The story behind the company that built TechCrunch’s bot

I am obsessed with chatbots so the last time I had today’s guest on I asked him about chatbots instead of about his company.

We even disagreed about the future of chatbots.

But today I want to get into how he built his platform, Chatfuel. It’s the most popular platform for creating fully-featured chatbots.

Dmitry Dumik

Dmitry Dumik


Dmitry Dumik is the founder of Chatfuel, the most popular platform for creating fully-featured chatbots.


Full Interview Transcript

Andrew: Hey, everyone. My name is Andrew Warner. I’m the founder of Mixergy. And you can tell sometimes when I get so carried away with an interview because I forget about the topic of the interview. That’s basically what happened last time. I am freaking obsessed with chat bots. So last time I had today’s guest on, instead of talking to him about how he started his company, I had to ask him what he thought about chat bots. Frankly, I think we debated a little bit. Is that right, Dmitry?

Dmitry: Yeah, just a little bit.

Andrew: Right? I think we had different points of view about it. I wonder if over the last few weeks since that last interview that you’ve changed your mind a little bit about what I’ve said.

Dmitry: I think we should check back in like a year.

Andrew: To see who’s . . .

Dmitry: Exactly.

Andrew: To see who’s right. All right, I would like that.

Dmitry: Yeah.

Andrew: Okay. Let’s do predictions for where we think chat bots are going to be. For anyone who doesn’t know what a chat bot is, a chat bot basically is you go into your chat app and you can subscribe or start a conversation with a company, right? So you might, for example, go to a Pizza Hut. I don’t know if Pizza Hut has one, but theoretically . . .-

Dmitry: Not yet.

Andrew: They do have one?

Dmitry: Not yet.

Andrew: Not yet. When they do have one, you’ll be able to go in there and say, “Hey, I want a pizza.” They say, “Where are you?” You tell them exactly where you are. Then they say, “Okay, it’s going to cost this much money,” and you tell them, “Fine.” And then you put your thumb print on the button on the iPhone and your pizza will be delivered to you and the whole thing happens within a chat experience. That, I think, is the future of chat bots.

Today, if you want to see what a chat bot looks like, you can see ours. With ours, you give me permission and every day I teach you something via Facebook Messenger. That’s the chat platform that we decided to use first. By the way, if anyone wants to check it out, we’re going to have it up on if you want to see how chat bots work and learn a little bit about them.

So anyway, last time Dmitry was here, we talked about bots and where he thought it was going and where I thought it was going, and I thought it was a really helpful conversation. What we didn’t get into is how he built this platform, Chatfuel, that is now, I think, the most popular platform for creating chat bots, and it’s all free and it’s all very easy. Chatfuel is a platform that allows you, in fact, to build your fully-featured chat bot in less than seven minutes. That’s one of the things they’re especially proud of that you can do it really fast.

So I invited Dmitry Dumik back to talk about how he founded Chatfuel and how he grew it. And this whole interview is sponsored by two great companies. The first will help you hire your next great developer. It’s called Toptal. The second will do your books right. It’s called Bench, but I’ll tell you more about those companies later. Dmitry, good to have you here.

Dmitry: Thank you so much for having me once again.

Andrew: How many bots have been created on Chatfuel?

Dmitry: That’s a great question. Now I think we are close to half a million.

Andrew: And how much revenue have you earned from all that?

Dmitry: It’s a funny question to ask a Silicon Valley startup. We’re a pre-revenue company.

Andrew: The answer is none. I smiled because I know the answer. The answer is none. The reason it’s none is because you’re focusing on growth. You want to be the biggest. You want to be everywhere. You want everyone to use you guys, and revenue will work itself out later. Am I right?

Dmitry: That is the plan.

Andrew: Okay. And it’s a perfectly valid approach, right? Facebook, actually, I don’t know that Facebook did that, but Instagram did that, right? Cool. Can you give me an example of a bot that’s been built on the Chatfuel platform, something that gives us a taste of what it’s like?

Dmitry: Sure. So one of the most well-known is the TechCrunch bot. Basically it sends you a summary based on your preferences of what’s going on in the tech world, and you can subscribe to the topics that you’d like to follow, like follow Messenger topic or anything else. Recently what I liked a lot, there was a TechCrunch Crunchies Award. So this event, they were selling tickets through this chat bot. So that was pretty cool.

Another example that did a lot of bots was a Christian Grey chat bot that was launched in support of “50 Shades of Grey” movie. And it did a lot of bots. It was featured on TV and many media outlets, so I highlighted it. The reason is that it evokes emotions. So it talks to you in a way that Christian Grey would talk to you, like, “You know what I want to do next?” “What is that?” “I want to put this on you,” and then a picture of handcuffs. The flow is just like–it makes me [inaudible 00:04:34].

Andrew: I remember that one. After our last interview, I took you out to dinner, but you’re not a big eater, so I ate for both of us. One of the things you showed me was that bot, and you said, “Here, Andrew. Take the phone. Say whatever you want to the bot and watch it respond.” I did. I tried to trick the bot, and it responded basically like a human being would. I didn’t get the handcuffs. That would have been cool if it would have said, “I want to put handcuffs on you.” But I did get the full experience from it.

All right. Let’s go back and see how you figured out that this was the business to get into. You’re a guy who before this was working at Procter & Gamble. Give me a taste of what it was like to work at Procter & Gamble as a systems manager, which is what you did.

Dmitry: Well, honestly, it was a lot of fun. I was offered a full-time job even before I graduated from university. So that was a lot of excitement for me, because, at that time, they offered me like a huge salary, like no way someone would pay salary like this to an undergraduate. So I had accepted it and I was lucky enough–I was working under great managers and I was very often new projects, like a plant assignment on a plant or production lines and on and on. My last project was about like creating an end-to-end reporting system for all P&G products, from small kiosks on streets all the way to the country level.

Andrew: Sales reporting?

Dmitry: Yeah, for products.

Andrew: I see. So anyone could go in and see how even a small kiosk performed as a sales mechanism for a Procter & Gamble product? That’s phenomenal. So then, considering how challenged you were and considering the work was, why did you go on as . . . I’m looking at your LinkedIn profile 2012, why did you cofound–is it Penxy?

Dmitry: Penxy, yes.

Andrew: Why did you start a company on your own considering you were doing so well?

Dmitry: That’s a great question. So, in November, 2011, I was approached by a friend who was working at the university, so he was like lecturing. His students were doing an assignment. As part of the assignment, they wanted to do a side project, a nonprofit project, and I was invited to join this project as a mentor. I accepted it. Two things happened during that nonprofit that we created. We created a very simple platform.

At that time, individuals in Russia, they didn’t contribute much in terms of like donations. Actually, they contributed 100 times less than an average American would contribute. We were like, “Why is that? How do we solve this challenge?” And the number three reasons, top three reasons, people said they did not have money. There was a lot of friction. You had to like pay somebody or like transfer money somewhere, and there was no transparency in this process.

We came up with the idea of an online platform, like a three-sided marketplace, where one side would be advertisers. We were lucky to have like Google, who would pay for people to download their Google Chrome, a couple of banks, Starbucks. On the other side, there would be people who would just go there and do those ad actions, like watch a video, download a browser, filling like an application. That money they earn by doing the ad sponsored actions, they could direct over to any charity.

Andrew: I see. So I go and download Google Chrome. Chrome pays you, and you give money to the charity of my choice.

Dmitry: That is correct.

Andrew: That was the model.

Dmitry: Yeah, exactly.

Andrew: Okay.

