E215: Joanna Lambadjieva

Richard Hill [00:00:00]: Hi there, I'm Richard Hill, the host of eCom@One, and welcome to episode 215. In this episode I speak with Jo Lambadjieva, CEO and founder of Amazing Wave, right off the back of last week's episode where I share what we're doing in the agency with AI. Jo brings a fresh perspective from working with dozens of Amazon and Ecom businesses integrating specific e commerce SEO processes with AI. We cover loads in this one: How AI can help uncover hidden inefficiencies and protect profitability, What's the best way to introduce AI into workflows without overwhelming people or creating resistance, Some of the most common mistakes Jo sees seven figure sellers making when trying to adopt AI and Jo's must have AI toolset and of course so much more in this episode. If you enjoyed this episode, hit the subscriber follow button wherever you are listening to this podcast so you're always the first to know when a new episode is released. Now let's head over to this fantastic episode. Hi Jo, how are we doing?

Jo Lambadjieva [00:01:04]: Hey, I'm good. Thank you so much for having me, Richard.

Richard Hill [00:01:08]: No problem at all. I am very excited for this one, as you might have guessed after having a little chat beforehand, but I'm very excited to get you on. Thank you so much for coming on the show. But I think before we get into all things AI, Ecom, Amazon, it'd be great to tell our listeners how you got into this world.

Jo Lambadjieva [00:01:26]: Sure. So I've been in digital marketing, e commerce for about 14 years now and before I started getting into AI, I kind of did a variety, I wore a variety of hats in the e commerce space. I used to head up the Amazon team of an Amazon aggregator. Before that I used to run a boutique vendor agency for Amazon vendors. And before that I used to work with wider e commerce brands or like D2C and retail space on their performance marketing. So I kind of come back come like my background is very much in e commerce, in strategy and processes. And so in November 2022 when ChatGPT came out, I started at that point trying to figure out how I can use ChatGPT for my team at the time. And then I was like, oh wow, this is so incredible.

Jo Lambadjieva [00:02:17]: And I started really geeking out about it and anybody that knows me or reads my newsletter can attest that I am 100% geek. And so gradually I started just trying to learn as much as possible about AI, specifically in the niche or the use cases of e commerce and then slowly grew my kind of, I guess Passion, knowledge and then yeah, footprint in that space and so that's basically it.

Richard Hill [00:02:47]: So full time in AI then.

Jo Lambadjieva [00:02:49]: Well, I think this is the thing, it's like, it's misleading. I would define myself more as an E commerce professional who knows AI quite well.

Richard Hill [00:02:57]: So yeah, so yeah, you're sort of an E commerce professional first, but you're using AI to help E commerce professionals, but using that sort of E commerce knowledge to align sort of AIs tech. The use of AI in Ecom. Yeah, very, very scary. It's so excited. Literally, it's just like, literally this is the thing that I sort of probably go to bed three nights a week and think like, oh, we could do this, we could do that, we could do this, we could do this.

Jo Lambadjieva [00:03:26]: It creates such a, it's, it creates such a crazy fomo. Like even I, I can feel it with myself. I, I write a newsletter all about AI and E commerce three times a week and I literally read everything there is about, just in general about AI, but also AI and marketing, e commerce, etc. And I still feel like I'm behind. And so, you know, you can get this feeling of oh my God, there is just so much opportunity, there is so much going on. What should I do next?

Richard Hill [00:03:53]: I think that's going to resonate with it with me definitely. But the listeners, I think, you know, ultimately you spend, you do and I'm probably not too far behind you, but definitely behind you in terms of the amount of time, I'm sure. But spend a lot of time thinking about, researching and trying and different things. But then I do wonder, are we doing enough? Are we doing enough? Are we doing enough? And I think for the listeners that are listening now that are the business owners, the marketing directors, you know, what would you say to those guys? How would you know? They're probably getting pressure from above to do more with their AI stack. Whether that's, you know, in the market, obviously, marketing, operations, all sorts of areas within the Ecom side, you know, maybe, you know, what stage of a Ecom business should the sort of senior leadership take take AI seriously?

Jo Lambadjieva [00:04:45]: Any stage. You can literally, honestly, this is my very, very, very honest opinion. You can be literally just starting out. You might be just not even starting out, you might just have an idea. Start with AI because AI can obviously we talk a lot right now about agentic AI automations etc and this is like I would say more , some more complex processes. But AI can just in general help you so much with anything from like ideation to Getting to brainstorming, to injecting data into your hypothesis in terms of like, how to, let's say, develop a business, how to find the next product, how to improve your product and anything. And so I would say at any stage of your business, you should be already testing AI and seeing the use cases that it can help with. Now, of course, the larger, I would say the larger your team is, the more actually I can really help improve efficiency, improve processes, and augment the existing talent that you have in your company.

