Why fight in a goldrush when you can sell pickaxes instead? That’s exactly what our next guest did. On this 131st episode of the E-Comm Show, Andrew Maff interviews Ruslan Leteyski, CEO of Zipchat AI. Starting as a product designer and struggling to build his own e-commerce brand, Ruslan decided to lean on his strengths and create AI solutions for sellers instead.
Ruslan dives into the power of AI and the success it’s brought his companies, providing a true testament to the lucrative opportunities within the AI market. But in a market so overwhelming, where do you start? As demonstrated by Zipchat AI, it all starts with listening to the market’s wants and needs.
Watch the full episode below, or visit TheEcommShow.com for more.
If you enjoyed the show, please rate, review, and SUBSCRIBE!
Have an e-commerce marketing question you'd like Andrew to cover in an upcoming episode? Email: hello@theecommshow.com
Selling Pickaxes During a Gold Rush: The Opportunities of AI with Zipchat AI
Andrew Maff and Ruslan Leteyski
CONNECT WITH OUR HOST: AndrewMaff.com | Twitter: @AndrewMaff | LinkedIn: @AndrewMaff
-
Ruslan Leteyski a seasoned entrepreneur who has started several companies focused in the e-commerce and SaaS spaces. He's the creator of Checkout X, a one-page checkout solution for Shopify, and Vanga AI, an AI tool for boosting e-commerce store sales that was later acquired. Currently, he's the CEO of Zipchat ( www.zipchat.ai ), which offers an AI chat service for e-commerce stores aimed at increasing sales, collecting customer data without invading privacy, and automating marketing. Ruslan is all about making online shopping better and helping businesses grow. Additionally, Ruslan shares lessons from his entrepreneurial journey on his blog: blog.leteyski.com.
00:00
I recommend products you know push them to buy products and also give you insights on what was wrong with your business what's great about your business all this kind of stuff
01:06
hello everyone and welcome to another episode of The E-Comm Show. I'm your host as usual Andrew Maff and today I'm joined by Ruslan Leteyski of the CEO of Snapchat AI Ruslan how're you doing? You're good show.
01:18
Yeah, I'm doing great. Thanks for inviting me here. And I'm really excited to talk to you. Yeah,
01:24
super excited to have you on the show. Obviously, the world of AI is constantly growing, and it is sometimes overwhelming. So it's always nice to have someone on the show to clear it up a little bit. I'd love to do the usual here and give you an opportunity. Let everyone know where they can like, teach us a little bit about your background. And a little bit more about zip chat. And we'll take it from there. Okay. Yeah,
01:48
perfect. So I started out in software, I worked as a product designer for a couple years. And then I moved into E commerce by building my own drop shipping store back in the days where you could dropship stuff from AliExpress and so on Facebook. Good. That is right. But you know, I'm with a background in software. So I was thinking instead of competing with all these amazing intrapreneurs that were that were beating the shit out of me in Facebook ads, I thought maybe I can actually help them be better and start building software for Shopify and for ecommerce stores. And this is how my journey in Shopify apps and ecommerce tools started probably like seven years ago. And over the years, I've built a couple of products. I've built a company called checkout x, which was one of the biggest alternative checkouts for Shopify. Then I built another company called Vanguard AI, which was doing automated upselling based on algorithms and this kind of stuff, which I sold. And now I'm working on zip chat AI, which is basically AI chat for your store. And I'm really excited about it. Because now we have the technology to actually do something that you were never able to do before on a store. And that is talk to your customers. Like if you go back to like the physical stores, when you're in the store, you know, like customers come in, they ask around, they have questions, you can talk to them. And you can understand your business and your customers and all these kinds of things. We're on an online store, you just seen some numbers on our analytics dashboard somewhere, or you have no idea like you only see a conversion rate. But that's no context. You don't know why people left like, why people bought this instead of this or like if they had any questions or if your website is broken, or these kinds of stuff. And here's HipChat can do that as code for your zip chat can talk to any of your customers real time in a human like conversation, help them out. Recommend products, you know, push them to buy products, and also give you insights on what was wrong with your business was great about your business, all this kind of stuff.
04:11
So is it? Is it essentially kind of giving you insight into the business by pulling data out of let's say Shopify, or is this more like customer facing? And it's kind of just providing them insight into the product line? Like, how's Where's all that data coming from? How's that aspect work?
