top of page

Voice design today is like web design in 1996 - Vittorio Banfi


Thank you very much. I I'm fixated with right this period my team is running crazy and buying all the red stuff. Hello everybody. I'm Victoria and the CEO and co-founder of Bot Society. I'm very excited to be talking today. So I'll be introducing very briefly what we do and then the topic is something else. So but societies our conversation design tool is like Photoshop for vice and chat we out creative teams figure out the best interaction possible before handing that over to the engineering team that will use is on technology where technology agnostic down the road so far. Our community has collectively designed more than 2 million messages on our platform.

We also have an educational program that I enrolled more than 70 universities. We have more than 1,000 students on and we have Ivy League's were very proud of the projects they're running and the courses they're running with bot Society. I would say using about Society our investors include Google 500 startups, and those are some of our customers that’s it. Let's talk about conversation design. So conversation designed today. It is the point of all I will make conversation design today is like web design in 1996. And so to make this point make a very brief history of human computer interfaces, very brief. I promise.

So this is the common line. How many of you have ever used a common line raise your hand love it, right? You need to type exactly what you want to do. Otherwise, it won't work. So there's a lot of cognitive load that's one way of interacting with a computer another way, you know 1986. The Macintosh is introduced as the graphical user interface. You don't need to remember every command you have screens and you have buttons and you have a pointing device. Is there your mouse or your fingers in the smartphone and you can play around with graphical elements. Now we're witnessing the third type of interaction which is the conversational user interface where the main interaction is a conversation.

That's the way that we look at it. And so I don't know if you ever seen this ad of the Google assistants assistant, but what's interesting about this is that he is not looking at the screen. So that's a Google assistant device with a screen. It's called It's called The Google Hub. There's also an Alexa version which is the likes a show but the point here is this device has a screen but is not even looking at it is but it's interacting with it. And so from a designer point of view what this means is the main interaction is the conversation and the graphical party kind of a secondary interaction in the same way that saying I'm graphical user interface. You still have some command line kind of things like when you search you need to be able to remember exactly what you're looking for, but the main interaction is the Article 1 here's the same thing. The main attraction is the conversation.

All right, where's 1996 went? All right. So 1996. Its 10 years of 30 years. So we worked 10 years some graphical user interfaces and then the web came along. Oh my God, we can build websites, right and then we got this. Whoa there so that's a menu. So it is a rotating magnet is a Space Gem website in 1996 amazing. Like everybody loved it and the background was moving is a moving GIF and this is a menu.

So so what happened like, how did we end up doing something like that after 10 years of work on graphical user interfaces. This is not like it's amazing. You can still you can still take a look at it as but as Not very usable, right? And so the reason is that we didn't know the best practices of graphical user interfaces just because there weren't enough people using them and so an example is like, okay, we launched our e-commerce website in 1996 the website on the left hand side and the right hand side technically.

They work the same but the button the buy button on the first website is yellow is on the bottom left doesn’t sell as many shoes as this website with a button on the bottom right corner and it's green right, but until the web came along and you had enough people using the stuff you didn't know that. Because there was no Collective journey of users and designers. So basically you have enough users using this stuff to give enough feedback to a community of designers which learns from it and iterate very fast. So from 1996 that accelerated very fast compared to before because you had this Collective Journey. So it was like a collective sandbox like a global sandbox where everybody could try stuff and get feedback very fast.

And so conversational user interfaces are reaching this Tipping Point where you have enough users using them every day in giving you feedback so that you can iterate very fast. And so they're not they're not new as well.

So if you have ever tried Eliza, that's from 9066. So it's even before and it's amazing. Like it looks like you can have a conversation about your fears and what you're thinking is kind of like a Psychotherapy and a thing. But so this was one of the first attempts and it still works very well. But if you ever use Siri, you had the Sensation that we're still in the Space Jam website kind of situation, right? We are like, oh this works this may work, but it's still not very usable and this is because we still didn't figure out like the 100% the best practices of conversational user interfaces, and I'm just going to make an example here to define a little bit better what I'm talking about. So this is an example of a question design.

So I'm building this bot to buy plane tickets. And the question is like I can help you buy plane tickets. Where would you like to go now? This is not a good design. If you if you design the but this way if you designed the question this way, then a lot of folks will say stuff like I want to go home ouch and then the bot won't get it. So a better design would be to say, where would you like to go popular destinations from San Francisco include New York and Los Angeles. Now, this is a better design for two reasons. The first one is that you're telling me basically that I need to come up with a CD right without telling me exactly what to what I have to say, which is also not a good experience, but I understand that that's the first thing and the second thing is that You're letting me know that you know that I'll be starting from San Francisco in this case. So we're basically saying to the user look we know where you're located.

Just tell me where you want to go. And that's another step that builds trust and allows the user to come up with an answer in a more meaningful way again, same technology different design totally different results. And so we're still in the early days kind of like the Space Jam thing and the more they will work on it collectively together. The mortise best practices will emerge as was one example, but as I was saying before we're at this Tipping Point where there's enough users. On this interface as that enables us to start this Collective journey. And so there's also I mentioned this during the panel, so I want to repeat it too much but like seventy seven percent of users speak with their smart speaker daily or more. So there's a lot of usage now and there's hundreds of millions of devices out there. And so we can start this Collective Journey.

So the difference is that a conversation interface is fundamentally different compared to the graphical one. So if you ever tried to visualize a conversation, it's really hard to do it if you don't have the proper tools, so but Society what we're doing is we're democratizing this so we're allowing creative people were used to work on different interfaces. We're giving them the design tool that allows you to easily visualize complex conversation flows in a way that it's intuitive and fast and here I have some ideas about I think the stuff that the designer Community will work on in the next year or so that we're excited about one of them is personality. I think there's still a lot of stuff that can be explored there. For example, I would really like to ask the weather to a grumpy guy. Like I don't like the weather to my right the stuff like that. We're going to see a lot of innovation there I think and experimentation. Another thing that we didn't experiment and have fun is context.

So how much can we use the data already know about the user to ease the experience? And generally speaking new experiences so new services or products which are now possible given this new this new interaction. So if you ever tried games for example, which is a classic initial use case that you have an emerging platforms if you tried games that are native to voice their very interesting. And that's because we're in this Collective Global sandbox. Well, we will experiment with users. So that's it. Thank you bonus bonus one bonus question one bonus question the man in green. He has been thinking and asking questions all day here. You are sir.

You're going to have to move up to the front here pretty quick. This is your tool also work as a CMS. So you can you can turn about Society into a CMS. We have an API that allows you to extract the content when you want and so you can effectively use it as a CMS. You can also not use it as a CMS. So it really depends on what your engineering team decides to do down the road. Once you're done with the design. They can pick the tool that they prefer.

So a lot of people use rasa which is amazing Dalek flow and stuff like that and we integrate with those



bottom of page