Guest: Shauli Mizrahi
Ever wondered why some customers stick around for years, while others disappear after just one purchase? Shauli Mizrahi, CTO and Co-Founder of Rep AI, has cracked the code. His AI-powered sales assistant technology is helping eCommerce brands achieve 5x ROI guarantees whilst transforming casual browsers into loyal customers. The secret lies in recreating that personal touch you get in brick-and-mortar stores—digitally.
Shauli's journey began five years ago when he recognised a fundamental gap in online shopping. His company, Rep AI, has developed behavioural AI that predicts the exact moment a shopper needs assistance, then steps in with personalised recommendations and support. Unlike traditional chatbots focused on customer service, this technology is specifically designed to drive sales, increase conversion rates, and boost average order values through intelligent and empathetic interactions.
The Abandoned Shopping Cart Epidemic
Before exploring solutions, we must first understand the psychological disconnect between online and offline shopping experiences.
"We love the experience of brick and mortar stores. How you go into a brick and mortar store, get concierge help, white glove support, someone helping you with whatever you need, recommending products, answering any questions. And when you dive into an eCommerce online store, you're there by yourself. No one is there to help you."
This isolation costs eCommerce businesses millions.
Customers arrive with intent but leave without purchasing because they lack the guidance and reassurance they'd receive in physical stores. The result? Cart abandonment rates are hovering around 70% across industries.
Research shows that shoppers crave two things traditional websites can't provide: perceived empathy and non-judgmental assistance. People feel more comfortable disclosing personal information—like sizing concerns or budget constraints—to AI than to human sales assistants, viewing AI as less judgmental and more anonymous.
The AI Sales Assistant Framework
Shauli's approach differs fundamentally from conventional chatbot implementations through what we might call the Behavioural Intelligence Framework:
Behavioural Tracking - The system monitors customer actions from the moment they enter the site, building detailed profiles based on browsing patterns, page visits, and engagement signals.
Predictive Intervention - Rather than waiting for customers to initiate contact, the AI evaluates whether each visitor is likely to purchase independently or requires assistance, proactively approaching those who are at risk of abandoning their journey.
Contextual Engagement - When the AI does intervene, it references specific products viewed, previous purchase history, and browsing behaviour to create relevant, personalised conversations that feel natural rather than scripted.
Comprehensive Knowledge Base - The system ingests entire product catalogues, website content, and purchase histories to answer detailed questions about specifications, compatibility, and recommendations with the depth of an expert sales associate.
One client experienced a transformation from a 3.5-month to a 15-month customer retention rate using this approach. The AI doesn't just answer questions—it builds relationships through understanding and empathy.
The Trust Evolution
The most fascinating aspect of AI sales assistants is how quickly human behaviour has adapted to them.
"We definitely saw a huge transformation along the years. We started in 2020 where we went to customers and said, hey, what about the idea of having your AI on your website? And a lot of the response we got was, oh, AI is never going to be a thing."
Fast-forward to today, and the same customers are sharing life stories with AI assistants, discussing sensitive topics they'd never mention to human sales staff. This shift represents more than technological acceptance—it reveals fundamental changes in how we perceive and interact with digital intelligence.
The evolution happened in stages: initial distrust, then treating AI like a search bar with single-word queries, and finally embracing full conversational commerce. Now, customers actively seek AI guidance for complex purchasing decisions, particularly for sensitive products where anonymity provides comfort.
Beyond Customer Support
The most significant mistake eCommerce businesses make is categorising AI chat as purely a customer support function.
"A lot of brands, we offer that to them or we have a conversation, a lot of marketers, head of e-coms will say, hey, OK, that's chat. Chat is not my business, Chat belongs to the CX department, the customer support department."
This misunderstanding costs revenue. AI sales assistants should be owned by whoever handles conversion optimisation and sales strategy, not customer service teams. The technology excels at:
- Product Education: Explaining complex features and benefits in detail
- Recommendation Engines: Suggesting complementary items and upsells
- Objection Handling: Addressing concerns that prevent purchase completion
- Choice Simplification: Helping overwhelmed customers narrow down options
Consider the difference between a customer asking "What's my order status?" versus "I need a gift for my wife who loves gin but I don't know anything about spirits." The first is customer service; the second is a sales opportunity requiring expertise, empathy, and product knowledge.
Implementation Without Risk
Getting started doesn't require massive datasets or technical expertise. Modern AI sales assistants integrate with existing eCommerce platforms through single-click installations, automatically ingesting product catalogues and customer histories.
"To get going, you don't need to have nothing basically. Just your own website. That's it."
The system becomes smarter through usage. Rep AI typically sees 3% of website visitors engage with their AI assistant, providing continuous learning opportunities. Each interaction teaches the system more about customer preferences, pain points, and successful conversion strategies.
Success isn't limited to extensive catalogues either. Single-product sites often benefit more than multi-product stores because expensive or complex items require more explanation and convincing—exactly what AI sales assistants excel at providing.
The Conversational Commerce Revolution
We're approaching a fundamental shift in how people discover and purchase products online.
Traditional search is migrating toward conversational AI platforms, such as ChatGPT, Claude, and Perplexity. Soon, customers won't visit websites directly—they'll describe their needs to AI assistants, which will research and recommend products on their behalf.
Forward-thinking brands are preparing for this reality through the integration of Model Context Protocol (MCP), enabling their AI to communicate directly with external AI agents. When someone asks ChatGPT, "Find me the best running shoes for flat feet under £200," brands with proper AI representation will be included in those recommendations.
Taking Action
The question isn't whether AI sales assistants work—it's whether you'll implement them before your competitors do.
Start by auditing your current customer journey. Where do people get stuck? What questions repeatedly appear in support tickets? Which products require the most explanation? These pain points represent opportunities for AI intervention.
Consider the mobile experience particularly. Desktop browsing allows customers to see multiple products simultaneously, but mobile users view one item at a time, making choice paralysis more likely. AI assistants excel at providing mobile guidance by asking qualifying questions to narrow down options efficiently.