Jo Lambadjieva [00:06:01]: But even if you're just a one person band, I would say it can like, you know, 2x or 3x your productivity and also give you new ways to think about existing product problems that maybe before you haven't really ever thought about.

Richard Hill [00:06:20]: So do you walk around talking to Your, let's say ChatGPT or your, depending on which one's your favorite. Are you one of those?

Jo Lambadjieva [00:06:30]: Yeah, yeah. I mean, I have to say yesterday I was going for a walk in the park and I just had a little conversation with ChatGPT and I had a few people give me a little bit of a weird look, but it's amazing. It's like you literally, for example, to give you a real use case yesterday. So I had a business call and I had to put together kind of a proposal, but I just really hadn't had the time to really think really through some of the mathematics. And so I asked Chachi PT to like be kind of my sparring buddy and like help me like go like step by step to think. Okay, what will make financial sense? How to like analyze the situation? What are the pros and cons? And I was like, okay, I'm just gonna do this while I'm walking because I think much better why I walk. And so we did it that way. And so like by the time I had the meeting, I was ready because I had this like second brain that helped me think through something.

Jo Lambadjieva [00:07:35]: Not give me the answer, but help me think.

Richard Hill [00:07:39]: It makes you think about things. Oh yeah, I didn't think of that. And then you can drill into that thing. I did exactly that. Literally. It's almost like we did the same last, last, whatever, Thursday I was on the treadmill in the gym and I'm just literally headphones in, talking to chatbots, asking it about an idea that I've got within the agency to do with AI and how, and how to do this certain thing and what, you know, what, what would be the next step with this and drill into the very actionable things that we can do. And have you thought about that? And looking at the competition doing this and, and it was just a half an hour. I had a few looks on the treadmill.

Richard Hill [00:08:15]: My, my wife, my wife walked past, she's like, am I on the phone? She thought I was on the phone to somebody. So she's like, who, you, who are you on the phone to? I was like, oh, I was on, I was chatting to chat GPT.

Jo Lambadjieva [00:08:24]: You know, it's amazing. Like, you know, me and my husband, like, he is, he's not very, like, he's not very big fan of AI, but like even he is like, oh, can you ask ChatGPT about this, like one thing. And so we have a casual chat about a recipe or an itinerary or whatever. It's amazing.

Richard Hill [00:08:47]: I think it's just more and more like you say when you start maybe spending an hour, two hour, three hours, however many hours a day and you, oh, we could do it for that. Like what? One of my newfound things is asking it to find working discount codes. You know, it's just like, find me a working. Everything I buy now, it's like, find me a working, it's got to be working discount code for this thing. We bought some software last week which was, I think it was thousand dollars and it saved us like $350. And that code, I, and I look and that code I couldn't find on Google, which is very interesting. But okay, so we digress. We're both, yeah, we both like to walk and talk to chat gbc.

Richard Hill [00:09:22]: That's okay. I'm sure there's some other people out there that do the same. Like you said, though, I think it's that, I think that's a great sort of analogy, that sort of sparring partner. And I think that's what it's just brilliant at, is enabling you and you know, our listeners to have a discussion with an expert in an area to get an opinion. Of course, you know, you've got to be careful what you, what you trust maybe and what's what. But it really gives you a bit of a head start in a lot of areas. So let's change direction a little bit. So I think, you know, we, we work with a lot of brands that are sort of 1 to 30 million turnover on the E commerce side of things.

Richard Hill [00:10:02]: And I think when brands hit this sort of 1 million pound mark, there's more and more sort of operational challenges with businesses there. You know, they might have been, you know, very busy getting to that point, but there's a few Things maybe going wrong around them, maybe operational, not always, but it's quite common. You know. Where do you see AI having an impact on that side of E commerce?

Jo Lambadjieva [00:10:25]: Yeah, I think again I would say AI can have impact on every single like element or process of in your organization. I think the reality is that I see AI kind of like in having three roles in an organization. The first one is to make things more efficient to save time. The second is to essentially create more data driven processes. And I think that's something that is widely under like talked about. That's not a phrase but people don't really talk about that benefit so much. They think about okay I can do more stuff and I can do them faster and it can be cheaper but actually the data benefit is massive. And third, and I think again like unpopular topic but very true is that AI is really good at augmenting your talent.