04:30
Yeah, zip child is a child, right? So like, it's a chat bubble, like any other chat bubble, and you can click it and talk to the website. Basically. It's simple as that. But the magic of this is, it's like a blank slate for every customer. So they can ask anything. They can say anything. And this is where the data comes in. So it's not data that you have on the dashboard of Shopify, it's data that customers will willingly provide by themselves like zero or party data is like, hey, is this product made in the US or right? Hey, like, how do you ship somewhere? Or hey, like I have this problem, or this kind of stuff. So these are all kind of like meta data that you would learn by talking on the phone with your customers that you cannot measure on, on analytics, or like on the CRM or somewhere else somewhere else.
05:24
Okay, so let's, let's take one of the examples you just said, right? Like, someone asked me like, hey, is this product made in the US? What if that information isn't on the website? Anywhere? What happens? Like how does the zip chat actually, does it? Like kind of alert you that someone's asking a question doesn't know the answer to or does it? Guess like, how does that work?
05:45
Yeah, so the difference between chat GPT, right, and zip chat is that zip chat is trained to first it sandboxed under information about your brand. And it's also trying to be a little bit like a sales clerk, right? So what is going to do is, is if he doesn't have the information, obviously, it's going to tell you, I don't have any information, it's not going to push it to you. But it's going to spin it in a way where it's going to be like, hey, I can like transfer this to our colleagues or you can contact us in this way. Yeah, they are they are. And what's really special about zip chat is that you can train it in a way that you train your employees, right. So if so many information is not available on the on the store, then what you can do is you can go and provide corrections. And you can say, Oh, well, you shouldn't answer this question like this, you should answer like this. And as HipChat is going to remember that and is going to replicate it for all the further conversations on top of that. And what we've built is really cool is the ability for your own customers to rate their replies. So you can see which replies are really great. And you can also see which replies are bad. And you can focus on those and like clearing them out and say, Okay, this is what the response should look like, all this kind of stuff. Wow.
07:10
So that's very interesting. So the, one of the things I'm thinking here, so it's pulling in all this data, it's talking to your customers, it's obviously speaking to them, as you know, kind of like as we a lot of people would know, is chat CBT. Or it's very, like, you know, sounds like a person for the most part. I immediately when I think of like, chat bots on websites, sometimes I think of like, like, let's say many chat, right, where it's just kind of like an if then statement, they select a button. And then it kind of just takes them down a rabbit hole. Can you explain the differences between that? And then obviously, in this scenario, where it's it sounds more like a person and it's actually AI as opposed to if then statements?
07:53
Yeah, so this is why we don't want to say chatbots at all, because people imagine like the generation of chatbots, where it was basically an interactive, you know, flow. And the problem with that is, you know, as if chat works great for the longtail, right? So visually, if you open a website, you should be able to see like the most important thing. So like the main question you have, but then you have a longtail of potential questions that are different for every customer. And you cannot build that into the main user experience of the website. And people don't want to search, right. So maybe you have an FAQ page somewhere or you have like a delivery page. We're not going to browse around like they they see a product like they click on an ad they see a product. And they say, Okay, I want this. And then I have one or two questions that are specific to my case, right? It can be a product question can be a geographical question, it can be something else. So the difference is you can ask anything, and you can have a free flowing conversation. And that's very different than what you used to have with the past generation of chatbots, where you had to click and you would get lost. Basically, it was a really low throughput way to interact with, with the brand, where once you've chatted, you can again, ask specifically what you want. So I think that's the main difference between many chat maybe I don't know what they're doing right now. Maybe they're also moving into that direction.
09:24
All right, cool. So it's pretty So would you say so you mentioned it kind of acts as like a little bit of a sales representative. It's also kind of on the customer service side? Do you find most people are using this from a customer service perspective? Or are they using it to kind of help from like an upsell side of things when someone's kind of interacting with it?
09:45
You know, originally everybody thinks we're doing customer support, but 75% of our customers say they want Zippy chat to increase their sales. And what we're seeing is that you're still going to have more knew most of the tickets besides maybe like, where is my order? Because we can look into the orders? And like, answer that for you. But many of the support tickets, they're like they require someone to figure out what happened, right? If everything went just as planned, then it wouldn't. There wouldn't be a conversation. But once you've chat does, it actually increases the amount of conversation you're you can have with your customers, so you have way more brands. And what we see is that customers that talk to our chat are three times more likely to convert compared to the ones that don't. So the ones that ask questions they, they usually, like, first, it's because of the chairs, right? Because they get us your questions. But second, like the people that have questions actually are interested in the product. So like, a little bit for them to be covered. And these are like the most high value leads you can have on your website.