Test with low-risk implementation. Many platforms offer free trials and ROI guarantees, removing financial barriers to experimentation. Track metrics beyond conversation volume—focus on conversion rates, average order values, and customer lifetime value improvements.
The transformation from 2020's AI scepticism to today's widespread adoption happened faster than anyone predicted. Don't be late to this particular party.
Today's Guest
Name: Shauli Mizrahi
Title: CTO and Co-Founder
Company: Rep AI
Bio: Experienced technology professional with expertise in AI-driven eCommerce solutions and behavioural prediction systems. Previously served as VP of R&D at Browsi, developing AI-powered advertising optimisation tools.
Contact: https://hellorep.ai/ | https://www.linkedin.com/in/shaulimizrachy/
Blog Summary
Discover how AI sales assistants are revolutionising eCommerce by recreating the personal touch of brick-and-mortar shopping online. Shauli Mizrahi reveals how behavioural AI can predict when customers need help, resulting in 5x ROI improvements. Learn why treating AI as sales tools rather than customer support creates measurable revenue increases, and explore the future of conversational commerce where customers shop through AI agents rather than traditional websites.
Resources
eCommerce Podcast Ecosystem
- eCommerce Cohort signup (free monthly calls for eCommerce entrepreneurs) - https://www.ecommerce-podcast.com/cohort
- Matt's LinkedIn Profile - https://www.linkedin.com/in/mattedmundson/
- The eCommerce Podcast Newsletter for show notes and extra insights - https://www.ecommerce-podcast.com/subscribe
Links for Shauli
[00:00:00]
Matt Edmundson: Well, hello, my name is Matt Edmundson and you are listening to the e-Commerce podcast. Yes, you are. Uh, I've been in e-commerce since 2002, and these days I partner with e-commerce brands to help them grow, scale and exit. And if you'd like to know more about how that works, uh, and if we could actually work together, head over to the website ecommercepodcast.net, where you'll find all kinds of information about that.
And also. All kinds of information about the show, all the past episodes, the hundreds of them that we've been doing this, now it's like six years. Can you believe it? Uh, so there's a lot of cool content on there. Uh, so you can check that out at ecommercepodcast.net. Uh, it'd be great to see you on that site.
So do come join us now today. I'm talking to Shauli all the way from New York who has braved the ills, uh, and swallowed some modern medicine and forced himself to come onto the show. Uh, what a trooper [00:01:00] this, uh, chap is, uh, Shauli. Welcome to the show, man. Great to have you. Uh, I would ask you how you're doing, but it's, uh, it's, it, it, it's, it's kind of okay.
Shauli Mizrahi: Yeah, exactly. Getting better. Getting better. Uh, thanks for having me. It's uh, it's a retreat. Um, real honor. Uh, so thank you.
Matt Edmundson: No, no, it's great to be here. Now. We were, we were talking, uh, Shauli before we hit the record button. Uh oh. By the way, ladies and gentlemen, lemme just say, uh, before we get into this, if you, um, if you are watching this, uh, on YouTube, you are gonna go, Matt, you're not wearing your usual clothing. I'm in fact wearing, uh, a Kansas City Chief's.
Super Bowl winning Jersey. And uh, is that because I'm a Kansas City Chiefs fan? Well, yeah, why not? You know, living in England, we like American football. Uh, but the, a good friend of the show, he's been on the show before and a really good friend of mine now, Brett Curry, um, from OMG Commerce sent me this and it arrived this morning.
He sent one for me and one for my daughter. So we have matching Kansas City. [00:02:00] Chief's football chose, so I thought I'd wear it in honor of him. And just give a brief shout out. Brett, you are a legend. Thank you so much for my shirt. Uh, we always like Gifts and Brett sent, sent me in the one for Zoe. So yeah, thank you very much.
Anyway, that, sorry, Shauli, I just got a bit carried away there. I, I realized, uh, with the, with the screen, I'm like, oh, visually things are a little bit different today. Um, and by the way, are you a Chiefs fan? Do you do American football? Am I offending you? Wearing this top?
Shauli Mizrahi: You are not offending me, actually. Um, I'm getting into American football. Uh, so this is something I'm, you know, always been a soccer fan, but, uh, soccer or English football fan. But this is, I'm doing my first steps into American football. I love, I love that. So far I'm still trying to get, understand the different teams and so on. Uh, I live in New York, so I'm, you know, more of a Giants fan here, but, uh, or like started, but uh, yeah, it's, uh, it's
fascinating actually.
Matt Edmundson: Yeah.
Shauli Mizrahi: I watched like last Super Bowl. Loved it. It was,
uh, [00:03:00] uh, I would say great time with friends and family, so it was nice.
Matt Edmundson: Yeah. Yeah, yeah, yeah, definitely. Yeah. And And can I ask what English team you support though? Actually real football. 'cause we played with our feet, but that's the next sense of question. We will, we will not get into right now.
Shauli Mizrahi: Arsenal,
Matt Edmundson: Awesome. Well it is been great having you on the show. Thanks for joining us.
Arsenal. Uh, if you don't, if you don't know, uh, already I'm a big Liverpool fan.
Shauli Mizrahi: I'm sure you're, I mean,
you.
Matt Edmundson: live in Liverpool. Yeah, yeah, yeah. So, yeah, it's, the reason why I chose to live in Liverpool was because of the football team.
Shauli Mizrahi: Right. Okay,
Matt Edmundson: So, which is, I, I thought was as good a reason as any to, to choose where to live.
Shauli Mizrahi: so you're not born, raised in Liverpool.
Matt Edmundson: no. No.
Uh, if I was born and raised in Liverpool, I'd probably be talking like deese. Exactly. Um, and so, no, I've lived here 32 years, 33 years now. Actually a long old time. Came for university and have sort of stayed. Um, but yeah. Anyway, we digress. Let's jump into it. Uh, [00:04:00] enough about football, all those, I could talk about it all day.