Dmitry: So we did an MVP. It two weeks to build it, and it took two weeks to earn like $5,000. Like again, pure virality, we were featured on radio. People supported that. They shared it. It was called One Minute, like you don’t have to do much to support a charity of your choice, just spend one minute of your time. So we took those $5,000 and we purchased gifts to kids in the orphanages for New Year’s holiday.

This is where two things happened to me. One is first time in my life, I was working in a team where everybody was so excited about the end goal and was so motivated that you didn’t have to spend like a minute of your time to convince somebody to do something. We were like working at night. Nobody was paid. Everybody was just so excited about the mission. I was like, “Wow, that is just phenomenal. I want to be in this environment my whole life.”

The second one is when we wrote those gifts to the kids and gave it out to them, I felt a sense of impact on a much different level than I used to have in a corporation. But my actions are changing something in this world.

Andrew: Yeah.

Dmitry: At a much bigger scale. So those two things triggered the chain of thoughts in my head like, “Okay, what is the point of leverage I can have so I can have bigger impact and deliver something that I believe in and impact more people in a way that resonates with me?” A friend of mine sensed this hesitation in me.

So he came over and I was like, “Hey, I’m doing this startup focused on education. You have this amazing experience. I’d love you to join the startup.” I was like, “Okay, let’s do it.” But then I was like pretty well paid. So it was like a big problem for me to kind of give up my salary. So I told him, “How do we deal with that?” He said, “No worries, we are about to raise $1 million, so we can keep you on a salary.” I said, “Okay.” So I left Procter & Gamble. After one month, we realized that investments are not coming.

Andrew: You left before you got the funding, before you could be sure that you were going to get that money?

Dmitry: Yes.

Andrew: I see. Wow. All right. And were you married at the time?

Dmitry: I got married soon after I left Procter & Gamble.

Andrew: So you had some responsibilities coming up in your life. It’s not like you could just stay in somebody’s kitchen and sleep on the floor.

Dmitry: Yeah.

Andrew: Okay.

Dmitry: So this is where my startup life started. And then over the next year, I think it was one of the hardest–well, maybe the hardest year of my life because it was a startup. None of us knew what we were doing, and we made all possible mistakes. We had like four cofounders, all very smart and guys with a lot of leadership, but we would end up fighting like what is the right decision to do. We would spend like hours discussing it instead of just selecting something and working on it.

We would launch too late. We would launch without actually like sensing the market. We would pick the wrong niche to focus on. We would be raising money without actually having the product. It was like super hard to raise money. We have like a big team of ten people to whom we would need to pay.

Andrew: I see.

Dmitry: Like all possible mistakes, like you name them.

Andrew: So what happened to the business?

Dmitry: What happened to the business? So I’m trying to think of–

Andrew: Here’s one of the things that happened. You told our producer this was 2013, right? Yeah, 2013 towards the end of the run of the business, you guys were desperate for money. You hadn’t paid your staff of ten people for over two months. You got an introduction to an angel investor who was interested in putting in some money. How much money?

Dmitry: He’s was interested to put in like $250,000.

Andrew: $250,000? You jump on a call with the guy and then what happens?

Dmitry: So we had like a few calls that he was interested. We had a final call. I was desperate because we didn’t have any money. We had a lot of responsibility, and it was like my last chance. So I jump on the final call with him and ask him if he’s made his decision. He said, “I decided not to invest. But I kind of like you, Dmitry, and I can borrow you $50,000 to you personally and then you’ll pay it back to me whenever you can in your life. Would you like to take them?” I’m like, “I should not take this responsibility personally.” And I said, “Sure, thank you.”

Andrew: You took it?

Dmitry: Yeah. I took it.

Andrew: You personally now are on the line to him, and then you used the money in the business and if the business fails, you still owe him the money.

Dmitry: Right.

Andrew: Okay. What did you guys do with the money in the business?

Dmitry: Yeah. I told you we made all possible mistakes? He saved the company, like literally. I am so grateful to him.

Andrew: Who is this person?

Dmitry: There was this company called

Andrew: OneTwoTrip?

Dmitry: Yeah. It’s a Russian travel ecommerce.

Andrew: Okay.

Dmitry: They’re pretty big. The guy’s name is Peter Kutis. So he is the founder of this company. Anyway, he borrowed me $50,000. We used it to pay the salaries to all our guys. Later on, this was like October of 2012 actually, not 2013. Later on, we won a Web Ready Award, which is like pretty famous in Russia. It was in December.

After that, we got a lot of interest from investors. A lot of investors came over to us like, “Okay, you guys have a cool project.” By that time, we had a product already, so we had some initial traction. So we were able to later on, we were able to close a round of like $500,000, April, 2013, from which we repaid this debt back to the guy.

Andrew: Okay.

Dmitry: About this time, I realized the best way to move forward would be to part our ways with our cofounders because apparently we were not the best cofounding team.

Andrew: Why? Give me one reason, one specific thing.

Dmitry: Yeah, sure. All of the guys were like extremely smart and had a lot of leadership, but then when you combine all of them, it’s like everybody has their own opinion where to go without a single authority to decide where to go. Our desire or intention to make sure that everyone is aligned kind of like in the same vision, we spent a lot of energy and a lot of time to debate instead of just failing fast and delivering on our mistakes.

Andrew: Why do you think that you didn’t either step up to be the leader or one of the other people? Was it that you were too inexperienced? Were you at the time in your life a little chicken to stand up for what you wanted?

Dmitry: You can look at it from different angles. You can look at it from this angle. The way I felt is I was the last to join the company. Those three guys had actually cofounded it and they invited me to be their CEO because I had this experience from Procter & Gamble times. However, in terms of like shares, I had the least shares in the company. I didn’t feel like it’s quite of me to step up, but you might also consider to the way I rationalize it. I could obviously do that without having a formal shares amount.

Andrew: Okay. So you stepped away. They continued with the business. I can see the business is still up and running, but I can see it hasn’t fully taken off yet. You then had this idea for something that did take off, which was essentially like Flipboard for Russia. Did you base this on Flipboard that existed in the U.S. or Flipboard, I should say for Eastern Europe. Did you base it on Flipboard in the U.S.?

Dmitry: No, not really. I based the idea of it. So I was like, “Here is this beautiful idea of Flipboard.” I used it for some time. We did not have a similar app in Russia. Flipboard is based on–is still based on social networks, like the signals they take from social networks, like this content, people share it on Twitter, on Facebook. In Russia, in Eastern Europe, we had our own set of social networks, like, or So, like Facebook was actually number five in terms of popularity.

Andrew: I see. Flipboard, for anyone who doesn’t know, it’s like a magazine reading experience on the iPad. I think they even have a iPhone version now that, like you said, the articles are curated based on the interest that you tell Flipboard you have, but also on the social signals that pick up from your friends what they’re sharing, etc. I see. So you said, “Here’s the idea. I’m going to run with it.” How well did that one do?

Dmitry: Well, it did pretty well. So the good news is that we learned from our mistakes. So I met with a friend of mine from a long time ago. We decided to partner up. We didn’t have any money, so it was like our last shot, maybe, at least this is what I told myself. I was like, “Maybe I’m not meant for the startup world. It’s time for me to go back to the corporate world and just do something there.”

I borrowed some money from my friends and he borrowed some money from his friends. We were like, “Okay, let’s do our last shot.” We knew that no one would give us money. So we didn’t spend a single second on discussion whether we should raise.

Andrew: Why wouldn’t someone give you money, because you were Russian?

Dmitry: No, not really. Both of us had our own startups and both of us failed, and we didn’t have any successful history of like raising money. Sorry, we didn’t have any successful history of like taking off with startups. Now I understand that we probably–there was a chance someone would give us money, but then it was just like, “We don’t have time to look for money. We just focus on the product and see whether it takes off or not.” We didn’t have any money to have a team. So we had very clear responsibilities between us. He was responsible for all the technical things in our startup, and I was responsible for non-technical things.