10:58
That definitely makes a lot of sense. Because I mean, you would think someone who's interactive, they're asking additional questions, there's a good chance they're gonna be most likely to respond and most likely to convert. So from a perspective of that, leaning more towards the sales, the upsell side, there is still obviously I'm sure, like a customer service element aspect to it, like you just mentioned, like, Hey, where's my order? That kind of stuff? What? What's the thought on how that affects customer service teams? Like how are customer service teams leveraging this? Is it kind of like they're the ones monitoring how it's responding, and then they just help as needed?
11:36
Yeah, we try to work together with customer service teams. Because again, like if it's a more complex issue, we forward the request to them, we actually have this forward requests functionality. Where is it, I'm up to here, give me your email, to like the real support, basically escalating to level two. So the way we work with them is we give them more time to actually take care of actual cases, instead of responding to like someone being lazy, right. Also, like we like some, it depends on the brand, right? We have some brands that they don't just do support the sales, right? So they want the brand to call them, the customer to call them so we can program the chat. And you can pre program it with words to be like a little bit trying to catch you to give you a phone number or like to schedule a call or this kind of stuff. This is what we do on our website, once you start talking to the chat, we don't want it to be two years for like restricted in a way to push you to book a demo with us because we want to like talk to our sales agent and to us. So it's some kind of you can imagine that like us as a hype man, like somewhere that's gonna bring you in the store. One or more two things is either directly to the product or is going to push it to the sales person that's going to actually do the sale, it really depends on the case. But because you can program it with words, you can really have different behavior based on the brand and the business process behind it.
13:19
So being able to program it was certain words, so like, kind of to that point. I know a lot of people who are just like, I don't want to talk to a robot, give me a person, that kind of thing. I know through chat, and specifically for AI stuff, it's very difficult to kind of tell who is you know, who's a robot who's not. But how does that kind of get approached? How do you do that get escalated to like, Oh, this is a person that, you know, or I want to escalate this to a person as opposed to dealing with the chat.
13:49
So first of all, majority of the people don't understand they're talking to the chat to chat AI instead of the support agent. Like we see so many messages like oh my god, how are you answering so fast at this time of day? Like you guys are the best brand is hilarious, right? But it really depends is a brand like how you want to position the chat. So you can say, Okay, I'm a chat assistant, I can transfer you to a human or you can make it to be more human like to act like as a level one representative. And the binding on that is going to answer in a way where it's saying, Okay, I don't have information about this. Let me follow your request, or please call this number. So you can get this done. Right. So it's very similar, like bigger brands, they do the same thing, right? They have level one and they're like, I don't know, like, let me escalate that to level two, level three support agent.
14:45
AI is crazy. It's blowing my mind on the amount of stuff that like can come. It's like all the things that are coming out of AI like, obviously you created this awesome app. Where do you think, where are we headed? Right? Like where are you? You heard it was HipChat, where do you think AI is headed? Like at this point? I know it's not, you know, if then statements which some people, they'll slap an AI label on it. And it's just an if then statement, but like, for, for where AI is headed between your company's HipChat and where you think it's going in the future, like, what are your thoughts on that?
15:19
For first, I forget to do like a super bold statement. So you can laugh at me, like in two years, but I think we can challenge Klaviyo in converse, like in marketing automation, because the way you do marketing automation is super old school. And now you can actually have a real conversation, unique conversation with each one of your customers. And this is just the game, not only on the website, but before purchase after purchase different channels, I don't think people in the future will be sending like email campaigns, the way there is going to be tailored messages, probably direct messages at the right time for the right customer. And I think we are in a good position to, to push into that direction. For example, something we're trying to figure out. And we're experimenting is it is when you're talking to the chat, like how can we send real time events to some kind of marketing automation machine? And new type of events that are not possible right now, for example, product looking at this email, or this phone numbers looking for a specific type of product? Or he's looking for a discount? Or is I don't like, needs a little bit of pushing new type of events that are not possible, because the current first event you get is like abandoned cart, right? Yeah. But yeah, like people, people are not going to shop like they used to shop before. People will shopping conversation, they will want more tailored messaging, they would want more tailored products, and the brands would be dynamic, they won't be static as they are right now.
16:59
So it's very, you're right, that is very interesting. So in terms of challenging, challenging Klaviyo, so you're thinking like, it's going to get to a point where emails will get triggered automatically based on things that we can't even, we can't even picture because it's not a property within Play via or SMS is going to get triggered. Because the AI knows that not only is this person looking around, but maybe they need a little bit more of a push. And that's not really a property that you can put within like clay vo or something. So it's gonna take some of the nuanced information, and then be able to automatically kick out like email campaigns SMS, that kind of stuff.
