Latest / Talk Commerce / Product Images Get You Chosen in eCommerce with Jehan Hamedi of Vizit

Transcript

Product Images Get You Chosen in eCommerce with Jehan Hamedi of Vizit · Talk Commerce

• Brent Peterson (00:56) Welcome to this episode of Talk Commerce. Today I have Jehan. He is the founder of Vizit. Jehan, go ahead, give us an introduction for yourself. Tell us your day-to-day role and one of your passions in life. Jehan Hamedi (01:08) Awesome. Thanks for having me on, Brent. It's really nice to meet you. Really nice to be here. Great to meet everybody who's listening. So as Brent mentioned, my name's Jehan, founder of Vizit. A bit about my background. Kind of Midwest guy, grew up in the Midwest, just outside of Milwaukee, Wisconsin. Moved to Boston for school, and have lived here ever since. I've I've worked in the AI industry now for about 15 years. building, designing, and scaling AI measurement systems for enterprise brands and retailers. the thing that I really geek out about and what I'm really interested in is consumer behavior, trying to understand why people do different things. you know, it turns out that sometimes it's really hard to kind of explain the choices that we make. but ironically, a big part of my career and background has been finding that. The more that we use artificial intelligence, the better we can actually understand ourselves and understand others. and so I just kind of started my career working in kind of text analytics. So understanding, you know, based on what people say and how they review products, what are the reasons why they're more likely to choose certain products over others, other brands, brands over other brands. and now at Vizit, we are we are kind of focused on the analysis and understanding of visual content and imagery. Brent Peterson (02:25) That's great. And passions outside of of AI? Jehan Hamedi (02:28) Outside of AI and work, well, let's see. I am a I'm now a father of two. both of my kids are under three. So I am kind of learning what it means to be an AI entrepreneur and of and a father. and I'm having such a blast. my kids are are so wonderful. and outside of that, once upon a time, well, actually, a fun fact about me, or I guess our family, is that my great uncle was Wasn't the brewmaster, but he was a senior brewer at Schlitz. So like I said, I'm a Wisconsin guy. And so there's Brent Peterson (02:56) Wow. Jehan Hamedi (02:56) like beer in our blood. And I've always been really interested in craft beer. so one of the it's kind of like a hobby of mine. I'm not a not a particularly talented brewer, but I would I love to, you know, try and sample other other people's art. I really do think brewing is is as much a science as it is an art. And so, you know, I you you'll you'll find me at different, you know, kind of beer festivals. You'll find me whenever I go to a new city. And we were talking about this just before we started Brent. Like when I go to Hawaii, I've got to have whatever's local, whatever is made here. I want to hear the story. I want to meet them. I want to try it. That's just something that's fun for me. but yeah, other than that, a lot of lot of like I look at pictures for a living, Brent. So that's that's that's what I do. Brent Peterson (03:37) Yeah. That's awesome. well before we start talking about beer, we we I do this thing called the Free Joke Project. I'm just gonna give you a joke. I'll tell you a joke. You just give me a rating eight through thirteen, and now I if I if I would have known I would have come up with a beer joke. There's probably tons of but here we go. This will Jehan Hamedi (03:53) It Brent Peterson (03:54) be a quick one. Bruce Lee was fast, but he had even a faster brother, sudden-lee Jehan Hamedi (04:00) Like an eleven. Brent Peterson (04:01) All right, thank you. Appreciate it. all right, so let's I'm I th you know I think visual AI is like so important and people are so used to typing in, you know, their thoughts in a prompt, but you know, I I think that the the ability for AI to look at images and it's you know it's been around for a while. I and when I say a while, you know well it has been around for a while, but give us a Give us a background of how you dived how you got into this and and what kind of problems you're solving. Jehan Hamedi (04:30) Sure. So as you said, AI is is particularly fashionable today, but it's been around for quite some time. And AI is a very broad, sort of encompassing you know, field of a lot of different, you know, technologies. but as you sort of think about image or visual related AI applications, one of the most prominent sort of innovations that that you know I think folks across across the river at MIT and others, you know, were were leaders here for years, was trying to understand what is inside an image. Like so detection and classification. This these are the most common applications for you know kind of computer vision related AI tasks today. So you know being able to detect, you know, you know faces to be able to detect, you know, specific objects, to be able to label images as being, you know, well, this one's outside, but this one's inside. This is a kitchen, this is a kitchen and a person, or a kitchen and a person on the left. And so as the field has really grown, you know, we in AI can get a really good semantic understanding of kind of what is inside an image. And Now there are lots of technologies out there that can kind of help you tag and you know describe what is in inside an image. But that technology falls short for specific applications, one of which is is the one that I'm really interested in. So what I'm really interested in is, and kind of what I have observed is that. imagery