Decoding digital body language in e-commerce

# Transcript Episode 2 of the Trusted Cofounder Podcast with Victor Kuwandira, founder of Inzwa, an AI engagement widget that reads digital body language on e-commerce storefronts to recover customers between the ad click and the checkout. Recorded May 2026. Speakers are inferred from context; minor filler ("yeah", "like", "kind of" when not load-bearing) has been trimmed for readability. Proper nouns (Inzwa, Kuwandira, Harare, Shanghai, Ylikiiminki are not in this episode but the spelling convention follows episode 1) are passed through manually. --- ## Jason Welcome to the Trusted Cofounder Podcast. I'm your host, Jason, and I'm here with Victor. Mind you, I have difficulty pronouncing his last name. Kuwandira. That's right. So, Victor Kuwandira. I have him here to talk about his idea that he's working on. It's not quite a company yet, but you've been playing around with it for a while. Let's start even before the company. I want to go deeper. I'm putting you a little on the spot here, but I want to talk about your family first. Tell me about your family, your background, and then let's lead into why you make the product you have today. ## Victor Thank you, Jason, and thank you for having me here. I'm Victor, originally from Zimbabwe. I come from a family of my mother, my sister, and I. My father passed away when I was pretty young. I did my education in Zimbabwe. I pretty much grew up on a farm. ## Jason We talked about this before, that I have this farming background as well. But you've done quite a variety of different farming. What kind was it? Crops? ## Victor Crops and livestock. Goats, cows, sheep, whatever you can name. Some hunting, some fishing as well. I was still young, though. I lived there up until I was about ten years old. Then we moved to the city. ## Jason What city was it in Zimbabwe? ## Victor Harare. That's the capital city. ## Jason Man, these names are so good. I love them. ## Victor Then I did my high school. In Zimbabwe, we basically have the ordinary level, which takes four years. ## Jason Primary school kind of thing. ## Victor Then you have the advanced level, which takes two years. ## Jason What age range is that, the advanced stuff? ## Victor Let me see if I remember. I think this is when you're a teenager. I remember I finished when I was maybe 17 or 18. Somewhere there. When you take the advanced level courses, you can specialise in commercials, in terms of subjects you can study. Or you can specialise in sciences or arts. For me, I kind of did both, because I had mathematics, then economics, and also business studies. Of course, this was still in high school. ## Jason That's already in high school, learning about business and economics. That's actually pretty cool. ## Victor It helped a lot. After that, I moved to China. ## Jason That's where Shanghai is. You mentioned this China thing, I totally forgot about it. So what part of China was this? ## Victor Shanghai. ## Jason Is that the capital? ## Victor No, the capital city is Beijing. But Shanghai is the commercial city. ## Jason 28 million people. Like everybody I know in China comes from Shanghai. It's so big. ## Victor It's pretty big. So I was there. I did my bachelor's degree in software engineering. While I was still there, I also started my own company back home. I was basically exporting smart monitoring systems, things like CCTVs, smart door locks. Stuff like that. ## Jason That's so cool. So you took this basically e-commerce and business knowledge from when you were young. But you also picked up a degree in computer programming, basically. ## Victor Right, right. ## Jason That's a great combination to have. I think one thing I strongly believe in is this trifecta of skill sets, and that is the business aspect, the other one is the technical programming, and the last one is design. Now, if you're telling me you can draw really well, then we have an amazing trifecta here. ## Victor I mean, I'm learning UI/UX design, but that's a bit… ## Jason Okay. So after starting this company and so forth, where did you go next? What happened? ## Victor That's when I moved to Finland. Right after that. And, because we're going to connect this probably later on, when I was doing my company in Zimbabwe, I pretty much learned a lot in terms of business. I was selling online, even though I also had a physical store in Zimbabwe, and I was also exporting to South Africa and Zambia. ## Jason So you're already playing with the e-commerce kind of solutions. ## Victor Exactly. I was playing with e-commerce there. So then I came to Finland. ## Jason During this time, you've been studying what? You're a student now, I assume? ## Victor I graduated last year in August. I was doing my master's in information systems and software engineering. ## Jason So you continue with the software engineering and the software development here as a master's degree. And then now, if I remember right, you got started with the Startup Express business accelerator. Tell me about that. ## Victor The Startup Express program was actually really good for me. I needed to sort of… because I had spent quite some time coding and engineering. So you start to lose that. ## Jason You're deep in that world. ## Victor Exactly. So the Startup Express, with the mentoring team there, we got to learn what it means to start a startup. Your idea, hey, there's something called problem validation. You've got to talk to people. So I had to learn, and also unlearn, certain things. ## Jason I feel the same way. I studied computers and computer programming. But then when I started building business ideas, I realised I was way too technical in trying to solve the thing, build it just straight up. And what I needed was the business sense. Are people even willing to buy this product? Are they willing to use it? I see where you're going with Startup Express. It was giving you that kind of perspective on the outside. Is that right? ## Victor Exactly. Exactly. ## Jason Okay, now how long was the accelerator? My guess is there's a classic problem-solution situation. So how long was the accelerator, and