Jo Lambadjieva [00:11:26]: It's not about replacing them. Honestly it's a cliche but it's true. It's not there to replace humans. Although it is also true that by having more efficient team you might not need to hire as many people but the team you have, it will make it even better. And so like from that point of view when you hit 1 million and even if that is just you and like let's say a bunch of virtual assistants or just you, well done human AI can, can really, really help you like as I said, become more efficient, systematize and automate to an extent certain processes that are very repeatable and it will make you make better decisions by having more data to work with. Now if you have a team it's even, it's even more I would say clear benefit because you know the humans that are working on a task now have a constant copilot. Like you know I've seen, I've worked for so long in the agency world and there's always a problem of, especially when you hire someone new, there is a question of hey, this person keeps asking their more senior people where should I find this, how should I do this, etc. Well just this one use case like this case study can be lifted off by AI for majority of the tasks because you can create yourself.

Jo Lambadjieva [00:13:01]: Like and I'm not even again talking about some complex agents or like automations. I'm talking about like a simple custom GPT or cloud project where you load up a lot of your sops and like kind of like a map of where everything is in your organization and every new newcomer can like literally tap in and ask where is this how do I do this? Et cetera. And it like dives in directly into your processes. So it's, it has so many use cases and it can save tons of time and make things much more efficient for, for everyone, no matter how large your organization is.

Richard Hill [00:13:38]: Yeah, I think it's, it's brilliant you mentioned Claude there. I think, you know, we were a big fan, massive fan to be fair. We have a lot of our team have full access and whatnot. So for those that don't know, I think, you know, maybe if you, what would your sort of take on Claude be in terms of, you know, its strengths and what you would recommend to use it for? You mentioned projects there and loading data into, into a project. Maybe explain that a little bit more for our listeners.

Jo Lambadjieva [00:14:06]: Yeah, sure. I mean for anyone who has ever listened to a podcast with me talking about AI, I am a total like Claude fan girl. I literally gush. It's embarrassing. But, but in reality I think Claude is I would say one of my favorite tools for many reasons. So obviously most people would work with ChatGPT and I, I always like to compare ChatGPT to a Swiss army knife. It can do so many things. It can do so many things really well and it's obviously the most popular tool for a reason.

Jo Lambadjieva [00:14:41]: But what Claude, like, maybe, I guess like doesn't deliver in terms of having so, so many functionalities though it's getting there. It compensates with the quality of writing and the quality of thought. So for me, Claude is the most human sounding writer there is right now, like that I am aware of in terms of AI tools. And it's the most superior in terms of understanding of like semantic understanding of text. And so if you have any sort of text based task, whether it's to analyze like a particular tone of voice in your documents, or it's to analyze customer reviews, or it's to categorize particular keywords in a particular idea or to just write anything. Claude I feel is probably right now the best tool out there for that. Aside from that, obviously Claude has things like projects, which is the, if you're aware of what custom GPTs are, projects are something similar. Obviously.

Jo Lambadjieva [00:16:01]: I'm not going to go into that in too much detail, but like they just came up with the MCP sort of availability which connects tons of apps and makes like, let's call, let's call it agentic or automated AI much more like easily available for teams. And it also does something really cool called artifacts, where you can in essence build yourself like small functioning applications. That you can use in your team. So just to give you an example, yesterday I built a tiny like artifact within claude which is a, an RFP generator. So like you just fill up like very simple few lines and then it gives you an entire like four page detailed document of like requests for proposal which normally would take ages for somebody to write, but with this it's super automated and so he has so many use cases and I think, and I also like just to finish with Claude, I also really like what like their kind of their idea and ethics behind Claude and the way they are developing it. I think the it's and I think this is going to matter more and more as we progress and AI becomes more developing like how these AI systems are built, what are their morals, ethics and how they approach, let's say anything. So that, that is something also I find really much better in claw than versus like ChatGPT or OpenAI and other.

Richard Hill [00:17:42]: Yeah, it's interesting. You pick, you pick Claude over others. That's exactly literally what we, we do. You know, we've. It's. Claude is the one that we use for words and for thought and it seems to continually just out, you know, even when the new versions of, you know, ChatGPT 5 came out what a month ago, maybe less. Um, so you talked about projects, you know, and sort of, I guess the idea is to you, you create a project within CLAUDE and then you give that project information to help it deliver more specific output, words, content. Have you got any real sort of like what's, what's some of the smartest, cleverest things you've done to feed a project? You know, I know like a lot of people, like they upload, you know, turnover, voice documents, site maps and you know, look at this website, look at LinkedIn profiles and you know, that's sort of probably quite straightforward stuff, but anything that make our listeners go, oh my gosh, that is so smart.

Richard Hill [00:18:44]: What sort of things could you train or give to the project?