Um. So you are in New York. SalesRep.ai is the name of your company, so guess what they do, ladies and gentlemen, AI is in the title. Um, and we were hit talking before we hit record and before I started this whole rabbit trail, um, about how, you know, you obviously work with a whole bunch of people around the world, various different sizes of E-com businesses, and I asked you a simple question if you could solve like the, what is the key problem, you know, the common problem that everyone comes to you with.
Uh, what would that be? And Shauli, your answer was,
Shauli Mizrahi: Helping with sales, basically increasing sales. That's number one. Uh, um, number one problem that we're trying to help merchant is to increase their top line and get more revenue into, you know, into their businesses.
Um, just wanna say one correction. My, uh, the company name company's name is, uh, Rep AI. It's, hellorep.ai
uh,
Matt Edmundson: [00:05:00] what did I say?
Shauli Mizrahi: sales reps.
It's fine. It's not a big deal.
Matt Edmundson: You know what? I've even got your website right in front of me. I'm really sorry. I ladies, gentlemen, I should probably go get my glasses checked, I think is the, uh,
Shauli Mizrahi: okay. That's okay. Uh, don't ever make I'm mistake again,
Matt Edmundson: yeah, yeah. We've got your address, dude.
Yeah.
Oh wow. That's unbelievable. Sorry. hellorep.Ai, I'm, very sorry.
Shauli Mizrahi: exactly. It's, it's all good. It's all good. Uh, so yeah, the number one problem we saw. My co-founder and I five years ago, uh, you know, we were big fans of the e-commerce space. Always. Uh, there, my co-founder has been in, uh, agencies and helping, uh, retail brands forever. Um, I come from the tech side and the number one problem we saw five years going back is that we love the experience of brick and mortar stores.
How you go into a brick and mortar store, you get. [00:06:00] Concierge help, white glove support, uh, someone helping you with whatever you need, recommending products, answering any questions. And when you dive into an e-commerce online store, you're there by yourself. No one's there to help you. You know, there's images nice.
There might be, you know, personalized, but that's where you go, right? All you have are the, the search and, uh, the menu. That's it. And then we thought to ourselves, how amazing would it be if we can bring that experience from brick and mortar, like really simulate the brick and mortar experience, bring it online. And, you know, you obviously can't do that with people because it's just gonna be a lot of people doing that role for any visitor that comes in. And we said, okay, how if we can bring AI to, uh, to the, to the mix and give that, uh, give that experience and really treat the AI or train the AI to do the exact, the same stuff. As, uh, brick and mortar sales assistant would do. And that's where we came with the idea. Uh, two years in of developing the, the, uh, technology, [00:07:00] ChatGPT came on with, uh, their amazing AI and let us actually like, um, wave or ride their engine and give that experience to, uh, to merchants. So we came out with our solution in 2023 and ever since, uh, it's been amazing.
Matt Edmundson: 2023. It's only yesterday. It feels like decades ago in the world of AI because it's changed so much since 2023.
And this is really, um, is, this is sort of the philosophical question, isn't it? About what you guys do, um, and whether an e-commerce site needs it. 'cause you're right, I think about the amount of e-commerce websites I go to and I don't know them.
I have no feel for them. I. Either like their product or I don't. I'm, and one of the questions that we have wrestled with, I think in e-commerce over the years is, how do I connect with customers? So you have a lot of founder facing brands who do crazy videos or do the social media just to have that [00:08:00] connection, you know, uh, with their, their customers.
But you're right in the sense of whenever I go into a brick and mortar store. Especially actually, if that brick and mortar store is in the states, it, it, sometimes in the uk but certainly in the states, somebody's gonna come up to me and go, Hey, how can I help you today? Um, or some variant of that question, aren't they?
And if they're, if they're likable, if they're personable, if they kind of know what they're doing, they're probably gonna engage me pretty well. At some point I'm gonna buy something. A good example of this Shauli, actually, uh, last weekend I was up visiting my son in St. Andrews, um, and his fiance, uh, up at, um, the very famous, uh, golfing town in Scotland.
And I wanted to buy a gift, uh, for our host family. They like gin. So there's this really lovely kind of, uh, have you ever been to St. Andrew's by the way?
Shauli Mizrahi: No.
Matt Edmundson: No. Okay. There's this beautiful kind of whiskeys shop [00:09:00] at the end of the street, and I, I, every time I go, I, I go there and I just walked in. He said, how can I help you?
And I was like, I just need some gin. And he said, go talk to that fellow over there. And this, this guy just talked to me about gin for like 15 minutes and I ended up buying not one bottle, but two. And I tried all his variant gins and I'm a bit of a, I, I like his stuff, but that would've never happened. On an e-commerce website, right?
You kind of, you look at the label, you look at the price. Do I want it yes or no? I dunno if I do, but, um, because of that personal interaction, I think it changed my shopping experience. And I've always, the reason I go into that shop is not because it's the cheapest, but because every time I've gone in, the guys just talk to you, uh, and make you feel welcome and give you some good pro product recommendation.
So I love what you are trying to do and how you then do that on e-commerce websites. 'cause people have been trying that. In various different gise, uh, for years, and it seems like this is perhaps one [00:10:00] of the key benefits of AI for e-commerce owners.
Shauli Mizrahi: Yeah, no, totally. So, um, I think obviously it wouldn't be able to to happen without ai. So AI is the key, uh, driver for this, for this to, to happen. There's no doubt that, uh, what, what drives that higher conversion rate is the, the level of personalization that, um, that you get from the sales assistant, from the AI in the end.
So what we do is that we track, uh, the behavior of customers from the moment they, they walk into the website, the moment they log in, the moment they, they hit the, the, you know, enter through a campaign. The moment. They're, they're on the website. We start tracking everything they're doing to, to build a customer profile basically on them.
