Categories
Voice

How creating chatbots can be a slam dunk

View the original post

Listen on Spotify: https://open.spotify.com/show/4n6icHZbDEcXRX4DafjtPw

Listen on Apple: https://podcasts.apple.com/us/podcast/writers-in-tech/id1467160757

Episode Summary

As a long-time NBA analyst, Eran Soroka knows a winner. Today, he champions the use of chatbots to help organizations drive conversations with their users.

He talks about the endless possibilities not only for commercial use, but also for the greater good: like helping Pakistani girls learn about feminine hygiene, teaching civics, and how his 9-year-old set up a chatbot to study English.

Eran also discusses the opportunities that conversational AI offers UX writers and conversation designers.

Episode Notes

Take our free UX Writing course: https://course.uxwritinghub.com/free_course

Check out the UX Writing Hub: https://uxwritinghub.com

Follow Eran Soroka: https://www.linkedin.com/in/eransoroka/

Try CoCoHub: https://cocohub.ai/

Transcipt:

Yuval Keshtcher 0:21
Hey, how are you?

Glenn the bot 0:23
Hey you vow, it’s Glen. I don’t know where Aaron is. He said that he’s in a secret mission in Dubai or something. Do you want to talk with me instead?

Yuval Keshtcher 0:33
Wait, wait, wait, but we supposed to had a call with Aaron from cocoa? How can I know? Who are you?

Glenn the bot 0:42
Well,

I work in cocoa hub for about three months. My mission is to manage all our HR activity, including ordering healthy snacks and organizing happy hours.

Yuval Keshtcher 0:53
What are you kidding me? Ted? What’s our drinking also in happy hours?

Glenn the bot 0:57
I can’t drink beer. But I definitely drank from the Kool Aid. And now I’m a big believer in the power of conversational AI. Wait, I hear somebody knocks on the door. Can I go for a second? Sure you do you know? Okay, turns out Taron is back. After all, he’ll take over the conversation from now I’ll go back to organizing their happy hour. By the way. I’m a big fan of your part. Have a pleasant talk.

Yuval Keshtcher 1:26
Oh, Glenn, you’re too good to me. Alright. So today, Aaron Sorkin is he Iran is actually a famous sports journalist from Israel used to work in the news. and Iran the fascinating transition from journalism to leading the marketing department of cocoa hub, which is a tool for purchasing tech tool for people that want to create chatbots. And the reason that I really like this tool is also because it’s for no coders. I love the fact that you don’t have to know a single to write a single line of code for it. And we’re excited to have him here. That’s so cool. How are you?

Eran Soroka 2:02
I’m great. I’m also like Glenn, I’m a big fan of your podcast, and I’m hearing the last things that are amazing. And I’ve learned a lot from this, too.

Yuval Keshtcher 2:11
Thank you so much. Thank you so much. Happy to have you here. And I know we’d love to learn about your background. So I know about your journalism background. But can you tell me more about what’s your story?

Eran Soroka 2:22
Okay, so I was hooked on journalism. From a young age, I started writing in a marriage with one of the biggest newspaper in Israel in ages 15 and a half, like, collecting results for the Youth League. And then I just progressed and diverged in the IDF radio, Okay, back to Mali was the I became the chief editor of sports section of Marines, which was a dream for me because I lived this paper as a kid. And then I was transitioned to hard news to being the head of digital desk, and the radio manager for channel 10 News. And then I became the chief editor of the channel, 10 website. And then like the head of TV documentary, and then the head of documentary series in on the main news broadcast of channel 10. And then I just decided that TV is great, but I am a digital person, I need to go digital stuff. So starting from last May, I’m working at cocoa hub. I didn’t know this company before, to be honest. But I was participating in a conference in Portugal in September 2017, which was called di G. And it will publishers met startups and met like moderators. And like, I learned about all these things about from the companies that do like automated videos to the companies that are monetizing content, then opened my eyes. And actually, my third three years, I worked at a chief editor at Jonathan’s website, I already became a high tech employee without knowing that because I started working with the product manager and the brand dead designer, and like programmers and developers and doing q&a and just creating, you know, designing and everything. So I did like camouflaging as a journalist simultaneously, and then I when I landed cocoa hub, I just love it. I just found out that in so much things in common between conversation design, and content creation. I’m an NBA analyst in the sports channel for the last nine years.

