago I worked another five six years as mobile ok and what I think next 25 minutes are going to be absolutely essential for your career because of this chart chat pod computer programs designed to simulate conversation with human users for example inside haze messengers or slap if you’re watching this I assume you know what a chat pod is but this is not a talk about chat pods okay II despite chat BOTS have a lot to do with this top as we can see suddenly in these few months interest about chat box has gone through the roof and if you repeat the search with conversational user interfaces or other related terms you’ll get similar try the thing is okay why is this chart like this probably because this are the childhood but this time the sternum acid that is the last year 2015 for the first time history users spend more time on messaging apps than in social networking apps the apps were traditionally they had spent most of their time and why is this so important because we’ve seen a paradigm shift every decade or so in the 80s I’m the common light to the desktop with all these metaphors about folders and windows in the 90s with the web in the 2000s with the smartphone and mobile apps and we mind say that last year we have another paradigm shift even if we didn’t notice messaging apps and this is important because did you sir is like a moving target and they have moved and we should engage them wherever they are and now they are here using massively messaging apps and this is a great opportunity because last few years engaging users through native apps have become very very difficult difficult in terms of money for example with apps becoming more and more expensive to create a meantime let’s think for example about that romantic idea of two developers in the coverage create in the next killer app that’s not true anymore just have a look to the top 10 20 apps on any store all of them have been created by big big content but these apps are increasingly complex in terms of interface tooled with more and more screen sizes and interaction design toggles and micro interactions and animations and that’s fine but if we want to offer a simple service just one unique feature or establish a personal relationship with our customers maybe a chat BOTS inside Facebook Messenger platform is enough we don’t have to create and maintain a complex interface and the deployment is not easier as well think that deploying Android apps sometimes is like a nightmare and get your iOS app proof may take even weeks sometimes we don’t have to worry about that with a chatbot but there’s another problem going on here it is called app fatigue most users don’t load 0 apps in a typical month it’s not a apps aren’t useful anymore more than half smartphone users access

apps every single day but people just don’t really have that many apps the apps people already have are normally more than love with this problem specifically chatbots can be very useful but again this is not talk about checkbox checkbox are just part what we are going to cover today text-based conversational interfaces but we are going to talk a bit as well about services like Amazon Alexa all it’s based conversational interfaces or about personal systems like Siri who can get information and others using voice but can deliver from each industry is indexed so it might be here in the middle of this is – another interesting question is ok why now conversational interfaces have been science fiction for a long time and the users have always been disappointed by them after a few minutes playing with this interfaces ok it was clear that they were not working so why now again because let’s call this day renascence of the artificial intelligence AI has come through cycles of euphoria and rejection for decades but last few years it we’ve seen amazing improvements on this field some of them rely on solutions like cloud computing or big data because advances in language technologies speak to recognition text to speech and the more conversations we have with this interface is the better to get the emergency of Semantic Web so machines can access this structure content makes sense okay on improvements on device technologies and connectivity we we have seen this article’s power yes new iPhone 6s has a computational power of 1,000 times the auto 11 computer about Allah but this is not just a matter of computational power it’s about having GPS having Wi-Fi 4G and Bluetooth accelerometer so these devices have context for that information and overall because the interest of companies like Facebook Google Apple Microsoft this companies are investing are ridiculous amount of money on these technologies and that’s tracking more and more talented people and more and more companies in a feedback loop so this is like a perfect storm if you think about this we have a complex scenario full of opportunities and finally technology a set of technologies ready to be used so it’s perfect so after this brief brief introduction about why suddenly gonna shift on interfaces are so trendy what’s happening on the apps market nowadays and the problems we are facing there and why conversational interfaces may play an interesting role here what else are we going to to see today okay I have divided the rest of this presentation into another two more small sections in the first one we’re not going to the crowd conversational interfaces we’re going to talk about the problem we have the UX field a big massive huge program the problem and how talking about conversational interfaces might be the perfect excuse to start solving it in second section we are going to dive in a bit more into the relationship between conversation interfaces and user experience and we’ll see some guidelines and useful tips and tricks I love this quote by Donna Norman by the way the real problem with the interface is that isn’t interface and I think that word he was trying to say is that he didn’t want focus on an interface to solve a problem he wanted to be focused on the problem itself and every interface normally adds some cognitive load like the feeling complexity layer on top of the program anyway as I promised the problem we have right now in the UX field as UX designers let me put it this way I have had an iPhone since 2008 and I still struggle typing on the floods have you gotten used to type on a flooded screen because I haven’t for me I think it’s even worse because half of my friends are Spaniards and the other half are English or English speakers so most of the times I write something on what’s happened by