like we we as a as a species human beings are are are a visual species we we remember things visually we communicate now increasingly visually and if you think back to any product that you've ever bought you know whether that's you know a the coffee maker that's sitting on on your you know your kitchen counter your shoes your shirt you know the fruit snacks that you bought or or you know the Giant jar of pickles that you got at Costco. All of those purchases have exactly one thing in common. And it's that you saw something, it sparked something, and you wanted it. And that moment is very hard to explain and even harder to measure. And so the thing that I'm really interested in as it relates to imagery is not describing what is inside an image. Like that's been done. I'm really interested in understanding what is a good image, what is an effective image at presenting a product in the most compelling way to get somebody to choose it. And so the state of the art that kind of I've been really excited and and you know passionate about driving is in image scoring and an understanding sort of for a particular audience in a particular market in a particular product category. How good is this image? How much conversion power does this have? And so that's been so cool and so fascinating to be able to measure because when you know what is good or bad about a product image, and you can understand why, you now have the recipe or the formula to be able to fix it automatically with some of the latest. innovations that everybody has played with now, which is image generation, right? And image editing that has kind of a, you know, a programmable AI backend. So it's just been just kind of like a a really, really fascinating journey that I've been on where I never actually, you know, aspired to you don't you don't wake up one day. well maybe some people to do today do, but you know, several years ago, you didn't wake up saying, I want to be an AI founder. You start you start by saying, I want to solve this problem. And it just so happened that AI was the best technical solution to the problem that we wanted to solve. We just had to advance the state of the art in order to do it. Brent Peterson (08:20) Yeah, that's super interesting. It's very subjective though. So like yeah, I I I do a lot of training just on on text based items and I try to to get to learn a client's voice. How do you Jehan Hamedi (08:32) Mm-hmm. Brent Peterson (08:32) keep it in line, Freud? Is it the same thing? Because at the end of the day, your your LLM is describing the what the image looks like, and then it takes that description and runs it against something to compare it to make sure it's It is good. I mean, I there's so many different definitions of good that it right? Like Jehan Hamedi (08:49) Mm. So Brent Peterson (08:51) how how do you create your baseline, I guess what I'm trying to say. Jehan Hamedi (08:54) Sure. So the right definition of good is not whether something is compliant or whether something is, you know, on brand or off-brand. It's all about the shopper. Beauty is in the eye of the beholder. So the audience, the purchasing segment, and their preferences are all that matters. So Vizit was founded on the belief that it is possible with an artificial intelligence system to simulate the beholder. And so we pioneered. Synthetic audiences to do this. And so what underpins Vizit system is what we call the visual kind of conversion graph. And so that is a, you know, mere decades worth of behavioral data, product data, sales data, social data, search data, all of these, like literally trillions of data points that go into models that we train in-house ourselves. That can learn from shoppers' real behavior in particular categories and then build an understanding of how and why they choose certain products over others based on those signals. So that's what we find as kind of being that sort of ground truth, like as you said, the the foundation. And then what you can do with that data and the models you train on top of it is create a new measurement standard. Predictive one, one that can look at a new image that it has never seen before, that you have never even put online before. And it can analyze kind of atomically all of the attributes, the pixel by pixel relationships of everything that's inside of that image, and make a prediction about how likely that is to convert that particular audience based on all of the patterns of high converting and low converting things that they've seen. And it's extraordinarily accurate. and it's been really cool to see because when you when you sort of think about the sheer volume of of content that companies are are our brands and retailers are producing today, and now they're you know producing with you know some AI assistance, it's extraordinary. But it's been unmeasured. Imagery is the last unmeasured component of The digital shelf. we have data, e-commerce managers have data on everything except that moment, the first impression that a shopper has with a product, that is instant. It's irreversible. and it kind of impacts all downstream spend and upstream spend. Brent Peterson (11:18) Is video how does video fit in? Because I would imagine video is less monitored or less tracked than images yet. Are you saying the images video is more or less than learning about how video behaves for a consumer? How does that fit in? Jehan Hamedi (11:33) Yeah, I I would say, yeah, I broadly speaking, like visual content is is is is a field of s inclusive of video GIFs, you know, still images, is a significantly unmeasured, under understood sort of component of the you know marketing and e-commerce you know tool belt. you know, I I think you you sort of