Building Your Business with Bots to Scale Marketing & Customer Service w/ Marlene Jia @TOPBOTS (Episode 39)#DataTalk

Listen to the podcast:

Every week, we talk about important data and analytics topics with data science leaders from around the world on Facebook Live.  You can subscribe to the DataTalk podcast on iTunes, Google PlayStitcherSoundCloud and Spotify.

This data science video and podcast series is part of Experian’s effort to help people understand how data-powered decisions can help organizations develop innovative solutions and drive more business.

In a recent Experian #DataTalk, we had a chance to talk with Marlene Jia from TOPBOTS about ways businesses can use bots to scale customer service and marketing.

Here is a full transcript of the interview:

Mike Delgado: Welcome to Experian’s weekly #DataTalk, a show featuring some of the smartest people working in data science today. We’re super excited to feature Marlene Jia (MJ), who is the head of revenue at TOPBOTS and has worked with clients such as L’Oreal, WPP and PayPal on their applied artificial intelligence and bot technologies. She is also the coauthor of a brand new book coming out called “The Business of Bots: How to Grow Your Company Through Conversation.” MJ, thank you so much for being our guest on today’s #DataTalk.

Marlene Jia: Thank you for having me.

Mike Delgado: Can you share a little bit about your background and the work that you’re focused on right now?

Marlene Jia: Well, my background is actually pretty roundabout. I mean I actually studied violin performance. So I went to conservatory. I realized that I wasn’t going to be a good enough artist, which is potentially what we’ll all be doing in the future 😉 …But, I realized I wasn’t going to be a good enough artist and decided that I should probably pursue a different career.

Luckily, I was decent at math, so finance ended up being my path post-college, and I actually met my co-founder in New York when I worked in finance, so I feel very lucky that I’m working with some of my closest friends as well as my most professional colleagues. So, previously, we were actually a firm focused on UX design and mobile development, mobile app development. And what we started hearing … like, we live in Silicon Valley, so we hear a lot about new waves of technology and what was interesting is that our clients starting asking us about it too. We were thinking, “You know what? There are enough clients asking us about these topics. We should really do a true deep dive.”

So, I’d say about almost two years ago, we started really looking into bots and looking into artificial intelligence and really refocusing the work that we do. And so, that’s kind of how we really refocused our attention onto bots. Bots are interesting, right, because bots can be seen as another distribution channel.

It’s a distribution channel for a lot of things that are related to conversation in some way. So, that’s very interesting to consumer brands. That’s really interesting to organizations because no matter what, you need to communicate with somebody on the other side and bots are just a way to kind of move that along, make it a little bit better, potentially automate it and actually bots … I mean, we as humans are good at one thing, we’re still good at creating this emotional connection, right? But bots are great and technology is great at calculation.

So, yeah, so that’s kind of a little bit about how we got into this space, and it’s just been so interesting. Very, very interesting. I mean, a few years ago, it was just talk and now I think brands are truly experimenting. They’re clearly launching their bots with the help of some platform pushes like Facebook, and it’s just been a really interesting time to learn.

Mike Delgado: Yeah, I mean it’s been exciting as I’ve been watching the trends. We’re seeing such a lift in the amount of messaging apps that are being developed in relationship to number of websites launched. We’re also seeing a lot more messenger apps, and then along with those apps, bots, right?

Marlene Jia: It’s definitely true and I think it’s even more true abroad. For example, you can see WeChat just crushing it over there in China, and I don’t want to stray on the topics here, but it’s just so fascinating seeing examples all over the world.

Mike Delgado: Yeah, WeChat is doing awesome — and WhatsApp of course. I’ve recently discovered the Line app, which has been absolutely amazing. I love the translation abilities to be able to communicate with somebody in a different language in real-time. I was recently chatting with somebody in Arabic just through a bot.

Marlene Jia:  Yeah, I would actually argue that there are more sophisticated examples of bot uses globally. I mean, just refocusing on WeChat, I was in China just a few months ago and everyone uses WeChat for everything. I think New York Times actually did a very good job of showing a video of this, so if you want to Google it you can see, but the point is is that you can pay everywhere with WeChat. You can communicate with a lot of different people. You can tip good content with WeChat, which is really fascinating because in the US, we have all this free content. Well, in China, you can tip good content, which is really interesting.