the time I have sent a couple of phrases I realized that the keyboard was on the wrong language and what I send it doesn’t make any sense thanks to the character and it is like ah again I I know this has been partially solved by August and since I gave this talk but I still struggle seriously for me sending a message to my mother for example most of the times should be as simple as things like hey Siri send a message to mom I’m on my way my iPhone should say something like okay something to your mom Oh Alexa play the white album by the Beatles and we have any other clue or feedback just to start listening to music or something a little bit more complex like this I think is that there’s no reason why most of these experiences can’t be delivered at the current state of technological development I know that theory still sting speed but we can find quite appropriate use cases where deliver experiences like these ones I think is that all the interfaces have been created so far have had something important going for them they were all artificial creations and they were all invented for the purpose of enable or enabling interaction between a human on a machine with a very specific technological context and it was always a human who had to adapt to that interface who had to learn a new set of rules it was never the interface who adapted to the human being and this is even more important now that we are entering into era of the Internet of Things we are going to interact with these objects more and more it’s not just our phone or our laptop now we are interacting with our SmartWatch with our thermostat even with our cars every single interaction has a cost I have to for example take my phone from my pocket I have to unlock it I have to browse and swipe and I have to use my eyes I have to use my hands there’s some cognitive load because of this interface saying this is scenarios for example voice is not scalable because it’s cost is much slower nonetheless our love for the digital interface is out of control did you want to meditate there’s not for that do you want to lose weight yeah there’s an app for that and as content Krishna wrote in best interface is no interface those five words there is an app for that in some way my heaven intentionally founds some creative limitations of us let me put it this way a few years ago most automotive brands created their own door opening up so if I have a smartphone appropriate card I was able to open my car with my so just for the sake of demonstration let’s see how this apps work let’s imagine this is me and I’m walking up to my car so I have to pull out my smartphone from my pocket and I have to wake up my phone I have to unlock my phone I exit last opened up I exit last opened group I swipe through millions of icons trying to find the appropriate up and when I finally find it I thought yup I won and I wait for the app to load and I still waiting and then I make a guess like hmm this looks like the control to open the door and then I tapped you on the button and I slide to confirm that action because we are talking about opening my car so this should require like a confirmation for sure and then I physically open the corridor so this is me this is my operative and everything in between is a digital stream based interface this doesn’t make sense at all couldn’t we just use the Bluetooth of my smart phone to send sign-on to my car sinn fein close to my car and I pulled the door the door just opens okay let’s see some more examples this is one of the first buttons of a car it was created like 100 on many many years ago and the guys who created this patent solved a real problem with

transportation and this pattern changed the world so 100 and many many years later how do you make a better car yeah just lovely interface of it sorry for this silly example let’s see another one this is one of the first buttons of a fridge it was created again like 100 and many many years ago the guys who created this patent solved a real problem having fresh food at home so again one hundred and many many years later how did you make a better fruit yeah us labrum interface on it and by the way is this Evernote okay let’s see the last example how do you make a better Tresca I hope you got this publish one hot I guess okay by turning it into a 30 mm trash can 32,000 pounds trashcan sorry so you can see if it’s raining outside when you are outside standing in the rain yeah guinness so sorry about this trunking and mentions as well that I love that from keynote let’s see a real example a more serious one let’s try to open the trunk of our car we might create an app with a pattern to open the trunk we might just add a sensor below the trunk so if I move my foot with my smartphone or my keys on my pocket even if I’m using my hands the trunk opens it is a better way to solve the problem okay so now look at this job offers they are real job offers from well-known companies of across the world what’s the problem with them the problem is that two different disciplines as user experience and user interface design have been blurred here I’m blurring them has played an important role in our approach to create technological experiences when we hire someone good creating user interfaces we won’t probably get new innovative solutions we will get more user interfaces not necessarily better user experiences and no orphans visual designers the thing is that somewhere along the way we confused the two and instead of creating the best solution for the problem started solving problems with screens because that was sometimes part of our job description and UX sometimes stopped being about people and started being about around the rectangles and parallax animations so we create some more than inspirational boards and user stories and persona scenarios as summer experience maps and then the creativity stops and we start creating wireframes and mock-ups and sometimes we do it again