Jo Lambadjieva [00:18:48]: I mean this is the thing, it's like, it's not, it doesn't have to be anything like over the top smart, it's about utility. And again, I think like just before I kind of give you some examples, I think here it's really important to talk about quality and it's really important to talk about expectations and it's really important to talk about hallucinations. So you know, because you know, I get asked these questions like what is the most amazing thing? And like, you know, what about you? You do this and this. Yes. Anyone in the AI space will come and tell you that this amazing like system can solve all of your problems. You basically fire your entire staff and you go to the beach and AI runs your business and you are hunky dory. But the reality is that for anyone that works with AI seriously every day, you would know that there is serious limitations and you would know that like AI hallucinates on the regular and also it's prone to give you let's say pretty like bland generic bullshit if you don't prompt it or if you don't like essentially build it in the right way. And so I think in the end for me the more most interesting use cases or the things that I work on are things that are like actually boring, repeatable but like real true problems that take time.

Jo Lambadjieva [00:20:20]: And so I think you know, when it comes to actual examples it can be anything from like it can be in any field. Right. So it can be like your HR team like building let's say a system which both creates like let's say job specs and then like having like a system that evaluates CVS and let's say transcripts from interviews and then gives you additional cues that maybe you have missed as a human in the meeting by like seeing that transcript and seeing all of the, let's say the vocal cues. Right. And so this is one interesting way. Another interesting way is like obviously anything to do with like invoices, etc. Like you know, you can create a system which like, you know, you feed all of that, it gives you like ready sort of reports and then you feed it into another system to do. Anything to do with content generation obviously.

Jo Lambadjieva [00:21:25]: So I mean Claude is a little bit limited in terms of the text based kind of limitation. I think you know, with ChatGPT you can be much more creative because you can integrate images, you can integrate videos into your workflows. Claude I think is starting to get some sort of image capability but you know, I haven't, I haven't even paid attention because right now we have nanobanana just came out and that.

Richard Hill [00:21:53]: Yeah, yeah, yeah, that was all over the socials last couple of days. Yeah, yeah, yeah, yeah.

Jo Lambadjieva [00:21:59]: And so you know, I think anything to do with creating let's say email funnels, like you know, concepting ads on social media, like actually analyzing competitors visuals, creating like analysis and content briefs out of that and then you know, this is actually better maybe done with ChatGPT and then creating like let's say the ads. So many use cases.

Richard Hill [00:22:29]: Yeah, yeah, that I think everyone should be thinking now, you know, that's with us that you know, when you talk about interviews. So you would you then take a audio recording of an interview and, and then question the AI about that interview, about that recording.

Jo Lambadjieva [00:22:45]: Well, for example, you know, most companies now utilize like AI, like meeting recorders. And this is, this is actually great because that I like, you know, that is one of the underpinnings of, let's say a good process to actually have. Whether it's your client call or it's your like co with a supplier or it's an interview with a candidate, there is so much data that it's basically ready there. It's like transcribed for you and you can use it to automate like a lot of the media tasks. But like, for example, when it comes to an interview with a candidate, you can absolutely like, you know, feed this candidate's interview to the, the cv, the job spec and then potentially like, you know, even use like the way they express themselves, how they led the conversation to like try and like do some sort of like personality hypothesis. Like, you know, you can feed things like who will be that person managing or who will be who they be managed by and like, you know, based off the vocal cue, how they express themselves, how proactive, like how introverted, extroverted. You know, there's just so many things you can analyze and that in itself is a much richer way to make a decision versus somebody. Like as we used to do in the olden days, you have someone on a call for an hour, they might be like, let's say not in there, they might be nervous, you might be missing a lot because you're prejudiced, because they're nervous.

Jo Lambadjieva [00:24:28]: And that in itself is your own human bias. But actually if you just focus on their answers and what is the information that they convey to you, then you start seeing maybe something more than just your own human bias. Because let's say they're not the most confident and extroverted person.

Richard Hill [00:24:46]: That's a great point, I think. Yeah, a lot of people are nervous, but that's not necessarily, that's just a, A lot of people are. And after that, if you're basing the interview on that first three or four minutes where maybe they're really nervous, then you're going to maybe have the wrong opinion really, of that individual. Yeah, we did a. It's interesting you say that we, we interviewed for a role about three months ago and both of the candidates interviewed brilliantly and we were like, oh, what are we going to do? What are we going to do me and my colleague Carrie Anne, who you've, who you've liaised with. And so we both interviewed this two candidates and they were both just so brilliant and so, so, so good in slightly different ways, you know, of course. So we did, we, we analyzed, we, we, we put the CVS in. We didn't record, we didn't record the interviews, unfortunately.

Richard Hill [00:25:32]: We wish we had now, you know, and, and got some additional sort of feedback and additional sort of questions that then we asked them in a second interview. So that helped us and pushed us slightly in a direction. So it was quite useful. Yeah, it just gave us that going back to having that sort of sparring partner, that recruitment sparring partner to just to think of things to ask. Maybe that. Because you're not quite sure about something or somebody and. Yeah, yeah. So there's so many areas, isn't there? So maybe then just to put you on the spot completely.