You have bot triggers, which is like, if you’re having a conversation with your aunt about … I don’t know like a cleaner service, then a bot might trigger a bunch of cleaner services local to you. The e-commerce bots out there are also extremely advanced. I think we could learn a lot from our global neighbors about how they’re using bots as well despite cultural differences and different developments. But I think it’s very interesting to look at cases abroad, because they’re doing a very good job. Arguably better than us.

Mike Delgado: What would you say for companies that are kind of getting their feet wet with bot technologies? They see the value in applying it to their business, but they’re just not sure of a way to go about it and not make the mistakes of creating a bot that seems very impersonal that’s just going to push people away from their brand.

Marlene Jia: Yeah, I think there are a few things to think about and I’ll kind of explain some of the important steps to take as well as some tactical things that I think brands could do. But a lot of times when brands build bots, I think the biggest problem is they don’t know exactly what they’re building a bot for. Bots are conversational and so people kind of mistake that as they can converse like a human, but in fact if you look at bots, most of them actually suck. There aren’t that many good bots. Let’s just be honest. But the bots that are good out there, they know exactly what they’re trying to do. So, I can give you an example:

Disney’s Zootopia is fantastic. When you think about a lot of what Disney does to engage with people, engage with their audience, there’s advertising campaigns, there’s commercials, there’s toys. Well, bot, you can benchmark against those previous engagement tactics as well, right? And so Disney’s Zootopia was this Facebook Messenger gaming bot. It’s very simple, it’s just a few different games where you work with Judy Hopps, which is the protagonist of the movie, to solve some mysteries in the city. And you know what the average duration of that bot is?

Mike Delgado: No.

Marlene Jia: It’s over 10 minutes. So when you compare the Zootopia bot against a commercial, that’s two minutes of engagement and arguably not interactive, right?

Mike Delgado: Right.

Marlene Jia: So in that sense, Disney’s Zootopia is excellent because when you compare them and benchmark them against previous ways of engaging with their customers, they just do a better job. So I think that using that as an example, a bot is a distribution channel of some form, but you need to understand why you’re building a bot and what you’re tying to do with that bot.

I think there was another example, I think Kia had a new car released, I think it was a hybrid car and they released a bot that was just an interactive FAQ in a way or an interactive guide to the car’s features. Now, you and I might be like, “Eh, that’s okay. I don’t need that bot.” But when you think about what they built the bot to replace, you can see that it’s better than marketing collateral. Like a PDF, printed collateral. Or even going onsite with a car salesman. I mean, it’s just a more effective way to show people what your car does and all the features without spending all that money on extra collateral or having an extra dealership position to have people on site.

So I think something for brands to keep in mind when they think about bots, just like any other emerging technology, usually emerging technology can replace a previous solution. Not just replace, but also augment a previous solution. Like you could still have printed collateral and still have the interactive bot, right? So, it’s just one, knowing exactly what your bot is trying to do. Two, understanding what bots can do today. Bots are not fully conversational, so you can’t expect it to do everything like a human would.

And I think once you have a better idea of those things, then some of the tactical things a brand could do is hey, go to a conference, a bot-oriented conference. You can meet a ton of players there. You can meet other brands, talk to people, get some information. You can ask for referrals about specialized bot shops or advisory firms. I mean, again, with emerging technologies, you’re going to have specialized shops know what the heck they’re doing compared to just a generalist consulting firm. And then again, that’s just my opinion, but I think if you really want to do something well, you need to understand who as the domain expertise, who is willing to experiment with you and help guide what you’re trying to accomplish, and then also, just know that you’re going to need resources.

When you’re dealing with emerging technology, you can’t just do a plug and play platform. I cannot tell you how many brands we talk to where they think they can just throw a little bit money, dip their toes, and expect to do something good with it. No, no it just doesn’t happen. So, sorry, I kind of went on a rant there, but those are kind of a few things that as a brand I would think about and then like the conference, the specialized shops, those are some tactical things that you can also do, because obviously if you’re a brand, you’re domain expertise is in the product you’re selling. You may not be a technology company yet, but I think in the future, every brand, every company is going to have to have some level of an understanding of technology to really capture revenue opportunities.

Mike Delgado:  I like what you were saying because it’s really important to have a bot that is domain-specific, right? Don’t just have this generalized bot that may be able to answer some questions, but really having domain-specific bots and then maybe having multiple domain specific bots that then can communicate with each other to help solve problems when those general questions come in.