then what was the problem that you identified that you wanted to solve there? ## Victor The program lasted three months, starting from January. Because I had run an e-commerce store, the problem was clear to me. But I still needed to validate it. The problem was that every e-commerce store spends quite a lot of money on ads, trying to bring traffic to their website. But when people get there, because of different hurdles or different dead ends that they meet, they end up leaving the website without purchasing. ## Jason I mean, we all do that. We go to a website, browse around, and then we leave. And you get the classic stuff where, if you're just about to leave and your mouse leaves the port, suddenly a pop-up shows up. Hey, 10% off, blah blah blah. They all have this thing of trying to catch you. So here's this problem. You had the name of the product. It was Inzwa? You made this product called Inzwa. Tell me, how does it solve this problem? ## Victor We have a widget that sits on your storefront that can proactively engage your customers and answer their questions. It can communicate in natural language. And when that conversation happens, we analyse it, of course, and then surface it to the merchant or to the store owner. Like, hey, you know what, today you had 10 customers come to your website. You were able to engage five. But the problem is they didn't buy because you had a missing variant. They needed a specific size and a colour. Or maybe this one, they were so confused about your return policy, so they ended up not purchasing. ## Jason So instead of focusing on the inbound, that's something a lot of places are doing. A lot of people are saying we'll improve your SEO, that's one basic thing. Or we'll set up a chatbot to find a product that might be interesting to the customer. You're actually focusing on the fallout. Right? So you're somehow identifying when that customer looks a little lost. What shows up to interact with them? Is it a chat? What happens there? ## Victor So we talked about pop-ups, right? There's a really negative sentiment towards pop-ups. But if you think about it, I've been trying to think about the psychology behind why we hate pop-ups, and there's a few reasons. Number one, it shows up at the wrong time. I just got on the website and bam, right there. And it's irrelevant to what I'm doing. ## Jason You haven't even browsed anything. Sign up for 10% off. ## Victor It's not that I hate signing up for that 10%, but it's like, hey, I just got here. I need to understand what is really going on. ## Jason I have the funniest experience with that. It's to the point now where I deny it right away. I don't want to put my email in there and subscribe when I find them. And if I do find it interesting, I go to another network. Either I use my VPN, or I literally go to a café and remember to log in to the same site, and then the pop-up will show up again at that point and I sign up. It's so silly. These pop-ups really get in the way. But you said the timing was important. How do you determine that? ## Victor And secondly, it's intrusive in a way that it blocks the entire screen. If you think about it, when you go to a physical store, a good salesperson is not going to come to you, run into you, and stand in your way like, hey, we got this 10%. But why do we do that online? It's super annoying. So we are using the same technology, but in a different way, in the sense that we take what we call digital body language. The user has maybe searched for something and it shows zero results. We can put a pop-up that doesn't block the entire screen, but a simple push, like a simple nudge, a toast, a notification, whatever you're going to call it. It's clickable and it has context of what's happening. For example, right now we use search bars and not all of them understand semantic analysis. Where I'm searching for cool sneakers for something, it just shows zero results. ## Jason Right. ## Victor But because we have an AI behind, a system that understands natural language, it shows the right information. Like, hey, are you looking for sneakers? Let me quickly find those for you. If you click, it opens the chat, and already the products are rendered, the product cards are displayed, and you can see them. Oh, okay, so this is what's available. That's one example. Another example: maybe you're on a product page and you select something that is out of stock. It can push a small message that doesn't block you from whatever you're doing. It shows you, hey, that variant is out of stock. Would you like me to show you a close alternative or set a restock alert? ## Jason Relating to the physical world, you made that body language comment. And I totally agree with you. I was somebody who really liked doing good customer service. It was fun when I'd work at an office supply store or whatever. I'm that guy who sees that the person is looking at cameras. The digital cameras. And I'm like, oh, I should go and help them out. They've been there a while now. Maybe I should go talk to them. You're making the digital equivalent and trying to identify the quote-unquote body language and build on that experience. Now, let's say somebody is looking for. I have ginormous clown feet, so it's hard for me to find shoes in my size, but it's easy to find shoes generally. So how does your tool help solve that aspect, that when I can't find the shoes, can I somehow send them a message? Or how does it work for your tool? ## Victor Exactly. So if you cannot find your variant, especially if you're already within the chat, it will be able to notify the store owner. You can request an item, because there's already the set-up where you are able to request like, okay, this is what I'm looking for. And the good part is that even though you don't necessarily request it as the customer, the store owner will be getting a summary of that conversation with the most important information. Because I also understand that founders are so busy, if they have a lot of traffic that they don't have time to review each and every conversation. ## Jason You need to somehow categorise and build up that