Yuval Keshtcher 4:16
Many of my friends are familiar with your voice to be honest, the ones that stay up all night to watch the NBA. So they’re familiar with your voice. I want to know what is conversation design

Eran Soroka 4:26
conversation design is the ability to talk to the computer in our language, not in coding but in our language for example, you do you say to the computer, I guest on your podcast, and the computer is understanding that I is me and I’m the subject and I’m being guests on is like you know the verb that is done, like the stuff of NLP and audio energy natural language processing understanding generation is like the base of conversation design. But over the last few years conversation design is really became a huge thing for people without any coding experience like me, because no code platforms enabled us to write chatbots. And that opened the field for loads of people because nobody was born conversation designers. But we were all born, having conversations from the day that we born from the day that our mom or dad are pinching us in the teeth. This is the first time this session we have at the age of like, one hour, one day or one minute. So we all know how to create conversations. And therefore, we are conversation designers. Right?

Yuval Keshtcher 5:35
So what are the best practices tips right now just for for people that are conversation designers, but don’t know the how to express their conversation design skills?

Eran Soroka 5:48
I think that the answer was actually given by one of your latter, pos guests, Neil Smith, he said that today, you have to talk to your customer. So I should, you should go and check. Maybe he’s a cost a conversation designer in disguise. Because today as a UX writer, and that’s why we think that UX writers, UX designers are actually conversation designers, we just don’t know it yet. You are trying to engage in the best way. With your customer, you decide what is your goal, your goal can be like a sales goal, it can be like Follow me on tik tok, or download my one pager or just talk with somebody just entertain somebody who is quarantined, or is one has no job, or wants any kind of human like experience. And you just think how you as a person would want the other side to talk with you. And it affects everything it detects the persona of the chatbot that you’re building. And in fact, like, how long are the prompts of the chatbot? Because if the checklist is like going now with a rant for like one minute, you will stop listening. So the prompts need to be like, relatively short and sharp, and to take you in a determined direction. They shouldn’t be like too wide. Because if you asking some person like what do you want to do? Any chat about it? No, you can answer I want to buy your product. And you can answer I want to like go and do some jogging, I want to listen to the new song. Imagine Dragons on Spotify, I want to like go and cook. So you shouldn’t be like too wide. To be like very focused on the goal. You can like, go to the goal in multiple directions in multiple ways in multiple phases in multiple personalities, but you should like decide on the goal from the beginning, and try to create the customer funnel all the thing that we know from the marketing, just by compensation, just imagine that you have the best salesperson in front of you in the store, and that he is successful in selling your product. How did he did it? And then you just create in your mind a conversation, that ideal conversation between you and the chatbot. And when you are trying like

Yuval Keshtcher 8:02
to be the buyer. Amazing. And what are the some kind of fails that you have when you try to create these kind of directions? common challenges,

Eran Soroka 8:11
there are a lot of challenges. I think that one of the things that created like Glenn is our chatbot said that he likes the the artist Bon Jovi that has a song you give chatbots a bad name. Because chatbots today are first of all, there’s confusion, what is bots and what is chatbots for the people bots, I like the automated responses that are like commenting on political campaigns. And like, Jason, you were trying to track you down and trying to fail you. And just this is the chatbot This is like automated bots, they’re good for a lot of things, but not for chetek. a chatbot is a bot that helps you that drives your drives the conversation through chatting with you. And that’s the main challenge to explain people that every goal that can be reached in a conversation, or almost every goal, maybe you can’t like take a person out of the prison if he was like Miss convicted. But almost every conversation that you can have, between humans and human can be implemented as a chatbot. And that’s a big challenge because a lot of people don’t know, how easy is it to write a chatbot out to people doesn’t know that they need a chatbot like, for example, I hear about companies that are still like in 2021 do you depend on call centers, and you can like wait five minutes in line and there was like a survey a big research by IBM, that a good chatbot can reduce 80% of the loads on the call center or the person that the sales team and I was like, oh really 80% and then I had like two examples of one place that chatbots reduced the time on line from five minutes to one and another place chatbot reduce the amount of personnel that was supposed to be employed to answer the phone from five people to one. So those votes are 80%. So this seems to be some merit behind this number.