So to understand exactly what they're interested in and what happens in the background. Uh, the AI tries to evaluate if. This customer will end up buying by itself, by him himself, or herself, or if that customer needs some assistance. [00:11:00] And, and when the AI gets that signal that the, this customer will not buy the AI approaches that customer with basically very similar to what the experience you had in that, in that, uh, a liquor shop, but with the whole context of what they've, what they've been doing.
So if you were checking out gin bottle, for example. AI would know that. So the AI would come in with an approach of, Hey, we saw that you checked out this, uh, bottle of gin, and if you bought before, the AI would know that as well. And if you checked out other, uh, let's say other brands or, uh, other, uh, sorry, other products or other categories, other collections, the AI would know that too.
So everything goes in, into that customer profile that's being, uh, rebuilt. Every, you know, every interaction or everything, you're, you as a shopper, you're doing, and AI comes in and that's the basis of the conversation. AI lu you into the conversation with that proactive approach. And then from that moment on, the AI will do anything in its powers, okay?
It has a lot of powers, but anything in its powers to make sure that you end up as a happy customer. So if you need product [00:12:00] recommendations, if you need to know the history of that gin distillery, for example, and that will push you in. Uh, to, to buy that bottle or whatever, whatever is possible, because obviously the AI has all the knowledge of the store, the catalog, uh, your purchase history and, uh, other people's purchase history.
So it's able to recommend products that are, you know, complimentary, for example, or similar items. So it has all those tools, and again, the number one role is to make sure you come out as a happy customer.
Matt Edmundson: Yeah, that's fantastic. I'm, I'm intrigued by it, um, because it. I to tell you how intrigued I'm by, I, I, I, I had a slight, a slight moment, uh, Shauli earlier on when I was driving, uh, from the office back to the, the studio here. Um, and I, I knew you were coming on and I, I was, you know, I've got to the stage now where actually when we research guests, I use perplexity.
And Perplexity has this feature where it talks to you now. Um, so [00:13:00] it'll go away and it'll do a search and it'll come up with all this data. You press the headphone button and it just, it just reads it out to you, you know, and in, in quite a human manner. Uh, and you and I, one of my questions that I asked, and I I wanted to ask you this question 'cause I, I, I thought this was really fun when I realized what I'd done, uh, as I'm driving home, I went, well, but do humans actually trust ai?
Um, and this is what I said to perplexity. I've got it up here. It said, do humans trust ai? Which is more of a philosophical question, I suppose. Um, but you're talking about some big upswings in buying behavior, which we'll get into that this kind of technology brings. But what is a human response to ai? It it, it was kind of lost on me for a minute that I was having this conversation with ai, right.
Shauli Mizrahi: Yeah. You can prove the point basically.
Matt Edmundson: Yeah, yeah, yeah. And then I just, I, I got out and I realized, I thought, oh, okay, this is interesting. And because this is just what I do now, I quite happily chat to ai, you know, a significant portion of my day inhabit, read content back to [00:14:00] me. Um, or if I'm going on a walk, I'll just have a chat with ai. And it seems like.
Humanity is getting more and more accepting of the, sort of the AI features. Uh, and I I wanted to sort of start there a little bit because I know you, your company started earlier than 2023, but obviously you connected with chat GPT in 2023. How have you found people's response to this?
Shauli Mizrahi: To AI or to, to the, um, to our solution in gen, like in specific our solution or AI in general.
Matt Edmundson: Yeah, specifically to your solution. 'cause obviously that's where you've got the data. I mean, you are using AI now to drive that and I'm, I'm, I'm curious to know what, how people are responding now with the,
Shauli Mizrahi: So we definitely saw a huge transformation along the years. We started with 2020 where, you know, we went to customers and said, Hey, what about the idea of having your AI on your website? And a lot of the, a lot of the, you know, response we got was, uh, uh, oh, AI is never gonna be a thing, [00:15:00] you know? We don't,
Matt Edmundson: Along with Bitcoin, that's never gonna take off either, right?
Shauli Mizrahi: yeah, or we said, uh, and they said, okay, yeah.
How do you, you know, how do you, uh, envision that? And we try to under explain it would be a chat on the website and people said, a chat, AI chat. No way. That's, again, that's not never gonna be a thing that's a. Complete failure, you should go to voice. 'cause Alexa is the next thing, by the way. So we had a, you know, a moment where we thought, okay, should we try to have the same solution on Alexa? Um, and basically allow brands to have an app on Alexa. That was
part of the path we took. Uh, but we strongly believed in AI chats on, uh, on the websites. 'cause we, you know, we had this, um, I would say vision that AI chat would be the thing. So we saw this big transformation from merchants that are. Uh, not believing at all to merchants that, uh, all of a sudden saying, Hey, we, we would be happy to be early adopters here.
It's [00:16:00] looks cool. Might be the future. Let's try it out to obviously today that, you know, we see huge brands, enterprise brands that are coming to the realization, okay, we need AI on the website and it's okay. That it's okay if we don't control everything. We would say it's
okay letting go, because that was the biggest hurdle over the years.
Like, is it gonna say things that we never like, uh, we're not gonna dictate every word it's gonna say. And, uh, it was hard for our big brands to, to accept that it was hard for s and bs to accept at the beginning. And now we're slowly getting to the maturity of that. So acceptance from merchants. And that's on the merchant side.
From the shopper side, um, I would say shoppers also made, obviously big transformation started from. Think people thinking this is a human because there's no way the AI would speak like that. So they, they started speaking to humans. Um, and that happened a lot until a lot of our merchants, uh, a lot of our brands basically said, okay, let's write the word ai, because people get too confused and it's, and then they, they get, uh, upset when they figure out the truth. [00:17:00] So they started using the word ai and then people were treating it like, um. A Google search. Red shirt. Red shirt. That's it. And that's what we got at the beginning. So people, you know, writing one word or two words, that's it. That's the, the most that, uh, they looked at it as, as a search, um, search bar. And we seeing that the full transformation today were people are basically giving them the ai, the full life story and it's okay. All their problems and how they were fighting with their wives and, and. And the AI listens, so it's okay. And, uh, they, I don't think they would treat the same. They treat humans the same way. I don't think they would be that open
with writing everything.