data. That's cool. I'm thinking about there's different levels of fallout. There's the toaster pop-up. I'm thinking about a physical person rising from behind a shelf. But yours, it shows a little pop-up and says, hey, can we help you out? And the person still drops out anyway. Does your tool take that data into account in terms of the fallout, or is it the conversation that you're building off of? Where does it interact with the customer, and how does the data go to the e-commerce owner? ## Victor Right now, because I'm trying to focus on explicit demand, on explicit data. Because you see the industry right now, what we have in terms of analytics tools, it's all implicit data, like those heat maps. They can show you that, yeah, 10 customers moved from here, from there, from there. But it doesn't show you why that was happening. Were they looking for something? Was it too expensive? The body language, right? We can see that. So you can understand that I'm really reusing the same technology, which is, at this point, everybody's making assumptions. We are all assuming that the customer left because maybe of the price, or because they were looking for something. I am also assuming that when I put that pop-up, but with better context. And then I'm validating when they actually engage the nudge, because it has the right context. We get the conversation and we put it to the store owner. Because when you look at it right now with Google Analytics or Hotjar or any other tools, they're showing you what happened, not why it happened. ## Jason They're showing the positive path rather than fallout. ## Victor Exactly. So I'm trying to put the complete picture. This is what happened, and this is why it happened. ## Jason Well, that's really cool. So what are your next steps? Are you developing this tool right now? ## Victor It's actually interesting because today I onboarded my very first customer. A pilot merchant. Actually, that was like an hour ago. ## Jason Oh, you've got to have it ready for the podcast. ## Victor We had been talking. I'm super busy because I built the MVP, I sent there. Right now it's still in a closed beta. I only want to work with five very specific businesses. Maybe that's the wrong way of doing it as a founder. I don't know, man. I'm just figuring out things. ## Jason Really fantastic. ## Victor The reason why I'm doing it this way, whether it's wrong or not, I'm okay with learning and unlearning things. I want to understand my product before I can do a lot more, to actually see what is it that they need, what is it that actually matters to them. And this is why I'm only going to work with, for a start, five very specific brands. This is another thing. I don't know if I'm starting to talk too much. ## Jason No, no, please. ## Victor My core philosophy is that Inzwa can only serve merchants as a byproduct of serving their customers really well. ## Jason Okay. ## Victor I cannot provide any value, or Inzwa cannot provide any value, to store owners if I don't understand their customers or their customer behaviour. So this is the thing. As I'm looking for merchants, I'm not just looking at, oh, this business does this. I'm looking at what do their typical customers face in terms of problems. ## Jason That takes some careful stepping. You have to figure out what niche and so forth. Because you were mentioning that you're worried whether it should be lots of people you're talking to, or just a couple of people, in terms of understanding the customer and understanding your sales process. I think it really depends on what kind of company you are. But in your case, one philosophy I like to live by is that if you're learning from the customer you're talking to, if you're learning a lot, then keep going with that. But because sometimes you'll get this wave of information, too many people, and you end up losing them as potential customers, especially with B2B kind of stuff. Those people are already kind of burned through and you have to recycle them and try to do a sale again later. And if you don't approach it right, you can really burn them. So I think that's, am I right that that's kind of like where your approach is? You can make a mistake with one customer, but at the same time you give absolute time and commitment to that customer, to help understand not just them, but the people that flow into their e-commerce website. ## Victor Exactly. Because the widget, which is on the storefront, if it's not helpful to their customers, then I don't have any insights to give them as the business. So which means I need to be able to find those types of businesses where there's a lot of friction, there's a lot of hesitation, there's a lot of confusion, where the return policies are such a hassle. For example, think about products that are super technical on the website. Sometimes you're reading those specs, but you're like, man, I don't understand what is this. I've got to take a screenshot and maybe go to ChatGPT and maybe I'll come back or not. So that's the whole idea. Okay, I can put my widget there because it would be helpful and be able to digest any information. It's a large language model. That's what they're really good at. You can feed them a bunch of information and they can spit it out in a very simple way the customer can understand. ## Jason Well, I'm excited to see where this product goes. Do you have a basic website up now, or are you just working with the customer behind the scenes? What's your situation? ## Victor I do have a landing page, inzwa.co. ## Jason So that's I-N-Z-W-A dot C-O. ## Victor It's available in Finnish and in English. ## Jason So we can just go to the site, learn more about the product. Is your contact information there? ## Victor Everything's there. If you're a business looking to collaborate with me, you can fill out your information there in a form. ## Jason Excellent. Victor, thank you so much for your time. I appreciate your comments and willingness to work with me here, especially with the podcast being so young. It helps a lot to have good feedback, and I think you did that. Thanks so much, Victor. ## Victor Thank you so much.