Richard Hill [00:26:03]: So if you're building out a very smart project within Claude to do with creating content, an E Comm store, what sort of things do you feed it?

Jo Lambadjieva [00:26:13]: Okay, well so again like here would depend on for example, what is the data available in the company. And so another important point that I like to talk about and I think people don't necessarily find this topic particularly sexy, but it's going to become much more important is your data hygiene or like, you know, how do you actually start thinking about organizing your internal data and systematizing your internal data? Like you know, a lot of people talk about prompting and prompting is really important and you have to learn it. But prompting is like going to be as good as you like giving enough context about your company, your target audience, like what good looks like your, like you know, your criteria. All of this is systematized data. And that systematized data has to be done like by the organization. And so going back to what will be a really good way to build a project goes back to like how can you systematize the data to feed the project so that actually plot has as much as possible information on what it needs to do and then the actual build is not so complicated because if you have the right examples, if you have a way to like, you know, gather like, you know, as much as possible templates, information about your audience, information about what makes them thick, information about like what are they, the use cases of whatever you're selling, their aspirations and so forth, like, and so on and you are able to translate that into the best possible format for an AI to read. And then you Have a very clear criteria of. Okay, I want like, let's say we're doing email funnels.

Jo Lambadjieva [00:28:02]: On my email funnels. You want you to give me like two variations. This is what is good subject line is this is how you build and construct the copy. This is all of the information we know about. What are the topics that our audience is like, you know, keen to, to learn or listen or hear from us. Then you have a really good project. But again, it doesn't, it goes back to not just what you want to do, but how do you create the building blocks in an ongoing, processed, focused way in order to like make that project be as good as possible and make it repeatable across as many different functions of the business as possible?

Richard Hill [00:28:45]: Yeah, no, that's great. Thank you. Thank you. I think people try and I think people do try. Well, they sort of over complicate it. You explained it very, very clearly. I think so. Slight change of direction then.

Richard Hill [00:28:59]: So I think obviously you do a lot with brands around Amazon. So brands that are already sort of very successful on Amazon, doing multiple seven figures on Amazon. What sort of things are you seeing in terms of what shifts are you seeing on Amazon with sort of firms that are embracing AI and doing well with AI?

Jo Lambadjieva [00:29:19]: Okay, well, I think the something that a lot of sellers may not be aware of, but actually I wrote about this last week and it came like really to light really prominently last week was that actually now Rufus has memory. Okay, so like for people who don't know what Rufus is, Rufus is obviously Amazon's AI shopping assistant, whereby you can go in and either like try and find a product through the homepage or ask a question about the product listing that you're looking at at this particular point. And I think Rufus is something that people should not disregard because I think it's going to become ever so more important, not just for product discovery, but more so for product conversion. And so one of the main things that you know is changing right now in like the whole E commerce space, but specifically in Amazon, is that we are looking at these AI shopping systems that are helping us make decisions about our like product choices based on a conversation versus us typing keywords and browsing through many listings, which has been the way we have been shopping for the last 20 years. And because of that, like, you know, the entire way we need to think about essentially approaching this customer or this customer finding us has to change. And then now what happens is that like specifically within the Amazon remit, Rufus now has memory, which means that, you know, you can Put all the nice keywords in your listing and optimize it as much as possible for keyword richness. And that still matters, and that's still important. But now it also matters what Amazon knows about that customer, what have they purchased, how they browse, like, what other purchases they have made.

Jo Lambadjieva [00:31:29]: Like, do they know that you have a family of five or do they know that like you know, you have bought, I don't know, your toilet paper about a month ago and maybe now you need a new one. So all of this is information which now Amazon and therefore Rufus is going to use to recommend products when you are, when they are, like when customers are browsing. And so a lot is changing right now in E commerce, but a lot of change is changing on Amazon. And I think if, if I have to give one advice onto Amazon sellers is to start thinking about customer, the customer, who their customer is as the starting point of absolutely everything they do, whether it's from a listing point of view, whether it's from a like basically advertising strategy or anything else.

Richard Hill [00:32:25]: Yeah, so, yeah. So are you finding then that brands or are you advising that brands should think differently in terms of the way that they do their product listings? Because people are obviously the way that people are searching is changing. They're asking longer, more sort of detailed questions. So you've got to cover that information in the listing to therefore get bits picked up in Rufus and maybe another, you know, in other LLMs as well. You're finding, you're finding the brands that are doing more with their sort of longer tail research and longer tail descriptions and long descriptions are doing better.