Yuval Keshtcher 10:11
It’s amazing. And I definitely know a lot of companies in 2021, that definitely needs to, you know, take off their call center to implement this kind of chatbots. Because I don’t know, like banks, government stuff, I don’t know all of them should use that. I assume

Eran Soroka 10:30
there is actually a company, a bank in Spain, I won’t mention his name. But he’s a very innovative bank. And they’re created the first digital bank in Spain. And they just won like two months ago, with the digital banking awards, they won first place for the best mobile app and most innovative banks. And they actually had a voice assistant, installed in their bank, and it drove a huge success burden a lot of customers, especially during the Coronavirus, time when people were locked in. Yeah, it’s not like a total replacement for banks. Because feel if there is somebody that you have, like a personal problem, you need some human touch, you can do it with the person, but the other 80% of how do I find this one? How do I what’s my balance now you can do is to check how you just connect it to your database or to your CRM, or just you know, I won’t get too technical here. But mainly, for example, let’s say that you entering a chatbot of a bank, and you ask him, what’s my balance, and he’s like security and everything. And you say my, my account number is 11111119. And my name is this and that. And it sends an API call to the database, and it grabs the figure and says you have $3 million plus on your balance. So that’s the way to do it. And this is like the simplest of flows. But it can do multiple things. It can like educate you. And if you want to know what is interest, if you want to, like get some advice on where to invest your money. If you want to know about mortgages, like a good chatbot can solve most of the things that people have to explain, although it’s written already on your website, or it’s already a knowledge in the system, as good as you find it, it will serve you.

Yuval Keshtcher 12:26
That’s amazing. I played around with cocoa hub a little bit. So you can actually like in the visual editor is pretty slick. So you can just connect different conversations and like logics to each other. And then you just plug and play to your database. And it works. That’s what you’re saying.

Eran Soroka 12:46
Yeah, we don’t have like the full integrations on site. Because every we can’t like, connect it to every company’s database in the world and also be like a breach of privacy. But you can like, first of all, you build your chatbot. And then you have your technical people that have nothing to do with actually creating the chart, but just connected, you just write a simple, it’s called an extension of the code. And it’s like connecting you to the database. And we are actually some negotiations with a bigger company of the ecommerce, that we’re trying to create a use case for them. And they just need to connect us to their API. So we are like, when you create the chatbot, the the part of actually connected to your API to your database is it is performed by the technical people. Because, you know, this is like when things like cyber security and privacy, enter the picture. And you don’t want this to be done by people who don’t understand this

Yuval Keshtcher 13:48
right to any technical people for the complicated technical stuff. But in order to design the conversation, you can do it with a product like yours that you can also play around with because I remember, maybe one or two years ago already. Time flies so fast. So I played around with the Google Translate. I’m not sure what’s the name, but it’s like a database of Google that you can play around with it and create a chatbot but you had to I it was technically it was complicated, had to like take 24 hours of my life and actually learn how to do that. So you say that this is not the case with Coco home.

Eran Soroka 14:31
No, we actually make it easy. It’s like I would say that it’s so easy that the nine years old can do it. And actually have the living example because my son is nine years old. Actually he has a test ethic and examined English and he created a chat bot. He has liked like to remember to like go over like 100 boards and he created the chat bot for not on the 100 because they had like only an hour to to win it. So like for 20, or 25 words, the chatbot asked him, How do you say a certain word Hebrew in English? And then if you say the wrong answer, it says, you know, try again. And if you say the right answer, it says, great. And he wrote it. Without that I just, I had to, I taught him once, like, Andy just created a nice set for two hours and created a chatbot, that helps him to prepare for English exam. And this is actually for me, this is the best example of the fact that this interface is simple, it’s not perfect, for sure, nothing is perfect, and we’re still improving it every day. But we’re still having this thing. And also we are, since we are Coco. Coco has conversational components. So even if you didn’t, don’t know how to write a check, but you have these tacos, which are mini chatbots, doesn’t have any chatbots that are achieving a certain goal. For example, you can have a chatbot asking you What’s your name, and you say, My name is jack Smith, and it says first name, jack, last name Smith and say yes, it’s nice to meet to jack Smith. And it also like it’s good. If you’re not cooperating, it knows how to try to convince you. So you can literally if you like have a flower shop in, I don’t know what Seattle, you can take a name component blend a phone number component, that email component, like connect 123. And you have a chatbot in literally five minutes, and you can deploy it on your website. And so I’m sorry to say that, I just wanted to emphasize how easy it is