Marlene Jia:  Yeah, that’s right. I mean, the way I like to think about it is, we can even use designing websites or mobile apps as an example. You’re not going to design a mobile app for commerce the same way you design it for a gaming app, right? You’re just going to have different functionality, the users are going to have different flows. That’s the same for bots in a way, to my point again as another distribution channel, right?

The nice thing about the distribution channel bots, though, especially chat bots, is that it’s going to be interactive. It’s much more real time, much more personalized, and I think that’s honestly what people are excited about is this new type of distribution strategy through bots. So you’re absolutely right about being specific about what you’re trying to do when you launch a bot. Just like anything else, really.

Mike Delgado: I want to ask you a little bit about how to make or your recommendations when a brand approaches you to say, “You know what, we want to create a bot that’s going to be domain-specific but also is going to be friendly, delightful, personable enough. How do we do that?” And I’m kind of curious about your approach to working with content teams and people that are working in AI and data science to collaborate on creating these delightful bots.

Marlene Jia: Sure. Yeah, I mean it’s definitely not easy, because again, this is conversational and this is kind of personalized to that individual consumer on the other side of the bot. The way that we like to think about it and again, I think us as an industry, we’re also still learning, right? This is still nascent. But what I like to think about … the way I like to think about it is, if you’re designing a conversational flow, you should first observe how your best customer agents are engaging with your customers.

Observe the conversation that goes on with your best customer agents, right, because you want your bot to be as close as possible to that style and personality while covering the topics that are important to the consumer. So, you try and find … for example a customer service agent, you find a person that represents your brand well, and you watch how they converse with the customers.

Once you have a lot of these scripts, you try to operationalize it, right? So, obviously a bot is not the best at being a human with another human, but what a bot can do is it can recognize commonly asked questions. It can recognize, hey, perhaps this question is better routed to this customer service agent so that this human doesn’t need to wait for 30 minutes on hold.

So I think with designing conversation, you need to understand the brand personality, then you can observe people who represent that brand personality the best and design conversations from there. Because I think to the point earlier, bots are still not that advanced. But what we can do is learn best from our human counterparts and then operationalize that in the bot for the best type of engagement possible. And here’s the thing, I think as people, we just want to know so if you’re on the other side of the bot, I would like to know I’m talking to a bot.

Because if you know you’re talking to a bot, you’re behavior is going to be slightly different and your expectations are a little different. If you talked to this bot and you expected that bot to be human, you’re probably going to be pissed pretty, pretty quickly, which is what a lot of brands have learned. And a very easy way to do that is maybe just have an avatar, right? Have an avatar be an icon representative and then you go from there.

Also, keep expectations clear. It’s actually funny, we have a bot on our site called Botty McBotface. Now, Botty McBotface is on our site to help you find content you’re interested in. The funny thing is, people don’t really realize that Botty McBotface is just there to help you find content. We have people come to Botty McBotface for love tips, to just talk to him. We constantly have to redesign our bot to say, “Hey, here to help you with the content.”

So again, it’s about setting the right expectations. And this is an iterative process and so it’s, in our opinion, sometimes it’s good to just release it on a small scale, get enough information and then keep iterating. When you do that, you have a better likelihood of getting your bot to a place where you do feel comfortable releasing it to a wider audience, right? So yeah, I mean look, I think there are different techniques to designing that content, but I have found that at least those steps that I had mentioned before is very helpful and a good way to think about the experience that your customer should have with your bot.

Mike Delgado: Yeah, we’ve been moving far past the old “Clippy” bot days.

Marlene Jia: Exactly.

Mike Delgado:  I love your approach at Topbots on how you’re very much focused on capturing the brand personality, reading through the scripts, understanding customer challenges, how they’re approaching customer service to help develop those scripts, and then also just being real with people who are interacting with the bot.

You’re just trying to make a helpful bot to help that person find the right content.

Marlene Jia: Right. Right. That’s right. And no matter what, you’re going to upset a few people, so I think we have to get over that. That’s just what happens when you develop … you know you’re trying to develop new solutions and you’re trying to do something better. It just takes time.

Mike Delgado: So, I’m kind of curious about when people are getting upset, are there triggers that you apply to alert someone in customer service? How does that work and how does the bot start to understand that someone’s getting upset or frustrated?