Jo Lambadjieva [00:33:07]: Yeah, I mean 100. I think it's, it's just about like the nuance. So it's not about longer, it's about being contextually relevant towards the specific situation of their target audience. And so just to give you an example, I just finished working with a big marketplace agency to build the system which optimizes their clients listing for Rufus and Cosmo. So Cosmo is the kind of the, let's say the AI algorithm that underpins like Rufus. And so the focus of this product listing optimization system was how do we understand the customer in such great depth that that insight is then able to be injected within the listing. And so this means understanding obviously who their ideal customer Persona is. But that happens through, you know, firstly using first party data and then secondly using as much as possible insight from reviews.

Jo Lambadjieva [00:34:13]: Then also how do we make sure that we inject like the existing Rufus questions on the list and also the 15, like key attribute Cosmo questions. So, so it's not about the length, it's about injecting as much as possible customer focused context. And one, one piece of, of research or, or like a white paper that I would really recommend people to read is, it's called a blue, the Rufus Blueprint and it was written by Danny McMillan, Andrew Bell and Oana Poderario and they did a, essentially a white paper on the Rufus patent and they did a really, really extensive job of analyzing what actually Rufus takes in consideration when, when essentially it gives information back to the customer. And it's really, really good. And that is kind of a great, it's a very long piece, but it's very, very detailed and hopes you really understand how this thing works.

Richard Hill [00:35:24]: Well, I think we'll, we'll link that up so you guys that are still with us look in the show notes. We'll link that up and you'll be able to take a read of that and make a cup of tea by the sound of it. It's a long one, but it sounds very, very interesting. So we've covered quite a bit of ground there. But I think, I think what the big one of the big questions I sort of think about is sort of this resistance from companies when it comes to sort of change and obviously a lot of change at the moment when, when we think about, right, we're gonna, we're gonna use AI for, you know, we can talk about, you know, operations, inventory, marketing and so and so on and so on. But what would you say is sort of some of the best ways to introduce new AI workflows, new AIs into a business without sort of overwhelming and worrying existing team members.

Jo Lambadjieva [00:36:11]: I think the best way is to educate your team members. I think that's actually the first step because one like you're like, let me break this to you is that your teammates are already worried. Like, you know, you don't. You basically have to live under a stone to not be worried about what's happening. And actually people should be worried. I think it's anyone who says ah, don't worry about it is lying. Like people should be worried. I want people to be scared because, because if you don't pay att, like AI can replace you or like a not a very good like person like your professional who knows how to use AI would replace you.

Jo Lambadjieva [00:36:55]: And so don't sleep on that. So like one thing I would say is like for, for people who manage teams, the first, the first step is to educate your team to empower them to encourage them to use AI to, to create a system where using AI in their day to day is rewarded, acknowledged and appreciated. Because let me tell you, like a AI optimized process comes best from the people who run that process because they are the ones who understand what good looks like, how this thing has been running, what are the like, you know, the cranes and nooks that anyone external wouldn't know about. And so if they already know, okay, this is what AI is capable of, this is what AI like what, these are the AI's pitfalls. But I can, let's say automate this and this part and inject a bit of data by using this and this tool and this will make this process better. Then you have an AI powered team with an AI first mindset that builds you AI processes.

Richard Hill [00:38:02]: So it's about enabling your team, giving them the time, giving them the space, maybe the, some resource to be able to play, to probably play is the wrong word but within reason. So they are embracing. They are able to get excited rather than worried or scared because they realize actually you know, I'm really good at my job. But by using this, this and this, we can now be maybe 20% more efficient or a lot more. Which means then I've got time to do this, which means, you know, ultimately we may be adding 5% to the bottom line of the business which overall we no longer need to get to 50 people to do the 50 million. We think we can do it with 40 people by upskilling and training. The existing team there may be happier because they're not worried, you know, and they're, they are upskilled so they're getting maybe rewarded in terms of financially getting paid, paid a little bit more or a lot more in some instances. So yeah, I think it's an interesting, you know, I see a lot of it surprises me but still, I mean obviously we, we do this all day, this is our day to day.

Richard Hill [00:39:09]: But you know, it does surprise me. You know we speak to hundreds of ecom brands and it's a lot of them are, you know, yeah, they're, they're maybe buying certain sasses and certain, the platforms and so forth that are using AI within, you know, email and you know, different, different sort of SaaS tools but then that's almost all they're doing rather than then what can they build or who can they work with like people like yourself and us to help them build and help them, you know, embrace and train them and so forth. Because you guys do quite, you do training for Companies as well.

Jo Lambadjieva [00:39:41]: Yeah, exactly. But again, just to, to go back to that point about buying SaaS that is AI powered. Every SaaS right now is AI powered. Let me tell you. Like every, every, every, every SaaS that I've seen right now in the market has an AI somewhere in the title. The problem with that approach is that sure, you can buy it, but again you are trying to take a shortcut of a problem that will continue existing, which is your team needs to know how to prompt, your team needs to know how to add context. Your team needs to understand why data is, and data hygiene is important. That doesn't come with a button.