And if it's, um, you know, if it's products that are, you know, a bit sensitive or sizes that are a bit sensitive, they're okay with telling the ai they're, you know, they're, they're sizes and everything. And, uh, it's been amazing to see that that full adoption, that that happened.
Matt Edmundson: It's interesting you say that because, um, [00:18:00] research here that I've got in front of me, what drives human trust in AI chatbots? Uh, there's four key areas that it says. One is, um. the technology has got better. The reliability and competence of that technology has also got better. And I think if you can connect that with integrity so you feel like actually this thing's on your side somehow you, you do well.
Um, one of the things they said, which this is the, the, this connects with the two things that you've just said. And I, I'm, this really intrigued me because. They said one of the things was perceived empathy. Chatbots that seemed to understand and respond to user emotions, foster a stronger sense of trust and engagement, and the non-judgmental nature of chatbots.
So you talked about putting your sizes in. Many people feel more comfortable disclosing personal or sensitive information to chatbots than to humans, as AI is seen as a lot less [00:19:00] judgmental and more anonymous.
Shauli Mizrahi: Yeah, yeah, yeah. Which is, uh, pretty amazing.
Matt Edmundson: Yeah. I, I thought that was really intriguing, the fact that you've mentioned that. I, I, you see, they're, they're the things that I, I, I would've gone. Yeah. Well, it's gonna get me to my end point on the user journey quicker. 'cause it's gonna, I mean, but the, these soft things, you know, like empathy and non-judgmental nature, make a big difference.
But how you quantify that, I have no idea. You know, and I, I think it's quite fascinating that this is what's coming out in the research now.
Shauli Mizrahi: Yeah, so I think it's from the, the merchant perspective, it's quantified by top line revenue. That's pretty much it. Uh, saying in the end, seeing in the end that there's more revenue, there's bigger lifetime value as well. So, uh, and seeing all that actually happening and being able to test it and AB tested as well. So being able to say, Hey, this is actually increasing my revenue. I can see that before and after. Um, and yeah, there's, uh, there's no doubt that it's [00:20:00] working.
Matt Edmundson: So if I wanted to, um, get started maybe with something like, you know, like what you guys do, let's go through some of the, the sort of the, the technical side of things, Shauli, if we can like.
Shauli Mizrahi: Yeah.
Matt Edmundson: Do I need to have like a certain data set? Do I need to be a 'cause years ago, I'd say a few years ago, years ago, like it was decades last week, last Tuesday, you needed a really big data set.
And the more data you had, the more accurate it was. 'cause that was all based on machine learning. Um. Do we still need big data sets? Is this something that would resonate for a smaller e-commerce brand? Does it work specifically for, say, clothing brands? But if you're a single product site, does it work as well?
What, what sort of things have you found work really well, um, with the, the AI system?
Shauli Mizrahi: It's a great question. I get that question all the time. So, um. People ask, Hey, is I have [00:21:00] only one product? Does that mean that, uh, I don't need it as much? Now, the the answer I would say is like, if you open the retail store, would you not have shop assistance there? And, uh, they would say, of course I would like, you know, it's a, it's a mandatory thing in, in a retail store.
There's no way to not have shop assistance because they help, they educate, they, they do whatever is needed to make sure that the customers, uh, in the end buy. And it's exactly the same. So if you have one product, most likely that product is going to be very unique product that needs some convincing, needs, some education. As opposed to a retail store. Uh, the clothing store, we have thousands of, of skews. Uh, it's very easy to understand, right? You, you see a, a shirt and that's pretty much it. Uh, but there's a lot of them. So over there
the problem is different. Might be choice, paralysis, problem over a educational problem, but either way. Shop assistant in real life would be able to help with that. And same in the same way. AI [00:22:00] shop assistant, it's the same, same thing. It helps with the same problems if there are too many products. Let me help you with recommendation. Let me help you narrow down results if there's a few products, but there're, let's say very expensive. High price points and there's a lot of convincing that has to happen. AI is amazing in that because it is there with its empathy and knowledge of the product and complete specs of the product. So able to under, to explain what's the difference between that and that, and why you should get, you know, the higher level, uh, you know, the better tv, for example. Uh, so the AI is there to, to do all that and help in any type of, any type of brand. So we actually see success. By all different verticals. So from uh, electronics to uh, to apparel to, um, dog food, anything.
Matt Edmundson: That's, um, and, and does it matter about the data set size? Like if I was just starting a new store today, [00:23:00] would it work or do I need like a million rows of data for it to have any kind of meaning?
Shauli Mizrahi: No. So, um, also good question. So what we do is we connect, uh, to the retail, uh, e-commerce store. So Shopify Connection, one click integration. We already have the entire, uh, catalog ingested inside. Uh, same with, uh, purchase history. So we get and take the purchase history at the last couple of months to have all that data ingested to understand what people buy. To understand, truly understand the catalog. Then the AI also cross all the pages of the website, so. Any product pages, collection pages, uh, even block pages, anything to have that information. And that's the starting point. That's in one click, nothing. And then of course, if, uh, the, the me, if merchants want to add more information, they're able to add in more Excel information.
But that's only if they, they need to and only if there's missing information. But to get, get going, you don't need to do, have nothing, basically [00:24:00] just your, just your own website. That's it.
Matt Edmundson: And I guess the thing gets, uh, AI is, is one of these things that gets smarter the more we use it, right?
Shauli Mizrahi: Yeah,
Matt Edmundson: yeah. Yeah, yeah, yeah.
Shauli Mizrahi: So it, it learns during the process. So a lot of things that we actually highlight is that once we start, when they, once the AI is on the website, uh, the engagement rate goes up and up, uh, and we typically see around 3% of the. Visitors on the website will talk to our ai. So that's, that's a lot of people and that brings in a lot of, uh, new knowledge about how shoppers, uh, buy and what they're interested in and so on. And also highlights missing information, things that shoppers ask about. And the AI doesn't have the information because it's not on the website. Uh, what we do is that we highlight that information in our, uh, management console where. Merchants can see, okay, people ask about this and that there's no information on the website to explain that at all. And then merchants can add that information later on using, uh, using the tools that we, [00:25:00] uh, give them.