Andrew: Okay.

Dmitry: So it saved us a lot of energy like who should be focusing on what.

Andrew: I see.

Dmitry: That’s a huge learning that we had from our startups. We released our product too late. This is where we started customer development process.

Andrew: Which company? You’re saying at Penxy you started too late?

Dmitry: Penxy, yes, too late. We didn’t do customer development before.

Andrew: You did or did not?

Dmitry: We did not.

Andrew: Yeah. You said to our producer, “Listen, we launched this thing way too fast. We had not enough,” not way too fast, you just didn’t take any feedback from your customers. You just launched it internally.

Dmitry: Before launching the product, yes.

Andrew: Yeah. Okay.

Dmitry: So this time I did differently. There’s this platform in Russia called YouDo, which is kind of similar to Amazon Mechanical Turk, where you can pay people to perform small tasks. So this is where I posted a task, which would read like if you’re like a girl 18 to 23 years old and you have a smartphone and you use, I’ll pay you like a $1 for jumping on a 15-minute call where I would ask you a few questions about how you use the social network.

Andrew: What did you learn from doing all that?

Dmitry: Lots of stuff. The coolest discovery that I learned is that when you use VK app to read news, a lot of people will consume content through this app. Your status on social network is shown as online. So people know that you’re online and they start to reach out to you. A lot of people flip out because of that because they just want to consume content.

There was a bot in API. So when you use their social network through API, your status would not be shown. So, from our app because people would log in using their VK credentials, if they use our app, they wouldn’t appear as online. So this is where I figured out people don’t want to be–

Andrew: I see. VK, by the way, it’s a social network in Russia like Facebook, and just like on Facebook I could read the news. I did have one experience in the beginning of using Facebook where I was reading the news feed and suddenly people pinged me because they saw I was online. Now I’m in conversations with people when all I want to do is mindlessly read on the internet while I’m eating. Now Facebook lets you turn it off.

You’re saying VK did not let people turn it off. You said, “This is a big frustration our customers have. I can solve that frustration by letting them read inside my app without getting messages.” And that little benefit alone solved such a big pain that you were able to launch the business essentially on the back of that. Am I right, or am I missing something?

Dmitry: That was one of the value propositions that we had, the one that people clicked with a lot, but obviously not the main one.

Andrew: What was the main one?

Dmitry: The main one was the way we presented content. So the main one is we developed a technology that would essentially crawl the whole social network, and then within 15 minutes of each post, of each content item, we had an algorithm that would detect how successful this one would be. We had an opportunity to remove all ads, all crappy content, and we would just select the incredible content people would engage with like crazy.

Then we had this technology that would crawl like the whole social network and at that time, they had maybe 80 million registered users with like a lot of content being produced. We knew like what is the best content and we were able to pull this content into our app and categorize it and present it in a way that it’s designed just to consume the content. So it’s beautiful, sleek, not from [inaudible 00:22:38].

Andrew: I see. Okay. So, before we continue with the story I want to talk about my sponsor and then I’m going to get back into the happy ending that came out from this because something really interesting happened and then something else happened that led you to Chatfuel.

But first I’ve got to tell everyone that we’re recording this around April 15th. Actually, I think tax day in the U.S. is a little after April 15th this year. The thing is that every April, people get very anxious about taxes and entrepreneurs especially because there’s so much that we have to keep up with–all the revenue that comes in, all the expenses. Throughout the year, we sometimes fall behind.

The thing is that we can’t fall behind, not just because every April 15th we don’t want to have to figure out what our tax bill is and how fast we have to scramble to pay our taxes, but also because for us, knowing what our income and expenses are is the way that we keep score, the way that we know whether we’re doing well. More specifically, if we can dive into our books and understand where our biggest expenses are, we can actually critique them and figure out whether we’re spending money on the right things or not.

If you’re out there and you’re listening to me and your books are a mess, I want you know you’re not the only one. I had my books as a mess in the beginning of Mixergy. Many entrepreneurs do. But you can’t allow it to continue. So you can either go out there and hire a bookkeeper or you can do it yourself. And that’s, by the way, both of those options are what many entrepreneurs start to do. They hire a bookkeeper and then there’s an issue with the bookkeeper. Sometimes the bookkeeper ends up running away with money or making bad mistakes or maybe just disappears because they have other things coming up in their lives. Or you do it yourself, which is a nightmare because you need somebody else to look at your numbers, to scrutinize them, to make sure you’re not making mistakes yourself.

That’s where Bench comes in. I like Bench because it’s software and a service, software that sucks your data in from all the different places where you spend and make money and then people who make sure that all that data is categorized well. Imagine, just like you’re playing a videogame and you can see in the upper right corner all the time what your points are, imagine if you had the same thing for your business, where you constantly knew what your revenue was, what your expenses were, where the money went. You’d be surprised by how that improves the way you run your business.

So Bench, as you guys know, has been sponsoring Mixergy for a while now. They are offering something they’re not offering anywhere else. Yes, they’re going to give us 20% discount off the first six months with Bench. But there’s something that I want to make sure you pay attention to, which is they’re giving you a free trial. There’s a reason why a service like Bench ordinarily doesn’t do a free trial. They’re going to have to do a lot of work to get your data into their system, a lot of work. Then you can frankly walk away.

That means they’re going to get your data in order. They’re going to make sure that your finances make sense. Then if you’re not happy with it, you can walk away, there’s no obligation. All you have to do is go to This is obviously not a deal that makes a lot of sense for them, except for one thing. They’re trying to get a sense of how effective the Mixergy audience is for them as an ad buying platform. So, they want to take away all the reasons for you not to try Bench to see whether the Mixergy audience is a good group of people for them to work with.

So if you’re listening to me, you might as well take advantage of this. If there’s anyone out there you know who needs their books taken care of, you might as well pass this on to them. The URL is–because .com apparently is not cool anymore–

All right. You built this thing. Did people actually bite? Did they use it?

Dmitry: Yeah. So we did the customer development. We launched our first version. We got like 3,000 users. We checked the conversions. We analyzed what went wrong. We fixed it. Then we did like a real launch, and we made it to a top three in the app store in Russia. We got like 130,000 in one week. It was like a, “What?” moment for us.

Andrew: That’s fantastic. And then where is the revenue coming from? Since you guys were a little behind with money, you needed to find some revenue, had no funding.

Dmitry: Exactly. So right after we got some users, we were like, “We have to monetize this thing now because we don’t have any money.” So what happened is we were like, “Okay, what is it that that we can we do?” Since we had access to VK API and people were logging into our app using their VK credentials, we would run ads within our app. Then the beauty of that was that whenever a person would subscribe to a brand group, brands wanted to advertise.

We had like Pepsi, like a bunch of big names using us. The reason they chose us was because we had access to VK API, we knew exactly who subscribed to their group. So, later on, we would provide them with a report of the people who subscribed to the group and they knew those people were quality leads who owned a smartphone and they could like chat with them.

Andrew: Were they subscribing to the group using your app?

Dmitry: Yeah.

Andrew: I see. It’s not like you could go beyond your apps data to tell them who was in their groups. It’s just within your app. What were the groups in your app? I didn’t realize your app had a group.

Dmitry: So it’s not groups within our app. It’s groups that brands create on social networks. In Facebook it’s called Facebook Pages. In VK it’s called groups. So the concept is the same. It’s like a branded page for a like a bank or FMCG company. Like brands want you to subscribe or follow or like their branded page so you stay updated on what’s going on with their stuff.