Jo Lambadjieva [00:40:19]: That comes with understanding and process and, and methodology. Like you know, like creating a ongoing, let's say initiatives that push your team to be constantly acquisitive and constantly figuring out what is happening with AI because AI doesn't sleep. It's like the fastest like developing thing I have ever seen. I, I seriously feel like completely overwhelmed by how much change there is in just one week in the world of AI. And if you are just right now starting, you're way behind. So you know, you have to get going. You have to get your like, your team to understand which large language model they should use for this part or like, you know, why this for example data that you Fed is obviously not getting retrieved properly. So.

Richard Hill [00:41:16]: Yeah, yeah, yeah, no, I think that's brilliant. So let's, let's think about some of these bigger, bigger merchants, you know, these sort of maybe 5, 10 million pound plus merchants that are, they're sort of implementing AI. You know, that's a very poor statement. But they're doing things. What are some of, what are some things they should really be thinking about, you know, to stay competitive over the next year, over the next sort of couple of years. They're bigger merchants. Where can they have the biggest impact with implementing AI in their businesses?

Jo Lambadjieva [00:41:54]: Yeah, I think back sort of same point, investing in your team. I think investing in data like creating sort of all of the architecture to ensure that your data is basically clean, stored in the right place and optimized for machine, for machine learning. Because like once you get to that point where you have more budget to invest, you can, you know, work with open source models, you can tune your own models, you can like essentially build much more complex applications that are particularly tailored to the needs of your business, to your customers and then it becomes an even more advanced game. I think the other thing that businesses really should focus on is understanding how search is changing. I think that's A really big topic. A lot of people talk about it, but I think it's one of those things that it's going to hit people in the face much faster than anyone expects. Only last week I, or like a couple of weeks ago, I was reading a. An article by Search Engine Land that said that 500, that AI search traffic has grown by 527% year on year.

Jo Lambadjieva [00:43:20]: And you know, that without the context is obviously like, with the context that's still pretty small number, but it's growing so fast. And what I think is really even more important is that's actually a stat by Ahrefs that I think they did a study where only half a percent of their traffic, like came from like AI, so search engines. But that traffic drove 12% of converters.

Richard Hill [00:43:55]: Yeah, a lot higher conversion.

Jo Lambadjieva [00:43:57]: And so you basically are now in this, in this world where you have AI search traffic and that is not maybe on par with anything that comes from, from Google, but the reality is that that traffic is so much more, has such higher intent and it's going to continue growing. And so if you don't figure out how to get recommended by chat, GPT perplexity, like AI, Google, AI overviews. I feel like very, very soon all that you have been working for the last, I don't know, however many years you've been investing in Google, it's going to evaporate. And so this is something really, really important.

Richard Hill [00:44:41]: It is it. It. I'm so excited about that. You know, that this massive. Massive. We talk about it so much, you know, this, you know, how many days is it away before we go to Google and there is no top 10. There is nothing there other than AI mode. That's the default.

Richard Hill [00:44:59]: Yeah, that's going to be. It could be so quick. It could just be. Or maybe not, you know, or this, maybe some middle ground, but. But obviously pretty much most days it's trying to recommend me in Google, for example, to use AI mode. So you click on AI mode and then, boom, top 10's gone, the ads are gone, you know, so then what? You know, you're relying on SEO, traditional SEO and ppc. So yeah, investing in now how to rank. Rank's probably the wrong word, but how to be visible in search, you know, on all these platforms, you know, as you said, Claude, complexity and so forth.

Richard Hill [00:45:36]: GPT, you know, when you chat to GPT like me and Joe do on your phone and ask it, you know, what do you know about my brand? You know, and then if your data is not in order you know, and you're, it's not been fed properly about your brand and so forth and you're not coming up. You know, I implore everybody listening to just sort of look what you look like in the different LLMs because, because it is, as you said, it's, it's a small part, but it's moving, it's moving, it's moving in that 1% to 2% to 5% to 10%, you know, and then the conversion rate. Absolutely. You know, they're, they're people that have asked very specific questions, have got a very specific answer if it's got the right answer. You know, the intent to buy from that search is just huge, huge, huge. So it is a semi obsession of some of my team in here. We've got a whole episode coming up on that, on that actual topic completely. So, so I'm intrigued Joe around like there's so many that you said, you know, there's so many things out there.