think of you know, you could take A photograph of a cheeseburger infinite different ways, trillions of different ways. And so when you sort of think about that, and then you think about how shoppers make a first impression in under 13 milliseconds, you know, that that content, that first content is so important. So in a video, you could spend a fortune creating a 60-second spot. or a two-minute spot. But if that first frame and set of frames isn't attractive enough to get clicked on and isn't attractive enough to sort of, you know, emotionally engage the viewer, you've lost them. And so it's the same thing with the still image. You know, if you're clicking through a product carousel, you know, that that is a very we we've all done it, right? Where we've searched something on Amazon or your favorite retailer and the search grid of items returns. One of them catches our eye, and we go, we click that. And you go to the product page, look at the first couple images, and you're like, you click back, and now you're you do the same thing with somebody else. and so, you know, people vote with their thumbs now. And in this world where these decisions are happening so quickly, you know, tens of thousands of times a day, think about how much the average person scrolls on their phone. those are moments that brands will never get back if they didn't make that right first impression the moment that a shopper first sees their product. Brent Peterson (13:20) Yeah, interesting. So tell us tell us some of the problems you're solving then for clients. Jehan Hamedi (13:26) Sure. So if you called a a brand manager today and you asked them, you know, let's just, you know, pick it pick an example. If you called the the brand manager of you know a a popular, you know, deodorant brand, and you said, What are the best, what are your best images of that brand? Let's just say Old Spice, for example, you would get crickets because they don't know. Brands don't have a way to measure images today at all. And so if you don't know how good your content is, just think about that. That's like if that content is the face of all of your products, and you don't know the performance levels of different content, that means that of all of the traffic that you are getting organically or inorganically, you know, you're paying for. Paid search, paid social, you know, retail media, you are potentially driving millions of eyeballs to content that is going to, that's not going to win and lose their attention. so the first thing that Vizit provides with our AI system is is sort of a a new measurement. We call them Vizit scores that allows our customers to understand how good or how effective or ineffective their content is today. Relative to their competitors in the category. And that helps you do a few things. It helps you understand like where you should prioritize content optimization and sort of investing in new imagery. but it also gives you a really great kind of new source of competitive intelligence. Because, you know, for years, people might lose a couple points of search share, a couple points of of market share, and they might say, you know what, we gotta We got to rebrand. We got to change our packaging. We if you're in CPG, or you know, we need to we need to change all of these, we need to make some big changes. But oftentimes, what I have found is that the fastest path to incremental revenue starts with one better image. So for our customers, we will not only help them understand sort of where they sit today, but the other beautiful thing about AI is it scales. So with Vizit's AI in particular. What we create for our customers is we allow them to apply a consistent conversion standard across all of the new content that they're making. So all the new images in every market on every digital shelf, so that they can ensure before they put content in front of real shoppers, it has already been pre-validated to be successful at converting. And then on the flip side, when you have a large catalog, thousands of products online, and that content, you know, is performing at varying levels, Vizit can identify, recommend, and optimize that content to then make sure that the the brand is basically maximizing their return on advertising spend, but also the the conversion rates with those products. Because ironically, imagery is one of the, if not the biggest driver of conversion. you know, text kind of gets you found, but images get you chosen is what I've found in my Brent Peterson (16:28) Yeah, and it's have you seen patterns then of what what merchants are using for successful images, or does it depend on the brand and the cohesiveness of the brand and the image compared to the brand, things like that? Is there any patterns you've seen there? Jehan Hamedi (16:41) Yeah, context is king. You know, it it certainly depends on the brand. it depends on the category, as well. And what we have also observed, which shouldn't, you know, surprise you, Brent, is that shoppers in different countries expect and want and prefer different things. So, you know, the best way to, you know, photograph cat food for, you know, a US shopper on Amazon is going to be markedly different than one in in India, than one in the UK, than one in Japan. And so the the thing that we tell our customers and the thing that we've learned from our customers is to focus on the shopper. So we can all sit in a room and have great ideas about, you know, the the the brand vision that we have or the promises that we want to to to make shoppers. And those are all really important, right? A brand is is a very powerful asset. But we need to kind of again turn the turn the table. Like can what if you could see through your shoppers' eyes