Building the Right Mindset and Motivations for Successful AI in Ecommerce, with Joanna Lambadjieva, Founder and CEO of Amazing Wave

Transcript

Our transcripts are generated by AI. Please excuse any typos and if you have any specific questions please email info@digitalshelfinstitute.org.

Lauren Livak Gilbert (00:00):
Welcome to Unpacking the Digital Shelf, where industry leaders share insights, strategies, and stories to help brands win in the ever-changing world of commerce.

Peter Crosby (00:22):
Hey everyone. Peter Crosby here from the Digital Shelf Institute. As always, with the three-legged stool of people, process, and technology, it's the tech that gets all the buzz while it's getting the other two legs right that really determines success or failure. Joanna Lambadjieva, founder and CEO of Amazing Wave and author of the AI and E-commerce newsletter, invests all of her energy in helping e-commerce brands focus on enabling people to transform their processes using the right AI tooling for the job. She shares her advice four times a week with 70,000 newsletter readers and now with all of you. Joe, welcome to the podcast. We are so looking forward to diving in with you.

Joanna Lambadjieva (01:04):
Thank you so much for having me. I was really excited for this one. I was like, oh my God, we're going to talk about so many good things

Peter Crosby (01:10):
Today. Yes. And a lot of that is because you really spend so much of your day talking about AI with the folks on our ... It's focused on the e-commerce of AI, which is so great. And through that work, you help them get up to speed on how to think about using it as well as staying up to date on changes. It's all so fluid and so fast that just having outside resources can be so important. So first of all, how did you get into this and how do you spend your time every day?

Joanna Lambadjieva (01:43):
Yeah, 100%. So I have been in digital e-commerce for about 14 years. I went through a variety of roles from strategic advisor for performance marketing for D2C to diving deep into the Amazon world. And so from all of this experience, I've realized in the end of 2022 and the beginning of 2023 when ChatGPT came out that, wow, AI can be so powerful to help with so many aspects of e-commerce. And so I am a little bit of a ... I would definitely admit I'm a little bit of a geek. And so the first thing that happened when I discovered ChatGPT was, okay, how can I use this to make my team's life easier? And I started reverse engineering or what our problems were and how ChatGPT can be applied in those. And so from that process of very much self-iteration, I realized, okay, this is really powerful.

(02:50):
And I said to myself the go of, I need to learn everything I can about this. And so for the last more than three years, all I have been doing is living, breeding AI specifically in the space of e-commerce. I started my newsletter AI for e-commerce about two and a half years ago, and now it's read by more than 70,000 people. And you should

Lauren Livak Gilbert (03:13):
Subscribe if you're listening, a huge plug, because I know you won't do it, but I will do it. It is fantastic. Subscribe to the newsletter. Sorry, go ahead.

Joanna Lambadjieva (03:22):
Thank you so much, Laura. No, I really appreciate it. And so just for me, it's been always very much a self-iterative process. My approach to using AI and learning about AI has been always from the point of view of what are the teams and my problems in e-commerce and how can I apply AI? How can I reverse engineer this? And so I write about this in the newsletter and that's what I basically help brands with is train teams, build these processes and solve real business issues through the lens of AI.

Lauren Livak Gilbert (04:00):
And so you're working with brands day in and day out, Joe. And I'm curious, how are you suggesting brands approach AI or think about it? Because it's a big change. It's not their normal way of thinking and it's easy to get stuck in, "Oh, well, this is the old way that we've done it. " So how are you helping them change their mindset about it?

Joanna Lambadjieva (04:21):
Yeah, 100%. I think we are in a really interesting inflection in time is a lot of the old processes and ways of things or ways we've done things have to be destroyed. They have to be disrupted. I know that sounds really like a mega announcement, but even just what happened last week with Amazon announcing Alexa for shopping, it just shows that even the big guys are there to self-disrupt themselves because they know that this big shift is coming. And so I think the first thing that I think both people and brands and basically anyone working in e-commerce should start thinking about is using AI as a thinking partner rather than replacement for thinking. Now this is, I know it sounds probably simple, but is a really fundamental shift in how you approach using AI. And I think a lot of companies are trying to onboard tools that are solving an issue, but what I see happening with people adopting these tools is they are not necessarily using them as kind of like a help, but rather as something that they just diverge work to without the necessary oversight in terms of quality.