Yuval Keshtcher 16:34
to be honest, in our first call, you actually showed it how to do it live. So I can say that it did worked. And that was pretty cool. And you can connect it to your database, and then you’ll have that information in the database. So that’s an A great way to make money out of it, right? It’s like a form, like an interactive form. People love this kind of stuff. And there’s a great opportunity. It’s a great opportunity. So what kind of opportunities are there today for writers in tech that want to get into the field? Maybe they are freelancers, maybe they want to work here in a company, maybe they are in marketing, and they want to create a chatbot for their clients or companies. So what’s the opportunity here?

Eran Soroka 17:17
Yeah, I call it the holy trifecta. For me, the three types of people that are the most closest, like the most family related, under one three of the GD tell professions. So UX writers know digital interfaces, content creators know to write context, content, and marketers know to market. So a chatbot, for most of its users is at your inter interactive interface, digital interface that tried to sell with content. So actually, people can get to these from different directions. And every time I interview people, I understand that their background is different. Some of them were nutritionists before, some of them were basketballs, some of them were comedians. And like, you can take everything, as I said, Because writing the conversation is so easy. You can actually bring your own best practices from whatever you do. We have great conversation designers in the world today that we’re screenwriters that we’re filmmakers, educators, lectures in the university, you it’s actually one of the best platforms for voice chatbots today was created by a former UX designer, you can really come into this from office, and then how you can work as a compensation designer, this community of conversation designers is relatively small, I believe that there are no more than 2000s around the world that actually do is do this. And not all of them are doing it like the best way to be honest. But once you know how to create once you know how to market once, you know, digital interface, you can really get into this and you can like start and go to a company that doesn’t have a chatbot and say, Okay, let’s take a use case, for example, you want to sell flowers. And I will write to you a chatbot that helps you sell flowers. Try to I can create it like in a couple of hours, we can test it, we can try it, and then try to put it on your website and see how it goes. And for the most times, I think that the research and all the data shows that a chatbot can improve your sales of the digital interface by dozens of percentage. And once you go this is so easy. And it’s like why didn’t see your hours of availability, you understand that? You need a chapter. And you know, there’s a lot of document conversion. We convention that that convert the user, we believe in converting people like into the religion of conversational AI. Once you see how easy it is how valuable it is, you probably want to go back. So writers in tech that wants to go into conversational AI. Just think of a use case. Just find your platform. Just try to create the best flow for yourself, and then go to this company or shop or store this business and just show the people there, how it can be improved, because for example, if they’re only open from nine to five, and person wants to schedule a meeting or to buy something online, and you can tell that you can have nobody picking up the phone, a chatbot will be there for you.

Yuval Keshtcher 20:28
Amazing. The first time I tried to do a chatbot myself, it was an Arnold Schwarzenegger quote, generator. All kinds of stuff, like asking them for a quote, and then like getting close, like get to the chopper and then read like a gift of honors my senior year. So that was fun. You can do it for fun, too. But totally, I started to have some case studies of using your product specifically that were all over the international news stuff related to why don’t you tell me yourself?