Andrew: I see. So you were helping them get more people into their groups, which is the equivalent to them getting more people to be fans of their pages on Facebook. And you could tell them specifically who did it so they have a sense of the quality. I can’t tell with my advertisers. I can’t tell them who went to sign up for Bench. You were able to offer that. Okay. So I can see where the revenue would come from, from that. How high did you get the revenue before you moved on?

Dmitry: Tens of thousands of dollars per month.

Andrew: Tens of thousands?

Dmitry: Uh-huh.

Andrew: Wow. So let’s talk on a personal level. You must have been in some senses doubting yourself when the first business wasn’t doing well, right? You were struggling. Now this business does do well, tens of thousands of dollars is significant revenue. How did you feel about life? How’d you feel about yourself at that point?

Dmitry: Pretty good. I felt fantastic. A couple of things–I was living in San Francisco already, so my partner was living in Moscow and our product was focused on Eastern Europe. This is where I got this second doubt, like, “Am I good enough to build something like international or not?” So this is like, “Okay, I want to go to the next level.”

Andrew: You felt that?

Dmitry: Yeah.

Andrew: Okay.

Dmitry: This is like the imposter syndrome that I certainly have. It’s like whenever I achieve something, like what’s the next level? Am I able to get there? The second event happened that made me think we should diversify in a way, it’s like the Russian economy crisis, where the ruble to US dollar ratio would like drop 50%. My call center was in the U.S., and my value creation center was in Eastern Europe. I was like, “Okay, now we are having two times less money than we used to have. We have to diversify.”

Andrew: How’d you deal with the imposter syndrome?

Dmitry: How did I deal with the imposter syndrome?

Andrew: Yeah.

Dmitry: I have a bunch of things to deal with it. Well, first of all, I study brain science. Like one of the guys on Facebook, when you posted the announcement, he asked like, “What have those books given me?”

Andrew: Yeah. I did a post on Facebook. I said I was going to have you on. Who has any help with guidance about what could make this interview useful. It was Kunal Sampat who said, “What are the books that you recommend on Facebook? What kind of impact did they have on the work you do with bots in general?” So what’s your response?

Dmitry: My response is like those books are all about brain science and the way we operate and why we do the things we do. They are so fundamental. It’s not only about bots and the business. It’s just about how you treat yourself and how do you respond to the external and internal triggers that you get in life.

Andrew: So give me one thing you do that allows you to deal with the imposter syndrome. Don’t make me read all the books. Let me have a little bit of insight that you’ve gotten from reading those books, and then we can go back and look at the book list.

Dmitry: One thing I like is the gratitude practice. Gratitude practice is like every day when I wake up, I write down three things I’m grateful for in my life. One is to me personally, the second one in my relationship, and the third one is professionally. Every time before I go to bed, I write down three things that I am thankful for. I’m grateful for what happened during the day. The reason it’s so helpful is because it changes your perspective.

There’s a bias called selective perception bias. It assesses that we train our minds to focus on certain things. There was an experiment where people were locked up in a lab for a week and they were forced to play Tetris seven days. Then they were let out and they were asked, “What is it that you guys see?” They were like, “We walk down the street and we see two beautiful buildings. If you just put them all together, they would nicely collapse because the Tetris patterns.”

So whatever we train our mind to focus on, we tend to see these things and we tend to have this lens, the way we perceive the reality through this lens and we can change this lens. So this regular practice, it helps you retrain your mind to think positively on whatever happens in your life.

Andrew: And so why does thinking positively help you with the imposter syndrome or help you at all in any way?

Dmitry: Because when you think, “I’m not good enough,” you’re like, “It’s a great opportunity for you to improve on this topic.” Whenever you feel like it’s a great opportunity, it’s like a positive sense, like, “That’s actually cool that I have it. I cannot fight against it. It’s working for me.”

Andrew: I see. So it trains you to not succumb to that thought process but to see the advantage in it. I see. Got it. Where is this book list? I’m looking on your Facebook page. I’ve Googled you. I’ve gone to your Medium page. I don’t see it.

Dmitry: It’s on my Facebook page somewhere in 2015, I think. I can send it to you so you can post it afterwards.

Andrew: Okay.

Dmitry: This is like the best books I’ve ever read in my life. The number one book is written by Daniel Kahneman and Tversky. Kahneman, won a Nobel Prize in 2002 for actually proving that people are irrational. He has this book “Thinking Fast and Slow,” which is super fundamental for understanding how we humans operate and that most of our decisions are emotional and not rational.

Andrew: All right. I want to get back into this because one of the things I liked about you when we had dinner was you talked about how you create habits for yourself, how when a habit doesn’t stick, how you adjust your life for it and you get really into this stuff in a way that I think is interesting. Let’s continue with the story and come back to that and also come back to our predictions about where bots are going. I feel like your prediction is going to be better than mine because you’re in a better place to see what’s coming. But I don’t think that four or even two years out our predictions are going to be that different.

All right. So then what did you end up doing? What’s the name of this company that you had that did news?

Dmitry: It’s called Myata. It was called Myata. In English, it means mint.

Andrew: Okay. Mata?

Dmitry: Well, if you translate it to English, it’s like mint as an herb, M-I-N-T. In Russian, that would spell like M-Y-A-T-A.

Andrew: Okay. All right. So what happened with it?

Dmitry: We sold it in February, 2015 to a group of private investors back in Russia.

Andrew: Okay. How much?

Dmitry: I cannot disclose.

Andrew: Give me a ballpark. Are we talking about hundreds of thousands of dollars, millions?

Dmitry: Hundreds of thousands.

Andrew: All right. Enough for you to say, “Okay, I’ve got something here. I’m going to start a new business. I’ve got an understanding of how this works and how I work. And then you somehow created a bot directory. Why did you create a bot directory?

Dmitry: So I was looking for something new to jump on. This is in like 2015, Telegram, which is a popular messenger app that has like over 100 million monthly users, they launched, they opened up their API for bots. Actually, one of the first–well, first western messenger to open up for bots. Kik had it before, but Kik had it in a private beta. Telegram went like all public.

So I remember like that exact day I was reading the announcement and I was like okay, I have this unique opportunity of building something and to be the first on the platform. If you were there when App Store launched and you had the opportunity to launch any app, what would that be, like knowing how huge this would become?

Andrew: I see. How did you know, by the way, it’s going to become big? A lot of people said, “You know what? It’s Telegram. It’s kind of small,” or, “It’s a bot platform. Try it. It doesn’t make sense. Move on.” What was it about this whole bot platform that drew you in?

Dmitry: It feels like, again, it felt like a new platform, like a new paradigm shift. That’s what people called it later. I didn’t have those exact words in my mind before. But it felt like again, a new entity is coming and this entity is going to be very popular because it’s facing so many users. WeChat had it already. Kik was experimenting with it. Telegram was experimenting with it. There were good chances that other players would jump on board. So, obviously, there was some sort of a risk, but you’ve got to take some risks in order to win big, I think.

Andrew: Okay. Yeah. I think the fact that it was already happening in Asia was really encouraging, that if you see that it’s existing there, then you could kind of see maybe not the exact future, but a possible future. So you said, “This is coming. I see an opportunity here.” What’s the first thing you built for it?

Dmitry: Well, I had a bunch of ideas, but none of them actually were as a good–I had this opportunity to launch something, like to be number one. What would that be to help me to be successful? And then I was like, “Okay, I have no idea what that would be because I didn’t know what’s going to work on this platform, and if I don’t know what’s going to work, then I have to somehow establish myself so I have visibility on what’s going to work.”