the moment they s see your product and you know a light bulb goes off? It says, I didn't even think of the fact that, you know, these types of lifestyle images tend to be more and less converting. I always thought those were like, you know, just okay. Because people didn't have the data, they didn't have the intelligence, they didn't know. And so the biggest mistake that I think brands make is that they treat their imagery and and track their imagery as a creative deliverable as opposed to a performance asset. And when you sort of make that mental, you know, leap, or you change your your sort of your your point of view, you start questioning things because You know, if you like right now after after our our session, I would encourage you just you know download an image, can be any image from any website, and right-click it. And you know, you're gonna find on if you're on a Mac, you're gonna find metadata like its size or you know, when it was captured, and it's you know, then we might even see some information about the RGB. It's descriptive, but it doesn't tell you like if it's good, doesn't tell you if it's effective at selling that product. And if you are the merchant, if you're the brand manager, if you are Anybody in that sort of chain of value, that is exactly the measurement that you want to persist everywhere that that asset lives throughout its whole life cycle. Brent Peterson (18:55) Are you proposing that we start putting metadata within the image to help the LLMs better read those images? Because it's still subjective, right? Like a can of hams is going to look different than a can of schlitz. And personally I think the hams makes a much better crown if you wanted to build a crown out on your head than a can of schlitz, but like the visuals is still gonna be up to the consumer clicking on it. And are i do Walk down excuse, walk walk me down that path a little more. Were you just using that get info as an example of how to get data from your image or g explain that a little farther? Okay. Jehan Hamedi (19:30) Yeah, just an example of like what what you kind of have, what what everybody has at their fingertips today, just as an example. But to answer your your question, in anybody who uses any type of AI system today, the single most important thing you can give an AI to do a job better or more effectively than you could do on your own is context. Like specifically context about the job to be done. And now like outside of what Vizit does, you know, that can be context about, you know, your your business goals, your competitors, your you know, the expected outcomes that you want to have, other reference material. all of this information helps the AI do a better job. And so what we have found at at Vizit is Vizit has some proprietary Context that is extremely helpful for AI systems in in our use case. We have context about the shoppers. We have context about the product categories. So the consideration sets, the attributes of those products. We have context about you know sort of markets, other things that sort of help us understand the universe of imagery that you are ultimately competing against, because image is a, you know, sort of a relative competition. and when you have that information, and in our platform, when you have when we generate scores and insights and like literally English language recommendations on how to improve an image, you now have an extraordinary amount of context. About not just what is inside the asset, like its features, its its elements, but you have context about how it needs to compete and how products are chosen or not in that category. That enables you to instruct an AI system to very intelligently create a version or a variant of an existing image that you have that's low performing, but to inherit. The qualities that would make sense given your composition and your goals, but learned from the cues that are driving purchase behavior in your category. So, all that to say, Brent, I think like the companies and the brands, the retailers that are ultimately going to win in commerce are going to be the ones that can harness and compile. the best context around the either the assets that they need produced or the decisions that they need made, and then create a feedback loop so that that context gets richer and, you know, pruned over time through through again, you know, activity data. Brent Peterson (22:10) And how important is the human in this feedback group loop? And I I'll point at P Max from Google. If you left it just to go on its own, it's gonna it's gonna find the best best people to send you to your site, whether they're gonna whether they're good clients or not. They're just it's just trying to measure click or a conversion, right? how do you I I guess if you have a product that's easier, but if you have a service where you're just trying to get a forum fill. Theoretically that feedback loop could go way in the wrong direction without being sort of checked periodically. Jehan Hamedi (22:43) Yeah, I mean, you know, for as you said, it is a little different with a with a with a product. So, you know, with with Vizit, you know, our our instrumentation is is extremely intelligent and it's it's learning from real data and real behavior and that's sort of how it's it's kind of producing, you know, the the the outcomes that it is for our customers. But as you said, you know, for other in other applications and other industries and other verticals, I I think you're right that, you know, the the the really important thing to start with when you're considering any of this is again, like what is what is the goal? Who's the audience? And what is it that's kind of make what what makes them what goes