Richard Hill [00:46:35]: You know, we, if we sat here now and did nothing for a year and just listened to videos and we'd never get through it, you know, in terms of everything changing and new models and new versions of this and this, this new UGC tool and this. But I think it's always good to hear from other experts like yourself on their sort of maybe top one or two sort of tools that they really, really recommend at the moment. And of course I'll probably ask you in a month and it'll completely change. But you know, in terms of like ecom stores, you know, there's a lot around obviously imagery, creating creatives for meta ads and UGC and any, a couple of things around that side of things, you know, that you're seeing that are really, really strong at the moment.

Jo Lambadjieva [00:47:18]: Yeah, absolutely. I like to keep it really simple. I obviously like test a lot of tools but rarely do I see something that I think is not going to be killed by the big guys. Because all of the tools are in essence AI wrappers or most of them are. And so I like to keep it simple and to invest my time to learn how to use these tools really well. So you know, you need a subscription to ChatGPT, you need a subscription to Claude. Like these for me are non negotiable. These are the ones that I pay.

Jo Lambadjieva [00:47:56]: You need a subscription to a, an agenti browser. Now whether that's Comet or like for me is genspark and sometimes Manus like, you know, these, these tools are incredible and they can cut so much time in terms of research and they get better and better each day in terms of images and video. I would say right now, like for light, such stuff, you know, if you don't want to invest too much, you're good to go with like the ChatGPT images and so on. But if you are really looking to create like AI powered ads or conceptualize with AI or video, I would say that VO3 and the latest Nano banana model which is Flash 2.5 by Gemini.

Richard Hill [00:48:55]: You gotta love that name.

Jo Lambadjieva [00:48:57]: No, I know, it's brilliant. It's brilliant. I mean the word banana is. Anyway, great. So now we have a reason to say it all the time.

Richard Hill [00:49:04]: I feel like we just got to buy it just because of the name. Yeah, yeah, yeah. Who wants Nana Banana in the office? Oh, sorry. Yeah, no, absolutely. I think, yeah, that's one I'm very excited to try. I've not, I've not tried that yet. Obviously there's been a lot of chatter this literally this week, isn't it? I think there's a lot of, lot of chat on that. Well, Joe, I feel that we could go on for hours.

Richard Hill [00:49:30]: I think, I feel like we just scraped the surface. I think we'll, we'll look to get you back on if that's okay. But really, thank you so much for joining us today. I like to finish every episode with a book recommendation. You have a book to recommend to our listeners?

Jo Lambadjieva [00:49:43]: Okay, well, I can, I have a lot of books to recommend because I love reading books. But if you're looking for an AI focused book, I don't have one. I don't really read about AI. I read about AI online every day. Any like. So just a general book recommendation.

Richard Hill [00:49:57]: Yeah, whatever book you, you think this maybe helps you. It doesn't have to be AI related. It should just be a, you know, it could be a book on fairy tales or anything that you, you would recommend that help you in any way.

Jo Lambadjieva [00:50:10]: Right. So one book that I really recommend to anyone who hasn't read it, so it's a very well known book is Atomic Habits by James Clear. That's a really great book for just understanding what your habits are and potentially how to change them. And I just literally this morning finished listening to a really beautiful book called, called the Covenant of Water. For anyone who loves an epic long book based on history and personal narrative, that's a beautiful book.

Richard Hill [00:50:43]: That sounds fabulous. I will, I'll have a look at that personally and we'll link up both of those recommendations in the show notes. Well, thank you for coming on the show. Joe, for those that want to find out more about what you do and more about sort of AI and so forth. Obviously I know you have a newsletter if you want to tell our listeners a bit more about your newsletter and sort of best way to reach out to you.

Jo Lambadjieva [00:51:02]: Yeah, sure. So I published the AI for E Commerce and Amazon selling newsletter three times a week. As I said, it's now read by more than 45,000 subscribers. I publish it very regularly. I do it with a lot of love. It's. It was always meant to be something that I invest a lot of of my person and so and there is a lot of big fans of it. So if you want to learn how to use AI and what's happening with AI within the remit of E commerce, I definitely recommend though I'm a little biased.

Jo Lambadjieva [00:51:36]: If you want to get in touch with me, you can either connect with me on LinkedIn again I post a lot of like very actionable tutorials and sops there. My basically my long answer name will definitely lead you straight there. And yeah, you can also check out my website, Amazing Wave Digital where I basically help E commerce brands build AI powered processes.

Richard Hill [00:52:02]: Brilliant. Well, thank you so much for coming on the show and I think I look forward to literally getting you back on like in a few months time because there'll be a whole new wave but the fundamentals will remain the same as I think we've hammered home in that episode. So fabulous. Thanks for coming on the show.

Jo Lambadjieva [00:52:18]: Thank you for having me.

Richard Hill [00:52:24]: If you enjoy this episode, hit the subscribe or follow button. Wherever you are listening to this podcast, you're always the first to know when a new episode is released. Have a fantastic day and I'll see you on the next one.