I was like okay, the way to do it is to launch a bot store on top of the platform. So I called my cofounder, Artem. It was like 2:00 a.m. for him in Moscow. I was like, “Artem, wake up. We’ve got to build a bot store.” He was like, “Okay.”

Andrew: Let me highlight that thought strategy. You were saying, “Look, my strategy is I need to understand what works and what doesn’t because it’s a brand new platform. How do I understand what’s good and what’s not? Well, I’ll create a bot store, a platform where someone can come and find the right bots, and then that will give me insight into what’s most popular.” I’ve heard several entrepreneurs come up with techniques to understand what works and what doesn’t.

One of the ones that stuck in my head was in early interview with Hiten Shah, I asked him, “What would you start if you have to start with something brand new?” He said, “The first thing is an analytics company.” I said, “Why analytics, there are so many out there?” He said, “Yeah, but then that would give me deep insight into what people are doing and what’s working and then I can understand what needs to be created.”

All right. So yours was a bot store. What was the bot store going to be, just a standard website with a list of links?

Dmitry: It actually started with a bot.

Andrew: A bot that recommended other bots?

Dmitry: Exactly.

Andrew: Okay.

Dmitry: So, by that time, there were only 10 bots created by Telegram’s team. We launched bot number 11, where we had the other 10 bots. It was so funny because this bot store contained only 10 bots, but this is how you start. We built it so fast. We built it like within like a few hours, and I sent it to a founder of Telegram. He liked it. He was like, “Oh wow. That’s pretty cool.”

So he also added this bot to his blog post about the launch of the platform. This is how we got our first 15,000 users. But then later on, I realized that traffic is dying out because people–there was no way to–how do you find a bot store if there is no discovery mechanism and the bot store is supposed to be a discovery mechanism?

Andrew: Yeah.

Dmitry: That bothered me in a big way, because obviously this whole thing was dying and wasn’t getting the exposure and something they wanted to have. So, luckily, one day before going to bed, I was lying in bed and I was like, “What is that?” Suddenly it popped out. We created this technique on how you promote your bot. So if you want your bot to appear higher within the bot store, you ask your users to rate it and review it. Then your users, by rating it and reviewing it, they would have to activate in our bot store.

Andrew: I see.

Dmitry: This is how we started to get a lot of users. Once we had this figured out, it just started to grow like crazy.

Andrew: And this is, the site is still up, right?

Dmitry: Exactly.

Andrew: I do see a lot of ratings, like Andy, which helps people learn English, has 23,000+ ratings. Every one of those people would need to be connected to your bot. Is that right?

Dmitry: Well, if they wanted to review and rate their bot, yes. 23,000, that is correct.

Andrew: To review you have to subscribe to the bot, so that helped you guys grow. That was super clever. By the way, Microsoft is trying to get into bots now. People don’t realize that Skype has bots. Microsoft has a directory very similar to StoreBot

My guess is they’re trying to get a sense of what’s working and what’s not and also tell people who are looking for bots for Facebook Messenger, “Hey, this also exists on our platform or here are a few others.” I forget what it’s called, but I’ve been using it to get a sense of what’s working for bots too because they have numbers that they’re publishing. Do you happen to know what that is? I’ll look for it and come back to it later in the year.

Dmitry: The Microsoft one?

Andrew: Microsoft, yeah. They have a directory.

Dmitry: Yeah, they call it a Directory.

Andrew: They call it Directory? That’s it?

Dmitry: I think so.

Andrew: Okay. Microsoft directory bot messenger Skype–let me see if Googling that will bring it up. It’s on There’s a directory in there, I’m pretty sure. All right. I’m going to take a moment and we’re going to come back to–there, it’s on The direct link is When you go in there, you can see the bots that already exist and then you can also get a sense of what’s especially popular.

Okay. My sponsor is a company called Toptal. As you guys know, if you need a developer, there is, I think, no better place to get a top developer than to go to Toptal. If you’re listening to me and you want to use Toptal, do this–email me. I’ll make an introduction to my person at Toptal. Email and I’ll introduce you to my friend at Toptal.

There are several entrepreneurs that listen to me who have actually used Toptal to build their businesses. In fact, one guy who came here brought in a couple of bottles of scotch to–do you drink, by the way, Dmitry?

Dmitry: No.

Andrew: You don’t. Yeah. I didn’t think you did. It takes your mind off of things, is that right? You’re too focused.

Dmitry: ROI is negative. You invest your time, your body feels worse. I’m not having positive ROI from this experience.

Andrew: That’s the best line, “I do not drink because the ROI is negative on it.” Well, Eric did come over. He did have a drink. I think we had a positive ROI from it, because I’m about to mention his site for free. I think that’s one of the benefits of drinking sometimes, that you get to bond with people, but it’s not the only way to do it, so I get it.

Eric Brown came to the office. He brought me a couple of bottles of scotch. He told me about his business. It’s Permea. You guys can see it at He said, “Look, it’s strategic portfolio management. I want to help companies organize the way that they operate in a smart way.” He showed me his software. It looked really good. I said, “Who built it?” He said, “Toptal. I hired them. I heard you talk about them and it worked out beautifully.” He’s got his business set up with Toptal developers.

If you’re out there looking to hire great developers–we’re talking about the kinds of people that can work at Google, at Facebook, at Chatfuel–you owe it to yourself to go to talk to Toptal. The first thing they’ll do is have a conversation with you where they understand what you’re looking for, then they’ll introduce you to the right person based on whether you need a full-time, part-time, what kind of platform you’re working on, etc.

All you have to do is go to or email me, What you will get is 80 hours of Toptal developer credit for free when you pay for your first 80 hours. That’s in addition to a no risk trial period of up to two weeks. Go check them out,, top as in top of the mountain, tal as in talent, These guys just keep coming back because it works so well.

All right. What happened to your bot store?

Dmitry: What happened to our bot store? Everything was great. We were like by the end of the summer, we had millions of users, I think tens of thousands of chat bots uploaded. The network effect was so strong, no one could compete with us. Still, I think no one is competing with us. The only competition we could have faced could be from Telegram itself because obviously if you don’t own the platform–if you don’t own the messenger, you don’t own the platform.

Anyway, so things were going great. We started to notice that certain categories of chat bots are taking off. Those chat bots were very simple chat bots, but what they were doing, they were just channeling media to people. So they would send some funny pictures, videos, maybe quotes, smart wisdom, whatever, this kind of stuff. People seemed to like it and they were super simple.

Andrew: Not the bots where you were supposed to communicate back, like you’re talking to a human being, but the ones that just give basic information.

Dmitry: Right.

Andrew: I feel like with every platform things start out that way. I’m thinking about the web really started out doing well with horoscopes and the weather. Then email decided this was a good way to get reach with people. So we’ll send them horoscopes and the weather. So you’re noticing that same thing because you’re looking at your bot store. What else happening?

Dmitry: We were like, “Okay, this is a great opportunity and something we should jump on.” We wanted to have a simple bot building tools that would allow you to create those kinds of bots.

Andrew: Why a bot building tool? Why didn’t you say, “I see the content works, I’m going to create a content play?” Kind of like Jason Calacanis is doing with now, where you take the biggest topics and you create bots for them.

Dmitry: Well, to me, when you own a platform like this, you have much more leverage, much more visibility on what’s going on. You have a lot of network effect. You are able to monetize it in different ways, just like another level of visibility for a business, I think because you don’t have to compete with another person who decides to do the same competition in time and space. Based on our experience with our app, which was like content delivery app in essence, the competition was very tough.