and again and again on no matter client in the street problem we start asking questions like okay action or scared or discussing about libraries we are going to use to create the interface and then the burger icon discussion and that’s a thesis not because the icon itself but because talented people discussing just about solving problems with a screen based interface is a disaster and I think that’s why I’m more than happy to work at a QA because I know that useful and elegant solutions are exactly what we’ve been creating all this time but the question is okay good we have solved any of those problems in the past on a better way without having used an interface and I really think that that’s an interesting interesting question so what is UX for conversational interfaces well I don’t know yet but what I do know is what conversational interfaces are for UX a great opportunity to rethink our approach to solve problems in the future and I really thought it was important enough to spend 20 minutes with it so if the first part of this presentation was against the screen based thinking the second part is in failure of conversational interfaces overall yeah oh it’s user interfaces as a better approach to solve problems as we saw with the normal squad the closer we get to a human natural interface the easier we’ll be solving programs tasks that would require

multiple complex steps like browsing text input tabs everywhere now can be a teeth with a single voice comma don’t worry anyway if your chat bot fun because many many of the tips and tricks we’re gonna see can be applied as well to text basic conversational interfaces okay but why voice specifically why I think it’s better than text let’s let’s think about this for a second the average person speaks around 16,000 worth today if we wanted to achieve that amount of words typing we had to spend eight hours in a row without doing anything else just type that’s crazy so the bottom line here is that in terms of pure volume with a liver and process audio far more than we do through all the other media at all voice also communicates gender identity personality mood emphasis with minimal effort ice-free and hands-free this is very important when driving for example on a passive way and without any specific training from two three years all key to any age anyone can use voice the amount of accessibility problems this solves is amazing and okay what can we do as UX designers in this context this is the cool thing the same we do every time we start a project first of all research all our research skills can be used to want to walk into this type of project statistics interviews tools personas customer experience Maps everything but it is with personas that we find the first particularity because given the fact that always communicates personality we should create personas but just for the users of the platform but for the system the platform itself we have to define the voice user interface persona maintaining consistency in the voice and Tai you’ll see how palm to writing becomes much easier if we have a well-defined persona for the system would my persona say what is your age would it say how old are you would it say I didn’t get that or I didn’t understand that the answers to these questions are usually obvious once you have defined the persona the challenge for the UX designer here is to come up with a persona this is a voice and a personality but best fit the brand of the company and is most appropriate for the purposes of the interaction have you ever thought about creating a persona for non-human entity it’s crazy isn’t on later on in the project the best way to test our persona is at the Wizard of Oz test where a user interacts through a microphone and speaker with a human pretending to be the system probably us again UX designers so in some way we don’t have to create a prototype to test the interim interface we are the prototype this is weak by the way I hope you have enjoyed this life because it took me one hour of my life I want to get back and once we have our personas but we do next as UX designers as we saw before we create scenarios for them with storyboards customer experience map but let’s see three dimensions that need to be considered with poise when we create these artifacts regarding with sonars for example the environment is it too noisy if it is don’t use voice because that user won’t be able to hear the system and the system will have problems recognizing the voice as well if the user is in the company of other people voice might not be the best way to interact other words my over here the system or the user or might be bothered by the noise or something like that or what about the content what is being communicated if for example the information that needs to be gated by the systems the user is sensitive like the date of birth test results in the Dutch then unless no one is around and can overhear the interaction voice probably will not be the appropriate means and how much is being communicated we can’t imagine some terms and conditions being just like spoken to the user and what about the persistence do we want to maintain traces of their communication once their communication has been carried out I think it would be better for me to text you my address rather than call you

speak it to you so you can use that in the future what about the user estate are the properties of the user for instance if the user is 90 years old and hard-of-hearing or trunk or already speaking with someone or does not speak the language well or is in a hurry and can’t spare the time to engage in known conversations okay we’ve seen some things to take into account about your scenario what about the channel itself we have to remember that people can speak a lot faster than they can time so all the content we can get using voice go for it don’t make the user type on a flat screen think that they can listen much more slowly than they can read so if we have explained some kind of game terms and conditions it doesn’t look a good idea using voice for that and they can talk much more quickly they can listen for example a girlfriend so the thing is that we might think that designing our polish user interface consists in taking a graphical