Matt Edmundson: Fantastic, fantastic. Ladies, gentlemen, let me just say, if you are enjoying the show, uh, and you are in fact in e-commerce yourself, whether you're just starting out or whether you've been around like me for a little while, whether you use, uh, show, Liz system or whether you don't, wherever you are. Uh, if you are in e-commerce, we have something called cohorts, um, which I'm a part of.
We have a UK cohort, a US cohort just started. And also one in Australia. So, uh, New Zealand, the a NZ one, as we like to call it. Um, it's free to join. Uh, there's more information on the website, ecommercepodcast.net. But if you wanna come hang out with other e-commerce entrepreneurs, ask your questions, share your insights and thoughts, uh, then do come join us.
It's quite a fun little place just to come and chat, meet people, and we're starting up new cohorts all the time. Um, so yeah, just go to the website, ecommercepodcast.net and, uh, find out more information about it. We'd love to see you in there. So [00:26:00] Shauli, um, as you were talking, one of the things that sort of struck me, I suppose, was, um, was persuasion, right?
This sort of, this old fashioned word of persuasion. Rob, Robert Kini, you know, the, uh, persuasion. Um, and if I'm selling a single product that's expensive and I need a, I need to persuade people. I can see how AI works, but I also get that there's a space for vending machines, you know, that sell bottled water, because that's a, that's not, it's not, is it a product?
Yes, I suppose it is, but it's, it's, it's just, it's water in a bottle, isn't it? You don't, I don't need someone to sell me that. I just need a quick, easy way to just, to buy it, which I think, what is, what Amazon's become good at over the years, you know, that's where you go to if you wanna buy your equivalent of the bottle water, I think is Amazon.
But one of the things I've noticed Amazon do. Uh, and I think most of us have noticed this over the last few months, is their AI shop assistant, haven't they? I, I, you know, so they're, they're obviously thinking, yeah. We, we, we should probably add this to our side [00:27:00] ourselves.
Shauli Mizrahi: Yeah. Yeah, they added Rufuss.
Matt Edmundson: Yeah. Yeah.
Shauli Mizrahi: It's doing a great job.
Matt Edmundson: I wonder, I dunno if they was rufuss, the name of the guy in Bill and Ted's, um, excellent adventure.
Um, the, the sort of the guy. Anyway, I'm just showing my age a little bit there. I just go, I just wondered if that's where they got the name from. I dunno why I thought about that. I just did. Um, but it, it strikes me that what I love with, with companies like yourself, Shauli, is you make technology accessible.
So it's not just Amazon that can benefit from it. Now sure. They've got gazillions to spend on the development, you know, and, and, and God bless them, you know? Um, but you know, we, we don't have to keep giving Jeff our money, do we? And so what you're doing is you're leveling the playing field and making this, um, accessible for e-commerce.
And this is one of the things that excites me about technology at the moment is, is just the ease of accessibility that we, that we have, um, and. I quite like that, if I'm honest with you. I think that's quite [00:28:00] a nice thing to actually happen. Um, and so yeah, I, I appreciate you doing it. Um, one question, uh, I suppose that comes out of all of this, then.
Where's it going to? Because at the moment, you know, we've got chatbots fairly soon we're gonna be talking to AI like I do with Perplexing. You're not gonna be able to distinguish it between a human voice and a, an AI voice. And then I see it is we're gonna have video calls with AI because it's crazy what Google can do with video now.
And so in your product roadmap, where's, where's all this going to do you think?
Shauli Mizrahi: So the more, obviously there's more adoption towards, uh. Towards AI and towards, you know, towards usage of ai. So we're on top of that. And one of the things we're coming out with soon in uh, Q4 is voice to voice, uh, full, uh, integration. So you're able to shopper, you're able to just talk to the same way you talk to per complexity, same exact idea, uh, just on your website.[00:29:00]
So just. On e-commerce websites, that's the next thing. But that's like, that's very short term, uh, roadmap. Uh, the longer roadmap and the way I see things, uh, develop in the future is that there's gonna be a, a portion of the, um, there's definitely gonna be a portion of the traffic that actually will never go to the website, but we'll have a conversation, uh, you know, using perplexity, using chat pt, they will have to talk to the website. So what we're working on and started working on is, um, it's called an MCP Model Context Protocol. It's the way AI actually, uh, talk to other ai. Basically, uh, in some big, some portion of the traffic will actually have an agent to agent conversation so that the shopper will actually use its own tool, complexity charge pt, and that tool will speak to other agents, other agents on the website. What we're working on is we'll be able to represent the brand and give them the agent to be able to do that. [00:30:00] So that's again, some portion of the traffic. Of course, a lot of the traffic would still come from Facebook campaigns, Instagram campaigns, TikTok, that's still gonna be, uh, that's still gonna stay there. But the chat or the search traffic is slowly migrating. People see that towards chat PT
Matt Edmundson: Mm.
Shauli Mizrahi: right? And we want to enable merchants to have their representation. In those platforms. So again, the AI of those platforms will speak to our AI that represents the brand, represents the merchant, and the merchant needs to make sure that, uh, obviously there's more, uh, more revenue for the merchant.
Matt Edmundson: This is, this is where it, it starts to get quite deep and quite fascinating. And so, um, and ladies, gentlemen, if you've not. If you've not heard of CPS before, it's worth just, actually just going onto YouTube and just watching a quick explain, a quick explainer video about what they are, um, this ability for AI to, uh, in effect talk to other applications or [00:31:00] other bits of software on your behalf and get information back.
It's kind of like APIs, but on speed in some respects, isn't it? It's kind of, it's the next sort of level. I think it's a really interesting technology if it catches on and it, and I see it starting to catch on and the fact that you are, uh, talking about this show Lee, and about developing these cps, I think is quite fascinating.