One of the reasons we decided to sell the company was because it was harder and harder to acquire new users, right? The cost of customer acquisition was constantly increasing, and the LTV, if your app is ad model is constantly going down. It’s something you cannot compete with in the long term because some companies, they’re willing to invest more for cost of customer acquisition because their LTV is higher and you have to compete with all of them, so your model is simply not working anymore. We saw that coming.

Andrew: Okay.

Dmitry: So we’re like we don’t want to be competing with them, but we’d rather facilitate the competition and let the thousand flowers bloom and see what’s going on. So we had a few options. There was a guy who actually launched a bot building platform. So we had an option to launch our own. We messaged that guy and he seemed like an awesome person and we offered him to join our company. We did a small acquisition, and we used the money that we got from selling the Myata company. This guy joined our team, and that’s how we got our bot building platform.

Andrew: Because he had a platform already?

Dmitry: Yeah.

Andrew: This was working on Telegram.

Dmitry: Exactly.

Andrew: I see. So why did he want to sell to you guys, because he wanted to cash out a little bit?

Dmitry: Not really. Since we had the bot store–well, both things. We had a bot store and I explicitly explained to him that it’s in my plans to launch this kind of platform and it’s probably going to be hard to compete with us because we owned a distribution store. I think it might make sense for us to join efforts and work as a single team versus compete with each other.

Andrew: I think that’s the way that Microsoft’s Bill Gates used to make deals, “Look, we have the platform. We’re going to get into this anyway. You might as well sell to us instead of competing with us. It’s going to be harder to compete with the people who have the platform.”

Dmitry: I think it’s like a lot of times things like that happen. For example, like Instagram acquisition, I think, was kind of the same deal. I don’t know. Anyway, we decided to move on together. This is how we got our first bot building platform. It started to grow because we’d combine our efforts, and it was a great synergy effect. Then later on, we decided to launch our ad network on top of it.

So we essentially like owned the whole ecosystem based on telegram. We owned a distribution tool, which was a store. We owned a bot building platform. We had a monetization platform, where you could plug in your bot and essentially start to monetize it. We had like great synergy within all three of them because we could cross promote them.

Andrew: I see. Okay. So why did you go and raise money from Y Combinator?

Dmitry: So we were growing fast. The way Y Combinator felt to me and feels to me, and actually is, is that they can take your growth and multiply it by ten, at least, if you work your ass off. So we wanted to be there and we applied to Y Combinator in November. We got accepted luckily. We kept on working. I think Y Combinator believed in us and in our promise that other platforms will open up because obviously Telegram would be too small for Y Combinator to bet on. So they were betting on Facebook, and we were getting ready to jump on Facebook.

Andrew: And what did they give you to help you grow? Is there one piece of knowledge, one introduction, anything you can point to?

Dmitry: Lots of things. We got a bunch of introductions. We got a lot of not just [inaudible 00:51:43]. Every two weeks you have a meeting with fellow companies that are in your group and all of you share your progress. You sit next to someone who’s saying, “During the past two weeks, we closed deals for like $20 million.” You’re like, “Oh, shit.” You feel like shit and you go out like, “Oh my god, I’ve got to work my ass off because everybody’s killing it and I’m like the worst one.” That’s so funny.

So it was a great experience, for sure. But ultimately one of the biggest things that Y Combinator helped with is fundraising, that’s for sure. You are able to get the best investors. The way the whole process is organized, you spend as little time as possible actually fundraising.

Andrew: Okay. How much money did you raise and from whom? I know I saw it, but I don’t see it now on your AngelList page.

Dmitry: Yeah. So, as part of our deal with Yandex, we cannot disclose this amount, but I can say that it was a classic Silicon Valley seed kind of amount, like a few million dollars.

Andrew: Okay. Yandex is a huge Russian company. Why them and not a US company considering that you want to be more US-focused, Facebook specifically?

Dmitry: So Yandex, we started a conversation with them in October. This is where they got interested. We kept the discussion going. The reason we decided to go with them, they have amazing technical expertise and they were willing to share it with us. I think it’s crucial for this kind of company like we are, where you have to have a lot of knowledge with the NLP engine, on how to setup AAI things. You have a lot of experience and people to talk to.

So it’s really helpful. It was like one of the cases where an investor is actually not only out to get money, but also with the experience connections and true knowledge on how to build things. Yandex has a lot of knowledge on how to build things because they’re the Google of Russia, right?

Andrew: I see. Did 500 Startups also invest?

Dmitry: They did invest, yes.

Andrew: So they invested in you and in one of your competitors?

Dmitry: So the way it worked is the following–we went through 500 Startups’ accelerator with the Myata app. This is where they invested. We went through their program. Later on, another company that is our competitor got accepted to 500 Startups as well. They went through their acceleration program.

Andrew: I see. So you guys went through different paths. Did Google want to buy you guys? Be open with me. They tried to buy you, didn’t they?

Dmitry: Wall Street Journal published a piece where they say Google wanted to acquire us. I am not able to comment on that.

Andrew: All right. I kind of feel like maybe they didn’t because you could–well, who knows? Interesting. Let’s talk about how you grew. I understand the philosophy. I understand how you got here. You grew doing what? How did you get users? How did you get people to create bots on a platform when the platform is so new? And frankly, as much as I talk about it now, I see there’s a lot of hesitation. People don’t want to get into bots.

Dmitry: Great question. How did we grow? We wanted to be the first one the platform. So what happened is in February, I think, we were like okay, we have all of these thousands of bots created on Telegram with simple channel information. What is the next level of bots for publishers, who were a big chunk of our business?

I came up with this idea of the first AI-based publisher bot, so a bot where you would ask like real questions like, “Who is Elon Musk? Show me the latest news about Elon Musk.” We came up with an MVP for Forbes. I reached out to Forbes saying, “Hey, guys, you want to launch the first ever AI-based chat bot?” They said, “Sure.” So we launched it during MWC of 2016. Oh my god, just a year ago, wow.

So we launched it and instantly a lot of publishers reached out to me like, “Forbes launched their own bot. We want to have a bot as well.” So it worked fine. Then this is where we got in touch with Facebook. They were like, “Oh my god, guys, you have this amazing bot. We have this early access program, and we want to launch some experience in F8. Would you like to be a part of it?” We said, “Sure.” So we were working on the TechCrunch bot to be launched during F8. It got launched right after it.

The same time we were working on our platform, on version two of our platform that we can launch and make sure that we are live as soon as possible and we give people this opportunity to create chat bots and play with them and experiment. We had a lot of experience from our Telegram times. So it wasn’t like something that we had to develop from scratch. It worked fine for us.

We knew that it’s got to be like super simple. I think we can do even better when we’re working towards it. I want it to be as simple as possible. I think what we have right now is probably the best experience out there in terms of simplicity and takes you minutes to start and go ahead and play with it and building our own chat bot and later on build on it and build sophisticated experience using plug-ins.

Andrew: It was you reaching out to Forbes.

Dmitry: Yeah.

Andrew: Forbes got a chat bot that other people said, “If Forbes has one, I want one too.” Were you just cold emailing in to Forbes? I forget. You told me who you contacted.

Dmitry: So I had a contact at Forbes. I was given an interview once. I reached out to her and she reached out to her chief editor.

Andrew: I see. Got it. Then they created a bot. You created one for them for free. What did the bot do? I forget.

Dmitry: It was a publisher bot. We had certain categories you could subscribe to. You could ask questions like, “Who is Elon Musk? Who is this? What is that? Show me the latest news about Microsoft.”

Andrew: All right. And then others publishers signed on. You get on stage or at least–were you on stage at F8?

Dmitry: I wasn’t on stage at F8.

Andrew: You were not. But they talked about Chatfuel on stage?