Marlene Jia:  I think right now, it is still what we call rule-based. And rule-based means that we as humans, we as the designers of the bots need to be aware of those triggers. And keep in mind that again, to our point earlier about being domain-specific or use case specific, every bot is different.

Confusion is going to be different depending on how you’re using or designing the bot. So a shopping confusion trigger is going to be very different than say like a gaming confusion trigger. Though most of the time, you’ll see certain questions, right? Or people will repeat themselves.

This is the clearest indicator of a badly designed bot … I’m not going to call out any names, but I can name a ton of brand bots, where asked a ton of these bots, because we’re trying to learn. Some of these bots, I’m just … you know, I type “I’m confused. I want to go back. Can I go back? Can I go back?” And for some reason the bot just is not designed to recognize when you repeat a question twice, perhaps you are confused as a consumer, right?

And the important thing about confusion triggers is that, look, because bots are very limiting, route it to a human. Route it to a true customer service bot. The point is to get data. Learn more about your interactions, so you can build in more of these triggers as you go. It’s just not going to be perfect. And that’s why I say a lot of these bots are iterating. We’re iterating on creating better and better bots.

That’s why setting expectations up front is very important. I actually think consumers are much more forgiving when you set those expectations up front and say, “Hey. We’re trying to create this solution that will better your experience in the very near future. Hold tight with us though.”

So, it just … these confusion triggers are pretty difficult to just put into one category, because again, a confusion trigger for different bots are just different so you have to build them differently. But that’s where also artificial intelligence is very helpful. Bots and AI are very different, but we can certainly use both to help each other. Bots as a channel for delivery and AI as a way of making bots much more conversational. So, we’re still iterating. But I think there are analytic companies out there like who’s trying to build in triggers into their analytics alerts so people can track them. But a lot of it needs to be defined first.

Mike Delgado: Right. It always goes back to what is your ultimate goal with that bot? And that’s then going to help define what are the metrics that are going to determine success. Now, we have talked a little bit about some use cases, definitely customer service you pointed out and sales, marketing efforts. Someone’s on a store website and you’re trying to guide them to the right product to help them. What other use cases do you find are really helpful for businesses?

Marlene Jia: Sure. Yeah, I mean, when you think about a chat bot, its conversational. And that is a good baseline where to start. And there are different ways that you can engage with consumers. Whether it’s helping them convert to a purchase, or it’s just informing them of a new product release, like we talked about marketing. Deal promos. I think actually … was it British Airways? They released BOTler. I thought that was very clever.

And BOTler will share flight deals. They’ll share flight deals, they’ll share information about the cities that you’re in. That’s just a way to engage with you. It may not be direct to the brand, but you can certainly see how that relates to the brand. So it’s just any way to engage with the consumer. I can give you some examples. Actually, from a charity perspective, there’s this bot called Yeshi. It was created by Lokai and Charity: Water. Yeshi is a bot, but she represents this young, Ethiopian girl who has to carry a few gallons of water from a water well that’s a few miles away.

And she shares her experiences with you as one who cares about the initiative. Sometimes when you donate money, she’ll actually say, “Hey, thank you so much.”

I think it’s actually a pretty well done bot because what does the bot do? It tugs on your emotional strings. It shares really interesting information about the experiences of people that you are trying to help as a person donating money into this charity. You don’t have to donate money to access the bot, but obviously if you’re accessing this bot, then you care about the initiative, right? And so Yeshi is this bot who’s just trying to share these experiences. And really I think ultimately their goal is just to make people care, right?

I thought that was an excellent way of using a bot. So bots, like I said, at a foundational level, it is meant to engage with somebody. Whether it’s a consumer or it’s just a general audience or it’s for customer service or post-purchase transaction.

There are a lot of different ways to engage with people and I think that as you see voice assistants, which are also a type of bot by the way, as you see voice assistants become more pervasive in our society, you’ll see that become a much more interesting way to deploy bots. I’m actually very excited to see how voice bots develop because frankly, it’s almost like having an assistant or having a friend or a companion. I mean in Japan, they’re already creating these companion-like assistants.

On the flip side, that also causes concern, because some interesting questions are … let’s think about this. All of our voice assistants are actually by default female voices. Interesting. What does that mean for us? In Japan, they actually designed pretty feminine … I’ll stay PG here. But they’re developing some pretty feminine bots that acknowledge you, give you compliments through the day. That’s also very interesting. One could argue that, “Who needs people, right?” When bots can fulfill all of your emotional needs or give you all the validation that you need.