into their decision making process? So I can think of an example, as you said, that if your goal is to, you know, have all these form fills, but you have an agentic system that is reaching out to hundreds and hundreds of people to then direct them to your site. But it turns out those people are not the, they're not the ones that, you know, are most apt to buy your solution. As you said, you now just you created sort of this this traffic. But you have the you have a mismatch between the content and the offer and the audience. Or oftentimes you might be getting exactly the right people, but then when they show up, you're not showing them or offering, you know, what they expect, and then they leave. So I think where a lot of the loss happens, and this is this isn't just like, you know, opportunity cost, it's like wasted spend. Unrealized incremental revenue comes from spending money to create demand, but not thinking through fully what that demand is exposed to on their shopping journey. Because if you have, you know, a big retail media budget, for example, or you know, a really large meta budget or a large Google budget. You can get eyeballs to kind of anything and you know it can be really targeted, it can be very competitive at large volumes, but for again, first impressions are everything. And so if you're not thinking through exactly how like the the journey that those that those folks are on, you might very well develop lots of demand without the ability to harvest it effectively. Brent Peterson (25:00) Yeah, and does this also apply then to the whole customer journey and not you know, like from landing to clicking on navigation to getting to the right product page or however determining which images are appealing and even on a mega menu, is it does are you getting a bonus for having an image within your mega menu to bring people there? It I'm assuming you can use all this data to help drive traffic to places you want to go and then get people to go other places where you'd like them to go. Jehan Hamedi (25:27) Yeah, I mean, we are entering a world right now where AI is not just personalizing experiences, but it is personalizing products. And some of the smartest, you know, companies that I've worked with over the years will inject AI not just throughout the customer journey or the content creation journey, but the product development lifecycle itself, because a product is an image, right? You see it. in the physical world, in the digital world, it's you know it's obviously contained inside some you know visual asset. And so the the the kind of look and feel of the the product itself, like the the packaging, the box that that product is in. If it is a box that you know has certain colors and certain, you know, text size and you know contrast between the text and colors, well If that's your box and your box was at Best Buy next to these three other boxes, which of those is going to get noticed? But then more importantly, which one is going to elicit more of a like desire to purchase, not just something you you you sort of see. And so that's where like you can start to get extremely smart about color choice, about sort of typography, about you know contrast between different you know visual elements. I mean look there's there's there is a statistically right number of chocolate chips to show in a chocolate chip cookie to make it look more more appetizing. I did that research. And so you can imagine that like times All the different things that you're thinking about. If if you are a You know, a pet care brand, and or you're an emerging pet care brand, you're we want to create this ultimate pet care brand. Like depending on your target audience, depending on the category, depending on the market that you're in, there is going to be a right cat to use. Like the best cat for your particular goals in your context can be measured. so it that's it's kind of a a thrilling thing to think about how much. we can learn from, you know, people's choices to help us as a as a collective industry make better content for for our our customers, make better experiences, and ultimately make products that are are superior and and you know desired. Brent Peterson (27:40) what what about so you mentioned image like in image changing, image enhancement. I've worked with the imaging tools like in Shopify or in Google. Some of them are pretty poor, and I would say Google does a terrible job on redoing your images in a Google campaign. is there a better Tools outside of those platforms themselves that you should be looking at. Like I I mean I I've used Banana Nano or Nano Banana. It's fantastic. I don't know why they don't employ some of those rules within their own system. And I've seen, you know, 16 by 9 images that were cropped to one by one, but they didn't bother to resize. They just cropped it. Like you would think with all the tools that we have available that there are our providers could be smarter. So I I don't know if there's a question there, but I'm I guess maybe what are what are the best tools you've seen when you wanna make those new images or do you provide that? Jehan Hamedi (28:34) So it's a good question. I mean there's there is some really awesome technology that that I've I've seen and been inspired by over the last couple years. I think the the back to the point about context, the devil's in the details when it comes to image generation and and image you know transformation or editing, making changes to images programmatically. Because general sets of aesthetic rules, general sets of, you know, compliance criteria just don't work for that like use case. Because to make a better image, you need to know, like, well, who's it for? What else are