Eran Soroka 21:03
Okay, so we’re very user oriented, and use case oriented kind of company, we don’t limit ourselves to one industry, or one particular region, one publishing channel. So actually, like in December, we launched a chatbot zoom channel. And then like we did a workshop, like dozens of people registered, like more than dozens of people showed up. And like about 100 people registered. And I just like trying to reach out to people afterwards, I just saw young intrapreneur women in Pakistan called Somali. And it turns out that ci runs for the last couple of years, Raji, which is a chatbot, that helps young women and girls in Pakistan learn about things like sexual education, menstruation, feminine hygiene, which I think that are rarely talked about even by men, but human. So when you’re chatting with a chatbot, the chatbot, one judge you, it won’t laugh at you, it won’t be angry at you for asking that. And the information will be there, the access will be there. And she actually because it’s Pakistan, and sometimes we don’t understand how much of a third word it is, until Subbu intends to cancel a meeting with you because there is not enough bandwidth for a zoom meeting. So they were going to like villages, they were going to community center, they put up a projector, they put up like a cartoon. And then they show the cartoon, they talk with the girls. And after like half an hour or two, they just come home and the girls, most of the times are staying with the with the last questions. And then something happened that they don’t know what to do. So the chatbot is actually a way for them. To go this and we partner with her, we let her use of course, our zoo, take Roger line, because everybody now is learning through zoom. So when she can spread the word frog via an actual avatar, and we will create a personalized avatar for her. Because we believe that this is like one of most amazing use case that conversational AI can help with. So we love it. And I think that this this project is so important that we’re only all in on helping her with this other project we are very proud of that we’re built on our platform was good. We were live on last October, there was a big AHRQ competition United States collect for medical ideas and devices concept. And some guy called granny Rockwell created a chatbot voice skill actually with with us with that AR is helping elderly people who get health tips and like make their medication regimen more clear and just send them reminders where to take the medicines. So he won the second place. And he was like for a lot of big universities and HMOs and United States, which was a big proof for us that this technology can actually help people. Right. So this is like one in the United States. One in Pakistan, we have some people in Italy and in other places in the world. Actually one great use case is that one former employee of mine via channel 10, is now a civics teacher. And he created a chatbot that helps the students in Israel prepare for the final exam in civics. And it’s one of the most successful chatbots built on our platform. And we’re very proud of him.

Yuval Keshtcher 24:13
Just waiting. What is civics?

Eran Soroka 24:16
It’s like a you know how the country works? Like kind of political, some kind of political science. Yeah,

Yuval Keshtcher 24:22
yeah, political science guy, the guy named this amazing, like a sandbox that people could just get in and create their own products. And they don’t need to code to do that. And any tool that allow people to do no code, stuff like that, I feel like create a lot of opportunities in the world. And I think that it’s amazing how different people taking into different directions and the only limit that we have here is our own imagination and they feel like we can do in the future some pretty awesome stuff.

Eran Soroka 24:51
We do rarely do. We actually we have things like touch points touch bases with people in the edtech in the final says in healthcare, in e commerce, in marketing, in HR, even in sports, like, think about creating, for example, you having a football team that you love, think about like having a chat, not a play on LBC that can chat with any Barcelona fan 24 seven, like it’s a has like one database of 100 questions that like defense can ask him from, what do you think from Cristiano Ronaldo, and who’s your favorite teammate? And what’s the biggest goal, and you just put all these questions into the Leo Messi. chatbot. And you put it on your website. And let’s say that even you can charge people if you want to, like monetize it. But this is like can be an engagement. This is engaging. We as humans, we seek engagement, we want to create great conversation, for example, this component idea, whatever you want to sell, where you in the end, you will need that person’s details, some kinds of details, he won’t like run after him, the students, hey, what’s your name, I want them to sell you a bottle. And know this, like, for example, to get the people’s address. And this is good. If you try to sell a bottle of wine or a house, you will need people’s details. So this is like a higher starting point from every other platform for building tablets.

Yuval Keshtcher 26:15
Interesting. And I feel like audio is also something that is really making waves these days, you know, all of the audio based platform I talked about of course club, so clubhouse in the past spaces by Twitter seems like when it’s like voice, interfaces, music, it’s another level of social media. It’s much more personal. It’s much more natural. And yeah, I wonder, I wonder how the future is going to look like amazing. All right,

Eran Soroka 26:43
going to sound like not only look like

Yuval Keshtcher 26:45
you’re going to sound like that’s awesome, I hope it’s going to be positive. I hope it will create positive engagement between human and technology. Because you know, there are also a lot of questions about ethics here and how we can create positive engagements and negative engagement and the case studies that you’ve told me about, like entertaining, elderly is or educating for door countries or countries about sexual sexuality. That’s a great case studies. And I really hope people will use it for the best I