Dmitry: They didn’t mention any company from stage. They didn’t like–just as far as I remember, no, no companies were mentioned. I think it was too early. But we launched our chat bot, it was like the number seven chat bot on the platform.

Andrew: What was number seven?

Dmitry: TechCrunch.

Andrew: I see. Because you launched TechCrunch, I remember, then TechCrunch became one of the models at the time. Then I had to go and check out what TechCrunch did and then I said, “There’s this menu. Who the hell is this Chatfuel?” Then I saw Chatfuel’s website.

Dmitry: There you go. That’s how we got our first number of users. This is how we kept growing.

Andrew: What about agencies? It seems like the seven shades of whatever–what is that movie called?

Dmitry: “50 Shades of Grey.”

Andrew: “50 Shades of Grey,” thank you. I’m so out of touch with pop culture, “50 Shades of Grey,” I’m assuming it was the agency that works with them that did it, right?

Dmitry: Yeah.

Andrew: How do you get all these agencies to sign up for Chatfuel to create their bots?

Dmitry: So we were lucky. We have a lot of agencies actually working on top of Chatfuel. I think the reason they like is it’s so simple to launch their experiences and not only launch but also maintain them because like chat bot is a live organism. You have to change it constantly. It’s not like something you can develop and then fill out change requests on a weekly basis or a monthly basis. You iterate daily maybe, sometimes. You check what’s going on and you see a dead end the user would fall in and you have to change it.

The simplicity of the platform, the ability of the platform to be used by non-technical people, typically it’s an account manager of a client who supports the client. They don’t need to go back to their development team. The multitude of experiences that you can create on top of Chatfuel, those were the things that actually brought a lot of agencies to work on top of us. We have some really big names to work on top of us.

Andrew: Do you guys have anyone internally who reaches out to these agencies?

Dmitry: We have a head of partnerships, Dmitriy Kachin, super awesome guy. He works mostly on an inbound basis. So we have like a lot of inbound.

Andrew: I see. They do what I did, discover Chatfuel and then they reach out and then they ask you guys for what? How do they know to even ask for you guys?

Dmitry: So they see Chatfuel. We have a certain section on our website, reach out to us. They reach out. They ask for like maybe some sort of custom development or like how do we work with high volumes and so on and so forth. We have agency onboarding checklists. So we basically see if this agency is serious about chat bots. They want to go all in there and invest in.

We teach them and share our best practices, teach them to how to communicate with clients, like what is the best price model that could be for them, what are the best experiences we’ve seen on our platform, what are the things that engage users the most, convert users the most. We share all this knowledge so it’s like a mutually beneficial process. We help them to build the best bots possible for their clients.

Andrew: Okay. What do you guys do to check out user experience? This is a brand new way of dealing with customers. What’s your process?

Dmitry: We have a lot of internal tracking tools in place. So, for example, we have a bot that notifies us whenever a certain bot is performing exceptionally well. So, like user experience or growth or maybe like users are actually unsubscribing from the bot, so we kind of try to analyze all those outliers. From time to time, we do deep customer discovery sessions.

So, in a certain vertical, for example, like publisher bots or FAQ bots or like marketing bots, we try to find like top experience on our platform and reach out to our users to create bots directly and we set up half an hour calls with them and we ask like what are their pains, what’s working find for them, what’s not working fine for them. So, it’s both quantitative and qualitative customer development, constant customer development process that we have in place.

Andrew: I see. Is there something that you remember that’s come up from having these pain conversations with customers?

Dmitry: Like yeah, of course.

Andrew: What? Give me an example of something that’s come up because of that.

Dmitry: I think the segmentation feature came up based on the discussions.

Andrew: Segmenting . . .

Dmitry: Meaning the ability to assign user attributes, infinite number of user attributes to users and then later on be able to reach out back to them based on those certain attributes.

Andrew: I see. So I can ask somebody, for example, “Are you a man or a woman?” Actually, I wouldn’t need to do that because Facebook would tell me that. I can ask them, “Do you have a company now or are you thinking of starting a company?” Whatever they respond, I save in one of the attributes as they’re using the attribute feature of Chatfuel. And then I can come back in and say, “I only want to message people who have no company and want to start one.” Got it. I see. That’s come up from that.

All right. Where do you see this going? Let’s talk about the future.

Dmitry: Sure. I think the next big step we’re going to see is that like currently most of the chat bots, you can describe the way they communication with users as a kind of marketing way of communication. The shift that I foresee and that needs to happen is we’re going to switch to a transactional majority of use cases. Basically, when you use your messaging app to interact with objects in the real world and your messaging app is acting as a platform based on which you have lots of transactions.

So, for example, when I owe you like $10 for the dinner that we have, I don’t have to pay you cash or send you Venmo or like any other tools. I just scan your QR code and I pay you. Or when we go to a dinner, I can pay for a dinner using the scan code. So, basically, your messaging app turns into your bank account. Each is from mostly a place where you receive marketing communication or inbound things to place where you have your transaction going.

Andrew: I see. But it’s going to be more than just in-person payments, right?

Dmitry: Of course.

Andrew: Here’s what I see. Tell me what you see. I know for sure that you have a clear sense of this. You know what’s coming up in a couple of weeks that I wouldn’t know about. I feel like this. One of the good examples of this is Hipmunk. Going to Hipmunk’s website is a whole new experience, especially for first-time people. There’s a lot of new stuff going on including the pain rating and so on. Installing the app is another pain in the ass. I don’t want to install your app just because I need a flight. I just need a flight to go to New York.

The ability to pull up the phone and just chat with Hipmunk and say, “I need a flight to New York,” it exists right now. They come back and say, “Where are you leaving?” I say it. They say, “Where are you leaving from?” I say that. They say, “Here are the five different flights that we think you’d want.” Then I get to either click to buy it or I get to say, “Tell me when the price gets lower,” and then they keep pinging me whenever the price gets lower.

That is an interesting experience to me. To me that feels like a good vision of what the future could be. You don’t have to install the app. You don’t have to go to a website. You just chat like you would with a human being, like I would frankly with Andrea, my assistant. I wouldn’t try to go figure out her interface every time I want to do something different. I just text her and say, “Can you contact the babysitter to say we don’t need a sitter today. Can you go to the grocery app and go buy me groceries?” That seems like the future. Am I right?

Dmitry: Yeah. That’s very right. What you described is the first next step we’re going to see. What I’m saying the next big step that we’re going to face is through your messaging app, you will be able to interact with a real world object, not only like a ticket you can purchase through a messaging app, but let’s say you go to the airport and you use it for navigating yourself in the airport or paying for things in the real world you are physically present.

Andrew: Why would I do that instead of paying using Apple Pay, for example, or Android Pay or Android Wallet or whatever they’re calling it now?

Dmitry: Because you have not only the payment feature, but you also have extended functionality. So businesses would want to do that because they will have an opportunity to reach back to you with some special promotions. So a benefit for them is obvious. For you the benefit is when you’re scanning a QR code, you have multiple options, for example buy like the basic package or premium package or notify me about a discount or subscribing to certain things.

Andrew: I see.

Dmitry: So it’s like enriching functionality on the user side and giving an opportunity to reach back to customers from the business side.

Andrew: Got it. I see. Okay. What about this whole web view thing that’s coming into Messenger and other chat apps? What do you see there?

Dmitry: Yeah. It’s a natural transition. The reason we are seeing this web view come into Messenger is that this experience has got to be as frictionless as possible. So, if a certain action requires, let’s say, ten interactions and can be done using five interactions if it’s a web view with some custom controls, why not because it’s faster.