Um, because you are right. It's the fact now, I mean, my fa the, the AI of choice for me is not actually chat GPT anymore. I use C Claude, I use clawed a lot and actually clawed. Um, were Claude the ones that sort of founded the whole MCP thing anyway, whichever way,
I've integrated a whole bunch with Claude.
I don't even have to open up my email anymore. I can just say to Claude, I'll just, what's in my inbox that I need to look at and just get rid of anything that I don't need to know about. I, I, and it goes away and it reads your email and he goes, well read these, and I've thrown those away, you know?
Brilliant. [00:32:00] It's, it's, it's the most extraordinary thing and I think. The fact that you, you can, you are thinking about doing that with e-commerce, I think is actually really quite exciting because I think you are right, we are using chat, GPT or Perplexity or Claudes now to research products. Um, and I think if there's one thing that we've learned from TikTok and one thing we've learned from Instagram is actually when you build these platforms.
Make it so people can buy in the platform. Right. If you, if you really wanna monetize it, make it so people can buy in the platform, which Instagram we're a bit slow to do, let's be fair. Um, and I, I, I think it's quite extraordinary. So yeah, I Do you think it's going that way? That we will be able to buy products in our chat, GPT in our clawed models?
Shauli Mizrahi: Definitely, definitely. It's already happening. I know that. Um. A lot of companies are moving that way to, to make sure that that happens and Chedi is all for it. Uh, [00:33:00] Chedi wants to kill Google, basically, uh, search, uh, you know, Google shopping and Google search in general and bring their, the traffic their way.
So they're definitely moving that direction. And, uh, that's why our, our job is to enable brands to, to have, to be able to compete with us. 'cause otherwise we always, you know, the, the brand that doesn't have that representation, which is lost. And, uh, yeah, it's definitely going that way. The portion of this search traffic, for sure. No doubt. Again, there's, there's other tr other, you know, other traffic sources, email, uh, social that's gonna stay in their, their domain. But the search traffic is slowly moving from Google. And yeah.
Matt Edmundson: Which is fascinating, isn't it? It just, I mean, you do wonder what Google, I mean Google's obviously got Google Gemini, which they're trying like crazy to get everybody to use and integrate and everything into Google for the same reason. They wanna hold onto, I think that search traffic. Um, and to be fair, [00:34:00] they probably got access to way more data than every man, you know, and they should be able to make it work.
But it's, it's gonna be. AI is gonna be the, the sort of the next thing, isn't it? Everyone's always wanting to know what the next thing is, where's the next source of customers gonna come from? It was emailed. Then Google Search came along and then um, Facebook came along, and then it was Instagram, then it was tick, and we sort of go through these.
Oh, right, what's the next one? I need to get my head around. And it, it, it very much feels like it's gonna be ai, doesn't it? With these, like you say, everyone's sort of thinking about this, like, how do I get chat GPT to know that I exist so that when someone asks about product X, you know, we, we come up. I think it's, um, there's a big opportunity for companies isn't there?
If they, if they can crack that early doors, I would've thought.
Shauli Mizrahi: Yep. Yeah, so I think it's super important to be on top of, uh, the technology and understand where it's going and be ahead of the curve. 'cause that can obviously, uh, help with competition and [00:35:00] uh, I think companies should always, merchants should always think of how to be ahead and definitely not be behind.
So don't be late for the party.
Matt Edmundson: That's the title of today's podcast, ladies and gentlemen. Do not be late to the party. So the um, how long do you think it'll be before you have the MCP technology? When are we gonna start to see something like that come to play?
Shauli Mizrahi: So it obviously has, has to go also with adoption of that, of the, the traffic debt buys, uh, through Cha pt, through Claude. Um. We're seeing that somewhere in 2026, it's gonna become something. Uh, so we're developing it ahead, so we'll be there. So once it's, uh, the adoption is there, once shoppers actually want to buy, uh, without viewing the website, then we'll be there and we'll be able to offer that.
So we'll already have early customers that are using that specific feature.
Matt Edmundson: Yeah. Yeah. Uh, fantastic. Fanta, I'm very excited by that. And I'll be keeping an eye out on the website to understand what is going on there. [00:36:00] 'cause it, I do think you're onto something. I really do. I think that's, uh, that's quite clever. Um, yeah, no, it, one of the things, I mean, sort of rewinding back a little bit, so we've talked a little bit about the future, if we come back to the present day.
Um, one of the things, uh, that I asked you again before we hit record was, what's the one mistake we're all making? Um, and you said that it was, that we're thinking that AI is just really about customer support only. Do you wanna expand on that a little second?
Shauli Mizrahi: Definitely. Um, so again, our offering is to have AI that helps with sales on the website. Uh, AI is a chat bot, AI that, that's on the text medium. And a lot of brands, when they, when we offer that to them or we have a conversation, a lot of, uh, lot of marketers, head of Ecomms will say, Hey, oh, okay. That's chat. Chat is none of my business. Sorry. Chat belongs to the CX department or the customer support department and will actually will not [00:37:00] even understand that it is there to also help with sales. And this is the main function. We actually, uh, promote AI that helps with sales. It also helps, helps with support.
No problem. If someone wants to know their order status, no problem. Uh, but it's there to mainly make sure that the top line is, is higher, that we increase revenue, increase conversion rate, and average order value. That's the main thing. Uh. That's usually, um, not, uh, I would say ignored or like that possibility is ignored. So I think that's the number one mistake is that they're handing off the conversation to support instead of keeping it themselves and saying, I'm the owner of this. I'm the owner of conversion, I'm the owner of, uh, of sales on the website, and I should use every tool possible, especially new tools that are not coming out, uh, that will help me do my job.
Matt Edmundson: Yeah. No, that's great. I, I, and I get that, I understand that, um, there has to be a mental shift, isn't there in the, in the. Ahead of the guy or gal who's, [00:38:00] who's running the whole shebang and, and, and getting up and running. One crazy thought that came into my head sort of sideways there. Um, Shauli, as you were talking, is there a site out there?