17:41
Exactly. Yeah. So nuance is a very good word, you know, like generative AI in general, it can capture nuance, and this will unlock a whole new universe of possibilities.
17:55
That's, that's crazy. So. So, I mean, I would imagine one of the issues with that is like, how do you keep that from like, over communicating with them, right? Because I know, like, marketers are notorious for ruining everything, I'm, I'm one of them, I get it, I know it. So like, if I had the power of AI to just be like, Alright, I'm gonna hook this thing up, I'm gonna teach it what I wanted to do, give it some of the brand guidelines, that kind of stuff. And it's just a kick out emails and SMS and all this fun stuff as I need it. Marketers are just gonna be like, do it all the time until it like just consistently message them until they're ready to purchase like that kind of thing. But how are you going to keep that from getting to over the top where you're just over communicating with someone?
18:45
Honestly, I don't know, look, like we're very early, we're still an AI chatbot for your store. So the first recession, we're having anything many other solution in our vertical will have is, this is not a support tool. This is a marketing tool. Right? Yeah. And then, like, how can we do marketing automation with that? And once we figure it out, then we're going to figure out okay, how do we not get you know, don't don't overdo it. But I think more power to the brands like with Klaviyo you can spam as much as you want but like people are gonna hate your brand afterwards. And you're gonna get you know, flagged as spam and all this kind of stuff. Yeah, so I'm really I don't know how marketers are going to approach that our goal is to figure out new tools. So you and like all the other marketers out there can experiment with and at the end, create not only sell more, but create cooler brand experience, right? Because at the end of the day, we're in the era where branding is everything people buy the story people like a brand they pick a team when they pick a brand. And this is why they buy your probably overpriced product because like you In the world of T one Amazon, almost everything is overpriced, right? Yeah. So like the people why the why The reason people buy more buy more expensive products is because they're either they have higher quality and a better brand, and the brother brand story and we want to help you elevate that. Awesome. Obviously
20:20
you're having all these conversations with, you know, different customers it's giving you obviously, I know there's like transcripts and all that basic stuff, but like it What about from a reporting side? Is there any reporting that's supplied, that you know, gives deeper insight into like, I don't know, customer sentiment or something like that.
20:38
So right now what you can use, you can, like see all your conversations, you can see what kind of the conversations had like bad replies, good replies, you can see where your customers are starting the chats from. So you can see that you can see what percentage of your orders in which orders are in particular, how to chart conversation, like, see a map of what people talked about, and what they bought and all this kind of stuff. And something else we're cooking is actually reporting on like the nuance and like, basically data mining your own stuff, because right now you can look at your child's, but the next level is to get like the synthesis, like, fix this in this on your website, and you're gonna sell more of these to your marketing message, you're gonna sell more these kinds of stuff.
21:28
It's pretty crazy. That's awesome. That's gonna be very excited to see where AI goes over the next couple years, especially if you think it's gonna go where it goes. So in two years, if we're doing that, I'm gonna, I'm gonna reach out to me like you were right. So now I want to know what else you're thinking. But, Rosalind, thank you so much for being on the show. I don't want to take up too much of your time. I know you're super busy. I'd love to give you the opportunity here to let everyone know where they can find out more about you and of course more about zip chat.
21:56
You guys can find ZipChat on the Shopify app store if you're running Shopify, or you can find it on zipchat.ai And you're going to find all the information that's needed there. If you want to learn more about me, you can just type my name into Google or perplexity or ChatGPT. And you're gonna find my blog I write about my journey there might be interesting. And yeah, perfect.
22:24
Ruslan, thank you so much for being on the show everyone who tuned in of course, thank you as well please make sure you do the usual rate review, subscribe all that fun stuff on whichever podcast platform you prefer or head over to theecommshow.com to check out all of our previous episodes. And as usual, I appreciate you all joining we'll see you all next time.
22:43
Thank you for tuning in to The E-Comm Show head over to theecommshow.com to subscribe on your favorite podcast platform or on the BlueTuskr YouTube channel. The E-Comm Show is brought to you by BlueTusker, a full service digital marketing company specifically for e-commerce sellers looking to accelerate their growth. Go to bluetuskr.com Now for more information. Make sure to tune in next week for another amazing episode of The E-Comm Show.
Community for E-commerce Sellers
Leave a Reply