Andrew: So the example that I might give for that to illustrate it is right now all I’ve talked about is buying a ticket on Hipmunk and I’m finding the lowest rate. At some point, I’m going to want to pick my seat. I don’t want the chat bot to say, “We have seat F and it’s a window seat. We have D, which is an aisle seat, etc.” I just want to see a picture of the airplane with the available spots. I tap the ones that are available, right? That’s what it is. Okay. I see it comes up on that.

So that’s the next part. Here’s where I think we disagree. I feel like right now nobody thinks to go to Facebook Messenger to buy a drink, to buy a flight, to do any of it. I think that we need a reason to keep reaching out to users. To me, the best reason is the same thing that’s worked on the web–subscriptions.

So it would be interesting if Hipmunk didn’t say, “Come to Hipmunk when you’re ready to buy a ticket.” I’m not going to think to go to Hipmunk’s chat when I’m ready to buy a ticket. It would be interesting if they said, “Hey, Andrew, tell us what your favorite cities are and we’re going to tell you every week when there’s a discount on those cities,” or if there happens to be an extreme price drop, like below $100 from San Francisco to New York, we’ll let you know about that. That would be an interesting way for them to stay in my life on a regular basis, true or false?

Dmitry: Well, sure, I agree with that. We should understand that even though we’re seeing amazingly high open rates and click-through rates right now, that’s going to change exactly the way it changed with email, right? The more notifications you get, the less actions you’re going to take over time because it becomes crowded in your Messenger. This is just one of the external triggers that we might use in order to engage with a user.

Another external trigger–again, this is something I see it going in the mid-term future–let’s say you come to the airport and your flight is cancelled or you miss your flight and you have this QR code of Hipmunk on the column and you scan it right there and through Hipmunk, you’re like, “Where are you going?” and you interact with it and you get a new ticket right there.

Andrew: Or even with my airline, that would be good.

Dmitry: Another external trigger is a physical world object that you can start interacting with or that would serve as an entry point if the context is right.

Andrew: I see. The reason that a company would want to do this is the company that has Andrew check for flights on their website, like if I go to Hipmunk directly, they don’t know my name. They don’t know my gender. They don’t know where I am. They don’t know jack about me. But if I interact with them first via Facebook Messenger, they get to know my name. They get to know what city I’m in and they get a little bit more context and that’s the advantage for them for starting out there.

All right. Mike Townsend responded to my Facebook post asking what I should be asking you with an interesting question that we’ve thought about internally here at Mixergy. I know that if you and I are in a chat on Facebook Messenger, at some point, it might be too much to go back and forth with text. I just say, “Hey, hit the call button, let’s talk.”

I would like my people here at Mixergy, if someone has a problem with our sequence in our Facebook Messenger account to be able to say, “Hit this button and call us.” Right now that happens over the phone. Do you think that’s ever going to happen in Facebook Messenger and other apps, where the user will be able to hit a button and talk to a real human being at the company that manages that chat bot?

Dmitry: Uh-huh.

Andrew: You see that coming?

Dmitry: That’s a great question. Technologically there is no constraint except for if Facebook decides to implement that. Chances are, I don’t know. But then from a company perspective, it’s probably something you want to avoid because currently IVRs exist. Typically people hate them because IVR is intentionally made very complicated.

Andrew: What’s an IVR?

Dmitry: Interactive voice responder. It’s like if you go to a bank, it’s like, “Press one to switch to Spanish,” or, “Press two to get to know your bank account.”

Andrew: Those are awful. Apple has a good one, where they say, “Tell me what you want to do.” And I say, “I want to repair a phone.” They get it wrong a little bit.

Dmitry: Exactly. They get it wrong a little bit. There’s a problem. They never intentionally make it complicated, because it’s super expensive for them to have a lot of people answering phone calls real time. They try to shorten the funnel as much as possible.

Andrew: Yes.

Dmitry: So only if you are desperate and you hate the company, you get through all the steps of the funnel, you get to talk to a real human who might be able to solve your problem. So it’s like technologically there is no constraint.

From a business point of view, you as a business owner has to evaluate whether that makes sense for your business to have this customer support–voice-enabled customer support real time. It might cost you a lot of money, but if you know your customers are paying you to see your strategy, why not? But you’ve got to justify it from the business standpoint.

Andrew: Right. I feel like what you do is wait for the right time in the conversation to jump in, right when someone’s about to buy or curious, then you make that available. Then finally, David Williams came in and said, “When do you think we’re going to have. . .” let me ask his question the way he asked, “Ask him if any plans to let people import natural language dialogue pairs to automatically train bots to have responses to a wider variety of inputs.”

I guess what he’s asking is–tell me if you understood it differently–he’s asking I’ve got a frequently asked questions on my site. Can I take all the questions that people have asked and the answers we give them and put that into our bot so that when someone asks the same question we already answered on the site, even if they don’t ask it exactly as we answered it on the site, Chatfuel will understand and spit out the right answer? Do you think that’s possible any time soon?

Dmitry: It is possible, actually very soon. We are working on this feature. As soon as we are happy with the performance of this feature and the quality of input versus output, we’ll roll it out.

Andrew: Okay. Wow. So you guys are working on it right now. You’re thinking months away or weeks away or years?

Dmitry: As soon as we are happy with the quality. The most important here is–we’ve seen a bunch of shitty built chat bots. You don’t want to launch a feature for the sake of a feature. You want to launch it and make sure that it works just right.

Andrew: All right. I’ve definitely gone over. The only thing that I couldn’t get into this interview is the way that you like hack your life. I’m fascinated by it, but I think we’ll have to save it for the next visit that you have here.

Dmitry: One more interview?

Andrew: Sorry?

Dmitry: One more interview?

Andrew: I’ll freaking do that. I was fascinated how you said, “I have this habit and then it breaks down, then I analyze why it broke down and then I go and I fix it.” So maybe next interview maybe we’ll save it for someone to do in person with you.

If you’re meeting Dmitry, ask Dmitry about this personal hacking and you’re going to have your mind blown. I think there’s a reason why you are where you are and also there’s a reason you’re not getting distracted by television shows and video games the way other people do. Your focus is well structured. It’s not just intense, it’s well organized.

All right. For anyone out there who wants to create their own bot, Chatfuel is the place that many people go and create it. I think it’s the majority of bots are now created on Chatfuel, right? If not the majority, clearly the plurality. You’re definitely the most popular platform for creating a bot and you’re 100% free because Y Combinator is footing the bill for you and Yandex and because you’re in Silicon Valley.

Dmitry: And other great companies we are grateful for.

Andrew: Sorry?

Dmitry: And other great investors we are grateful for.

Andrew: And other great investors, right. I didn’t even get into a list of other people. Let me see who else–Knight Foundation, for example, took part in it. All right. I won’t get into the whole list. I will say this too. We’re experimenting constantly with messaging bots here at Mixergy. I’ve got my own attitude on what you think works.

If you want to try ours and see how it is, it doesn’t exist right now. So, Dmitry, don’t go and check it out, but by the time this interview is published, they will be able to go to I bought a special URL just to make it easy,, where you can see how our bot works. If you want to hire someone who’s learned from us at Mixergy to create a bot for you, you can do that too, and if you want to learn how to create one, we’ll help you do that also and that’s

I’m fascinated by this stuff, Dmitry. I feel like you’re onto something huge here. All right. And I’ll thank my two sponsors, and Thanks, Dmitry. By everyone.

Dmitry: Andrew, thank you so much. Pleasure. Huge pleasure as always.

Andrew: This was fantastic. Thanks.

Dmitry: Take care.

Andrew: Hope to see you in person again. Bye.

Who should we feature on Mixergy? Let us know who you think would make a great interviewee.