user interface and simplifying it for use through voice after all okay while only a very small minority of people can claim some talent as graphical artists the vast majority of us can safely claim to be competent talkers or at least competent enough to design a simple attraction between a human being on a damn computer so designing our boys user interface my seem at a gut level to be easier than designing on graphical user interface but the opposite is in fact the case voice user interface design is a lot harder than graphical user interface design okay anyway let’s move on about the tag time line narrative when you are reading text on a web page for instance you can easily escape a header and with your eyes to the section that you’re interested in but so with a voice interface where you must passionately listen to one word before you can hear the wanted follows and so on so on so on so avoid long prompt welcome to your new voice interface blah blah blah we have to avoid that use sort meanings around two three options it’s fine more than that might be problematic like select manage my booking online check-in table put important information first we already mean this but it’s not but ideas about checking everything just in case allow interruptions if we are dictating an email and we leave the app and we go back the following day the platform should say something like yesterday you were creating an email did you want me to delete it send it continually stating something Rita offer shortcuts if the user knows exactly where to go instead of saying hey welcome to very thin outline help online checking and it user goes straight to online checking on a low process let’s imagine that I ask Alexa for mattress it doesn’t have a screen so it might say something like I take an on paper and say continue we’re ready so I have time to TM paper my phone and say ok continue another one unity directionality just a stylish one-way street is a one-way medium when you hear something you countlessly go back and listen again contrast is to read my piece of text where you can Ridley’s cut a couple of paragraphs or even pages back and we read everything and so so off to repeat if the user hasn’t understood something hey could you say that again Oh what was that again overheard for example if we can’t get an instruction or a comment from the user we may say something like did you want someone from our sales department to call you on office summaries any point like hey hey hey wait a moment what have I ordered so far you have order like two pizzas and four peers and okay continue um perhaps the most frustrating thing about using a voice interface the feeling of not knowing where precisely during the interaction

what exactly the system expects to to do next invisibility on many websites we have normally on the top right corner something like step one step two payment put in confirmation of double-wrapped we don’t have that with police interface where the user can quickly feel lost for lack of mental markers positioning them when they precisely are in the exchange with the system so mark the exchange by online checking please tell me the name of the first passenger with that online check-in we solve this problem another one trace the path for example London restaurants what kind of food would you like today Chinese Spanish Mexican with that London restaurants we have solved the problem again and use earphones and ear common or auditory icon is feedback about an operation that just happened for example when we send an email on Outlook this or when something went wrong like uh or or something like that whatever but silence the silence might be understood if something went really really wrong or the app has paused or something like that so to summarise time line arity this is a whole long prompt useful meaningless put important information first bla bla you need directionality offer to bid of help of summaries and invisibility mark the exchange trace the path use arrogance um perhaps here resides the one fundamental advantage that graphical user interfaces have over voice user interfaces the feeling that a graphical user has of control over both the medium and the interaction we don’t have that control feeling when using a voice user interface so take this into account when creating those fantastic customer experience Maps the user is gonna feel amazing yields incredible lack of control over the interface okay and just one last thing this one related with the user itself when using this channel it is called anthropomorphism entrepreneurism encounter with this human beings are genetically wired to equate the sound of human speed with the intelligence and sensitivity of a human being Clifford Nass wrote the consul’s knowledge that the speeds can have non-human origin is not enough for the brain to overcome the historically appropriate activation of social relationships by voice the impulse to react to a computer that is engaging you through a speech the same way one would react to a human counterpart is much too strong when almost stable to suppress so what does this mean sir the expectations how okay if conversational interfaces are things that we talk to instead of we use what exactly are we talking to this thoughtful is fundamental to successful UX as the notion of a desktop is to the graphical user interface it’s a real our interface going to talk as if it was a semi gold in a box able to do almost anything like how 9,000 or is it going to be something like r2d2 like an extra smart Roomba this is important it’s an important question to answer because as human beings we are going to project our feelings to that interface and that’s what UX is about it’s not about drop-down menus and checkboxes as we saw in the first part of this presentation UX is about emotions reactions feelings and about how we manage them so I’m gonna leave you with that question today thank you very very much for watching if you enjoyed this and these books like the ones I commented on the first slide like best interface is not interface don’t make me tab I love that one the conversational interface at least a part blog a chatbots magazine is mason’s well and many many other good stuff out there so thank you again and see you sir