Eran Soroka 27:20
can only preach conversationally, I forgot. One of our core fundamentals is an ethical use of conversational AI. And not like use it for bad purposes. Can you like to try it like people and create a chatbot that will curse your most hated person in the world all day you can do. But as I say, like, you know, nuclear energy can save the world and can demolish Hiroshima and Nagasaki. So the technology is out there. And we can also we can, the best thing you can do is like educate people how to use it for good purposes, and give them the good the socially good use cases. And we don’t have control of on what people will do with chatbots. In the future. We already know that. Like I won’t mention names of politicians in certain Middle Eastern countries that use the chatbot to like spread hate, but Okay, that’s part of the deal. You will do it with even if you don’t do to chatbot you will do it with text messages and text messages you will send like, I don’t know, pigeons that will spread to say, but so technology is there for the taking. And whatever people will do it, it’s their business. But we are really trying to take our powers and they wrecked to do directed to do good things to help the society to help the underrepresented communities to help their underrepresented people like the end users that the Forgotten users, the girls in Pakistan, the girls that doesn’t have a computer, Pakistan, but then she goes to a community center and log into the internet and ask Raji, What should she do when she had a period? You know, when you’re young, you’re thinking what yourself is trying to save the world. And you can save the world. But if you can help a girl Pakistan, avoid infection and live longer, and become independent and become more educated for me, for us as a company, this is the best thing you can do with that technology.

Yuval Keshtcher 29:16
Amazing. I also don’t give too much credit to humanity. I guess people would just take it to the wrong way without asking. But I hope regulation would take parts here and we’ll make sure that technology is not going to ruin other people’s life. But that’s a discussion for another day. All right here. We are about to finish the episode. It was amazing to have you today. And as always, we need to find a name for this episode. Of course, it could be something like avoiding the hate pigeons.

Glenn the bot 29:49
Actually,

Eran Soroka 29:50
actually, when I because I did some research about my interviewees interviewers, I heard that you have this company this thing so it actually reminded me of a story about the time Was 17 years old kid trying to get accepted to the IDF radio, which is the best place for journalists in these early army. So I was sitting with some of the big shots of the Israeli media back then trying to test me and tackle me and failed me. And in the end, one of them asked me, What headline would you give this audition, and they answered swimming with the sharks. And they all laughed, and at the end I got in. So for this one, this episode, they talked about like two options from a world of content. Since I was working in the news desk, we can go by breaking news, creating a chatbot was never so easy. Or by my love for the NBA, we can go with how creating chatbots can be a slam dunk, or how to become a chatbot champion. So all of those options are on the table, and you can also

Yuval Keshtcher 30:44
create your own. That’s amazing. You have so many opportunities here. I guess we’re not going to go for the hate pigeons.

Eran Soroka 30:53
We hate hate pigeons hated, doing door delicate and they like, you know, Google News and words that are pG 13.

Yuval Keshtcher 31:03
I like the breaking news one and slam dunk one. And we’ll figure we’ll figure it out. Thank you. And I’m actually doing a workshop together about creating conversational interfaces going to be a on the 29th of April, I believe. So we’re going to meet by then. And all of our members are going to be there and people could also joined. So we’re going to publish it in our newsletter, by the way, this podcast brought to you by the UX writing hub. So check out our website, our newsletter, we have also a free course. Check them out and sign up and then you will get notified before this event. And in case people want to reach out and talk to you and want some kind of more information about chatbots. Where would be the best way to do that?

Eran Soroka 31:45
I’m on LinkedIn. I’m very active on LinkedIn. Actually, I’m just building a network of marketers and people and content creators and people. So I would like dinner on savoca and Twitter. So Rockman So okay, ma n. And also my email is pretty easy. Aaron, era n at a dash AI calm. That’s pretty easy.

Yuval Keshtcher 32:06
Amazing. All right. So I’ll add everything to the show notes so people can reach out to you. And, you know, thank you so much. It was a pleasure to have you and see you in our next event.

Eran Soroka 32:17
It was very fun. Thank you very much for having me.

Yuval Keshtcher 32:20
Ciao.

The post How creating chatbots can be a slam dunk appeared first on UX WRITING HUB.