What does that mean for our society? I mean, I think it’s very fascinating, especially as bots of all sorts become more conversational. I think it will be interesting to see how that affects us as people and how we interact with each other.

Mike Delgado: You shared that Charity: Water example. That’s actually the first example where I’ve heard of a bot really pulling at heartstrings. Really it’s that empathy. And to your point about voice and AI with like Alexa and voice automated bots. Those are also ways like where relationships are beginning to be formed. I just read a study here in San Diego where they brought Alexa to a senior living community and it was actually very, very helpful for people that had vision problems or people who felt lonely. Just having Alexa in the room to be able to tell them what the weather was like, order things, and take music requests was helpful. It was like having a voice companion — and it made people feel less lonely. So yeah, it’s really interesting to think about bot ethics and where all this is going.

Marlene Jia: Right. And I think it’s very important for us as technologists, in fact I think technologists are very accountable for how this develops. Like, we need to think about how we’re designing these technologies to include these relevant issues that we’re talking about right. And it’s actually kind of funny, you build in these acknowledgement bots and then it makes people happy. It kind of makes you wonder, “Wow, I could probably be doing a better job.” Being a better friend, it doesn’t seem that hard, you just need to do it a little bit more on command. We’re just sometimes limited by our own cognitive efforts, right?

Mike Delgado: That’s right. That’s right. MJ, we’re coming to a close. I wanted to ask you before we get to the final five, what excites you about the future of bots and where do you see things headed?

Marlene Jia: Oh my gosh. That’s a really tough question. I think the last thing that I said about how our society is going to develop, it both excites me and also makes me nervous. I mean, with all other technologies as well, AR, BR, artificial intelligence, it begs the question of are we going to need other people, right? And that’s both a very exciting thing, because that means that we can be much more independent, but it’s also kind of a scary thing because how are you going to maintain relationships?

Relationships are not easy whether it’s with friends, business, we don’t always get what we want and I think with bots that personalize to us, we kind of get spoiled. So I think an interesting thing for me to think about is as bots and as conversational agents develop, or even ancillary technologies develop, how are emotional and physical needs satisfied and what does that mean for us as this world develops? So, I’m very excited because I’m very lazy about doing some rote tasks of mine, but I’m also very nervous because I don’t know what that means for us as a society. So, I don’t know, we’ll see. We’ll see.

Mike Delgado: Okay, so before we get to this quick fire final five, I wanted to share something. I was watching an interview with some bot technologists over at … where were they at? Facebook and Slack and the developer over at Slack was saying he doesn’t use the Turing test to determine is this bot human? He has what is called the beer test. And is this bot friendly enough that I take him to a pub? So he calls it the beer test.

Marlene Jia: Hey, that’s not bad. You know, here’s the thing, most bots … sorry, let me take it back. A lot of bots will actually pass the Turing test. It’s not that special. Like you have Tinder bots that fool people all the time. So I think that people need to take the Turing test with an understanding that that’s actually not the complicated part. I think prolonging longer meaningful conversations is what’s really going to be a good test.

Mike Delgado: Okay, very cool. To make sure that you’re not a bot, MJ, because you answered all these questions phenomenally, we have our own little Turing test.

It’s called the final five and there are just five random questions. I wrote these out based on what I saw on your Twitter profile. So, the first one was what’s your favorite restaurant at the moment and what do you get there?

Marlene Jia: You know, I actually love cooking. If in the near future there are machines that do most of my job, I will go to culinary school. I will go to culinary school. I love cooking, so I actually really enjoy creating my own recipes. I love cooking Chinese food, I’m Chinese and I just have a strong emotional attachment to my mom who is an amazing cook. There are some excellent specialized restaurants. Like, I love … I think a lot of the restaurants I love kind of appeal to the hominess, so comfort foods. So I love ramen, dumplings, and love Din Tai Fung.

Mike Delgado: Oh, yeah.

Marlene Jia: So mine are specific, but I absolutely love, I just love cooking. I love culinary experiences, yeah.

Mike Delgado: That’s cool. MJ, can you make soup dumpling?
Marlene Jia: I cannot. It’s hard. It’s not easy. That’s why you don’t have Beijingers making good soup dumplings, it’s only the Shanghaiers that know how to make good soup dumplings.