I dunno if you know the answer. And ladies, gentlemen, if you know the answer, I'd be curious to see it. That actually you go to the site and it is purely AI driven. There is no, there is no navigation, right? Where you just, you literally go on the site. It's an e-commerce site. It's all conversational. Has anyone done that?
And I'm curious to know what the results are. So there's no nav, there's no products, there's no store, there's no shop by category or, or whatever. It's just literally, welcome to the store, tell me what you want. I'm your concierge and I'm gonna go away and I'm gonna start and, and I'm gonna start to build this for you.
Shauli Mizrahi: So I specific, I personally don't know of any website that has made that shift yet. I truly believe that this is the future. Okay. The future is there. [00:39:00] So, um, our solution, by the way, allows brands to have that ability. So if they on mobile, that's how it looks at the moment. The, the shopper speaks to the ai, it goes on full screen and mobile. And the AI does everything for the shoppers. So adding to cart, showing products. Even moving to checkout. So moving between different pages so the AI can actually, uh, completely, uh, replace the navigation of the website. There's no need for that anymore. So the ability is there, uh, and I truly believe that we'll start seeing more and more websites that will say, Hey, you know what? I don't need the actual website anymore because especially on mobile, by the way. Desktop, it's a different thing. You can see a lot of products all at once. It's nice on mobile. It's not on mobile. You can see one product at a time and you have to switch, you have to flip between. Dozens, hundreds of different products, right? And, uh, the ability of having AI is that AI is able to narrow down those results, have a conversation saying, you know what? We've got thousands of products. Let me ask you a few questions, uh, [00:40:00] about yourself to try to know you so I can, uh, recommend you something, right? So that, uh, experience is obviously much, much better than viewing thousands of products on mobile. And this is the shift that's going on. Truly believe that next year we're gonna see some, uh, websites that are gonna be a hundred percent conversational.
Matt Edmundson: I'm, I'm actually, as you're talking, I'm actually thinking, I, there's a few websites that I have where I'm thinking, how can I make that happen? Just as an experiment. I'd be really curious to see what the, uh, what the results were. So. I'm, I'm gonna have a little ponder on that. One of them is actually on Shopify, so I'm kind of like, I wonder if I could, anyway.
I will just now go down a rabbit hole in my head, which is probably not helpful. Uh, but surely, listen, I I, as I say this regularly to guess, but I, I, I genuinely mean it all the time. It just feels like we're just getting started with the conversation. Uh, and I'm, but I am aware of time. So, uh, if people wanna reach you, if they want to connect with you, if they wanna find out more about hello rep.ai, [00:41:00] um, how do they do that?
What's the best way to reach you?
Shauli Mizrahi: So go on our website. hellorep.Ai. Um, and if you talk to sales and from certain, uh, stage, you're gonna be talking to our sales and we're going to instruct you on how to, uh. To, uh, just optimize your account better. By the way, it's a fully self serve platform plug and play. So, uh, you know, it's super easy to, uh, to, um, interact, to go live. We offer 30 day free trial, super important, and we offer, uh, ROI guarantee annual your results. So we guarantee ROI of five x on on your results basically. So if you can't see an increase in sales. At least five times the, uh, price point, the, the tier, uh, um, value. Then you won't have to pay us. Simple as that.
So there's no risk on your side, and we offer 20% off, uh, to any of listeners of the show. So please mention that you, uh, you, you know, you heard on us on the show, um, [00:42:00] and you'll get to 20% off. Um, yeah,
Matt Edmundson: Brilliant. There you go. Can't say further than that. Uh, so yeah, so the website again is hellorep.ai. Um, Shauli listen, there's a couple of things I need to do before we sign off. Number one is I want to ask for the question for maps. This is where you give me a question. I. And I go away on the wonderful social media, and I'll answer the question over there.
As I've said to people over, over the months we've been doing this, ask me any question you like, what's your question, my friend?
Shauli Mizrahi: Since, uh, you're fan of, uh, American football now, uh, I'm gonna ask you, um, what, um, what, what do you think is the, why is American football better than English soccer or vice versa? And, uh, what tips would you, uh, give to someone who's also new to, uh, American football and, uh. Doesn't know much about it and how to get more, to be more of a fan of, uh, this new sport for me at least.
Matt Edmundson: Okay. Very good question. Very, very good question. [00:43:00] Uh, I'm, I'm, and again, I'm trying to think how I would answer this, but let's not do that now. I'll do that on social media. So if you've not done so already, come follow me on LinkedIn at Matt Edmundson. Uh, and we will post the content there at some point.
Uh, the, the second thing I like to do, Shauli, right at the end of the show. I call it now, saving the best tool. Last, what's your best tip, your best strategy, um, for our listeners regarding this whole topic of AI in shopping, uh, that we've not already covered. I want to give a, give them the most insane value, the best value to those that listened right till the end.
What would it be?
Shauli Mizrahi: I would say always listen to, um, to what's out there. Always explore new tools. Be ahead of the curve. Don't be a lagger. It's very important, uh, and just evaluate and truly evaluate tools. In their sophistication, in the value that they bring. So value is very important. Don't just jump into because to something, because it has a nice buzzword, but actually like try to and see the value.
And if it brings value, then [00:44:00] it's great. If it doesn't, then explore other options. Um, that's my tip for delay.
Matt Edmundson: I love that and this, this describes half my life at the moment, is just trying different things with different. Uh, uh, different pieces of software and just figuring out what's working well for us and ditching what's not working well for us, um, which is, is super, super helpful. So, uh, Shauli, thank you so much for joining us.
My friend genuinely loved the conversation. Uh, I thought it was fascinating, uh, really intrigued to see where it goes. And all the best with hellorep.Ai and, and let me know when you get the MCP thing working.
Shauli Mizrahi: Will do. Will do. Thank you.