those is that audience seeing or evaluating me against? How successful have those things been or not been? How you know. successful have my, you know, different kind of visual iterations been and why. Like there's a whole context packet there that is required in order to to to do this successfully at scale without you know creating, you know, five fingers or like you said, the weird crops or hey, this is a great image, but now my product looks totally different. Like you actually changed my product. I can't do that. Legal won't be happy. So our our solution to this at Vizit was sort of we we built it for our particular customers and and our use case. So we have a tool that's called Vizit Spark. and what Vizit Spark is specifically engineered to do is it's the first, you know, generative AI engine that is is learning from and sort of infusing conversion data. And understanding into the assets that it's producing. So when a customer and Vizit has a low Vizit score and we have all of the context information about why, we can instruct an AI system exactly how to improve that image given all of that detail and reliably produce higher scoring, higher converting images in just a couple seconds. So It's one of these things that's why it's been so thrilling to work in AI for as long as I have for 15 years, because you know, a few years ago, it was not possible. Right. Like this, you know, that's what's so cool about artificial intelligence is like the state of the art is is is advancing by the day. I have friends that are PhDs that, you know, used to, you know, casually keep up with the new papers and the new new publications and research that's coming out. Now there is just so much. It's like, you know, you you know, I we need AI to understand all of it, right? but what we have what we've found, and I'm kinda you know commenting for my application, but there are, I'm sure, dozens of other applications where you know the nanobananas of the world are going to be extremely effective at sort of generalist use cases. And then there will be other applications for for generative that are are are really going to depend on that data in the context that is something that aga again like you know is kind of a combination of of customer data, public data, proprietary data, you know, third party or licensed data, you know, what have you. Brent Peterson (31:35) Yeah, very cool. geez, we could talk all day about this. so yeah, yeah. So Jehan Hamedi (31:38) It's fun. Pictures are fun. Brent Peterson (31:41) where where do you see the market heading now? This we get this good we're s we we're believe it or not, we're gonna start thinking about Black Friday. Do you see us in I do you see this just really taking off and and and a lot of this I I heard a new term, not a new term, but qualitative. analytics based on usage of sites compared to just quantitative, you know, like that you're you're measuring the quality of the user on the site rather than just the what they're doing. Are you seeing that is that like the next big trend we're gonna see or what what do you think's gonna happen this next year? Jehan Hamedi (32:15) Yeah, I mean I I think it it it goes back to this this notion of of of good is I I think AI makes it possible to measure the unmeasurable and it allows you to scale that intelligence across, you know, a global catalog, across, you know, tens of thousands of products, potentially millions of images, depending on how large the the retailer or or the brand is. So, you know, I sort of see two things. The first thing Is that for black for Black Friday for you know second half of this year, brands should be and will be asking the question, is this content going to work? You know, how effective is this content going to be? Because you know, that Black Friday moment, so as you you well know, Brent, like so much rides on on that. And if you sort of Instead of thinking about the campaigns and all the spend, turn the table, think about the shopper. What are they literally seeing on a product page that is going to compel them to consider and ultimately choose a product? And so what I would do is I would make sure that before I spend a dollar on any of that media activation, I would ensure that. All of the traffic that I'm going to be driving now has the best possible chance of producing real sales because I've validated my content decisions beforehand. If you go and fix all of this stuff in January, you know, it's too late. You know, you'll be converting any organic traffic to those products like later after the fact, but you will have have missed the window. And so my sort of macro prediction is that I believe that in the next couple years, if not even sooner, every product image will be scored before it's shipped. Just like, you know, the same way if you were a brand, you would never put out a financial report without auditing it first. You just it you just wouldn't do that. It doesn't make sense. And so why would you put, you know, Content in front of real potential customers, millions of them, and bet half your marketing budget without being totally sure that it's going to be effective. I think that world is over. And I see just a total new transformation towards more predictive conversion intelligence to understand, you know, before you put things into market. Will they be successful? And if not, can you reme you know fix it before you know deploying that spend? And then as soon as things are in the market and you're tracking and learning from what competitors are doing for how shoppers, you know, in the wild are are reacting to or perceiving, you know, your products, you can then make more intelligent, focused content optimizations. Brent Peterson (34:57) Yeah, I mean such good that's good such good feedback. I you know we have a few minutes left, Jehan, and as I close things out, I do a I do a segment called the Shameless Plug. But before I do the shameless plug, I want to plug humans because I do believe that we are going down this path of just of just AI doing what we're doing. But I do feel as though If we do if we automate everything to the point of where we don't need a human, and you just have a machine determining what is the best image, you are really just term determining what is the best image that a that a machine thought a human would want, and then a human validates that by clicking on it and and consuming it. But it doesn't really tell you if there's anything new, like that like AI is only ever going to look sideways. It's never going to look forward because it doesn't have that shared knowledge. It doesn't have that compiled and scraped, like you can call it scraped knowledge that the L that the LLMs have. So I I would just like to plug human beings for offering new solutions outside of what any LLM would offer. But that I'll give you an opportunity to do a shameless plug for whatever you'd like to plug now. Jehan Hamedi (36:10) Yeah. Sure. Well, I'll I'll I'll second that. So I, you know, I've the I bet many of the folks listening to this have been to one or seven, you know, agentic commerce conferences or talks just in the last couple months, right? It's really important. but as it relates to commerce, you know, this idea that agents are going to just take over and do all of the shopping for us, like without any supervision, I I don't subscribe to that. and the the the data is very real. 95% of of consumers that utilize AI as sort of you know an assistive technology on their shopping journey will take additional steps before they ultimately make that purchase. So you might be in Chat GPT and ask it, you know, what are your best recommendations for you know for this type of product? And they give they You know, they're the ones who are filtering those results, but people want to validate. So you'll still go to the product page, look at it. You know, you'll read the return policy. Like, is it actually, is it what they said it is? you know, these you'll read a review, the most recent review. And so it's one of those reasons why, like, you know, Walmart put out his research that they they had a three times higher conversion rate. When people clicked through from Chat GPT to the actual product page, because people want to see things. People don't buy products, they buy outcomes. I'm a big believer in that. And you need to like see and imagine the the the outcome, the the life state that you're gonna have when you purchase something. And so, you know, I I think that AI is an extraordinarily powerful tool. But again, back to the founding of of our company. Like we we still talk to humans all day long. We we have we have a company of humans. and you know what what AI helps us to do is is automate things that you know either take a lot of time or we can kind of add data in context to to make ultimately make better decisions. So my shameless plug, you know, obviously I I've talked about about my company Vizit a couple of times, but The the plug that I would that I would challenge everybody who's listening to this this podcast on is ask your brand manager, ask your e-commerce manager, how good is our content? How good are our images? Are they better than our competitors? And if they can't give you an answer, they're you're doing something wrong. Because there is a there is a better way to think about sort of the effectiveness of your content. And today, you know, we are all wearing more hats than I think most, you know, workers have for a long time. Like there's a lot that's expected from e-commerce managers now now today. they have to do a lot more with less. and so, you know, in this world where, you know, we have to work increasingly more efficiently, and you know, every, you know. dollar spent and every investment we make, like it, you know, is really meaningful. It's important to have the data and the intelligence that, you know, will will kind of enable you to to make the best decisions because again, first impressions are instant and they're irreversible and they cost a lot if they're wrong. Brent Peterson (39:16) Yeah, that's awesome. will you be at ETAL Boston coming up in in August showing off your product? Great. I'll be there as Jehan Hamedi (39:22) I will. I will, yeah. Brent Peterson (39:25) I'll I'll be there as well. It's perfect. I I I only mark and say that because last year I I hosted a panel at E-Tail and I had the lady, I think she was a CMO or C R O at Edible Edible Brands. They own Row T and a bunch of other brands. And Agentic was such the buzzword. Everybody was talking about agentic commerce, and I it was 400 people. I said, raise your hand if you as a merchant or you as an agency are employing a agentic right now, and there wasn't one person that raised your hand. So I do have some panels again this year. I'm looking forward to asking the same question to see how much we've turned. I do agree though it Jehan, that where you're going is much more effective and a great use of AI in terms of how to get people to convert and how to make sure that your content is really good. So thank you so much for doing that. Jehan Ham yeah, you're welcome. Jehan Jehan Hamedi (40:18) Thank you. Thank you for having me, Brent. It's been a lot of fun. Brent Peterson (40:21) ha Jehan Hamedi is the founder of Vizit and a Schlitzbier fanatic. So much, thank you Jehan Hamedi (40:29) Okay. Brent Peterson (40:29) so much for being here. Jehan Hamedi (40:31) You bet, Brent. Thank you.