Mike Delgado: Yeah, I sit there over at Din Tai Fung and I watch them make it, like they’re rolling the dough. They’re putting everything in and I watch it, and I’m like, “Oh, my gosh.” So good.

Marlene Jia: That’s okay. We can be experts at eating them.

Mike Delgado: Okay, so second question is what is your spirit animal?

Marlene Jia: Oh, that’s a good question. You know, if it could be imaginary, I’d say the phoenix. I like to renew myself and constantly learn. If it were a real animal, I’d say the German shepherd. Because I’m pretty loyal, but I’m fierce.

Mike Delgado: Yes. Oh, my gosh.

Marlene Jia:  Yeah, I also own a Dutch shepherd, so I’m very partial.

Mike Delgado: That’s cool. Okay, third question. This could be embarrassing… I don’t know. Favorite band in junior high or high school?

Marlene Jia: You know, I don’t think I was cool enough to know bands.

Marlene Jia: I was a … I practiced violin so much, so I was just like these people … contemporary bands, ugh. I was so elitist about music, I swear. I’m still a little elitist about music, so I very much appreciated the classical arts. I’d say if I had to choose a classical, Beethoven is … he’s really unparalleled. He’s really unparalleled. His music, when you listen to it, you’re just like, “Wow. I don’t think there could have been anything better where it is. Like, I can’t replace any of the chunks there.” So, if I had to choose then in high school, I’d say Beethoven.

Mike Delgado: Okay, okay. Now, if you had a bot selecting your music right now, maybe on Spotify or Pandora, what would be like a favorite playlist or type of music you like?

Marlene Jia: Oh, wow. It would be pretty jumbled. It would be mixed between a lot of electronic chill, jazz and classical music. Yeah, it’s gonna … very spiky. Very, very spiky. Probably be hard to concentrate to.

Mike Delgado: Your bot would be struggling, your bot would just be working overtime.

Marlene Jia: Yeah, exactly. Yeah.

Mike Delgado: Okay, last question is the dessert you can’t say no to.

Marlene Jia: I don’t like desserts.

Mike Delgado: You don’t like dessert?

Marlene Jia: I do not like desserts.

Mike Delgado: Really?

Marlene Jia: I have a savory tooth. Give me that burger. I would prefer or give me that taco. Oh my God, I love tacos. Give me ramen, oh my gosh. I am more of a savory person, no sweets [crosstalk 00:34:27].

Mike Delgado: Yeah, yeah.

Marlene Jia: On occasion, I like cheesecake and I like tres leches if it’s done well.

Mike Delgado: Yeah, yes.

Marlene Jia: But, I love pizza and burgers and steak. I love savory food.

Mike Delgado: Yeah, okay, okay. So, dessert is out. All right, well, MJ, this has been fascinating and wonderful to talk with you, learn about bot technology.

For those that are new to Data Talk, we do this every single week where we talk about bit data analytics and things going on in data science. If you want to learn more, you can always go to and before we go, MJ, where can everyone learn more about you?

Marlene Jia: Yes, we actually provide a lot of content about different topics regarding bots and AI. You can go to and if you guys have any questions you can also email me. I’m So, pretty simple, hopefully pretty simple and if you want, you can even talk to Botty McBotface.
Mike Delgado: Awesome, I put that on the screen. Marlene, thank you again so much for your time. It was great hanging out with you and looking forward to keeping in touch.

Marlene Jia: Yeah, thank you so much for having us at Topbots too.

Mike Delgado: And I definitely gotta check out Botty McBotface. That sounds so much fun.

About Marlene Jia

Marlene Jia (MJ) works with brand clients such as L’Oreal, WPP, and PayPal on their applied AI and bot technology initiatives. Her expertise in enterprise software and best practices helps corporations successfully evaluate, develop, and integrate emerging technologies.

Prior to TOPBOTS, she built go-to-market sales teams at high-growth companies like UStream, Wizeline, and Sales Bootcamp and was COO of Xanadu, a leading strategy and design firm in emerging technology.

She’s recognized by INC as a top 10 keynote speaker and recently presented at AI With The Best and Insight Exchange Network. Marlene studied economics at Northwestern University, starting her first company there. Her first venture had millions of dollars in sales and she has been a serial entrepreneur ever since then.

Make sure to follow MJ on Twitter, LinkedIn, and her work at TopBots. Also check out her upcoming book: Business of Bots: How to Grow Your Company Through Conversation

Check out our upcoming live video big data discussions.