(05:42):
And therefore you get subpar quality that a lot of people think is done and then you figure out it actually is just mediocre. And so a wave of issues happen. And so the first thing that I would say is that people should really think about how they use these tools, what they use the tools for. And then from that point is like building an entire, let's say, knowledge basis of context of how they do processes of what good looks like and then using AI with that context to get better outputs.

Lauren Livak Gilbert (06:22):
Do you have any examples of a brand that's kind of done that well or really shifted their mindset from not using AI or using it in the wrong way to using it in the right way?

Joanna Lambadjieva (06:33):
Yeah, 100%. So actually about a year ago I had the awesome occasion to be on a podcast with the, I think it was the head of e-commerce at Havayanas and just the way that they have adopted AI. Again, it's all about the mindset. It's not so much about the tooling, but the mindset internally of how to think about processes, how to think about upskilling their teams was just already like that was early 2025 was ahead of the curve. So I think that for me was one great example.

Peter Crosby (07:11):
And can you dig into that a little bit? What was it they were doing that really impressed you so much? Because you talk about as thinking partner and I understand the concept of it, but the actual, how do I put that

Joanna Lambadjieva (07:26):
Into

Peter Crosby (07:26):
Action in daily use? Yeah.

Joanna Lambadjieva (07:30):
So I think the reality when it comes to really digging deep in terms of what I guess any brand does, and obviously I can't dig deep specifically for Avayas, but what brands do well when they adopt AI is starting from the process. And again, this is very counterintuitive because so much of the conversation that we have right now on LinkedIn and conferences, et cetera, starts with the tool. Obviously we are all very excited about Cloud and Cloud cowork and CloudCode, but of course these are great tools and they allow you to do a lot, but for businesses, brands and organization, the first step of AI adoption comes to what is your process? Have you documented your process? Do you understand your process so well that if you sit down and have to explain it to an intern, you can really break down every single input, every single output, every edge case, all of the data that needs to be included.

(08:35):
And then basically if you have that, can you then translate it into chunks that AI can process? Because if you start thinking about your usage of AI through this lens of a perfectly explained process, then you've done pretty much 80% of the job by explaining very well what the process of how you want the job to be done, then any tool will do a very good job because essentially AI, large language models, they thrive on context, but that context can only be explained by the person who knows what the context is. My context will be different than your context or your context. My company will have different objectives or different how good looks like depending on each edge case and even in terms of teams. And so it's basically down to the specific team, the specific individual, a specific team role to sit down, codify this knowledge, this criteria, and then use AI to help with the actual implementation, but only from that point of structured knowledge.

(09:51):
Yeah,

Peter Crosby (09:53):
We've been thinking about that a lot and the white paper that Rob and I recently wrote was talking about the context layer for brands that are marketing in order to represent all the use cases of their products and everything, that you need this new layer of content that adds the context. And in essence, you're talking about somewhat a similar concept but aplied to your own work with the AI providing more and more. We often talk about the new creative brief and in a way it's sort of doing a creative brief for your AI for whatever it is you're trying to accomplish.

Joanna Lambadjieva (10:34):
100%. I think again, we are really stepping right now in a really interesting time where work is really a mixture between human and machine input and output. I think we need to start thinking about not AI as just a tool that is kind of the last sort of step of the chain, but it's sort of like a partner along the chain. We talk about human in the loop when you talk about a lot about AI a lot, but it's true when we start thinking about an entire process and what is the place of AI and what is the place of human, which is basically overseer, thinker, creative mind and AI as, again, a helping thinking buddy or a creative brainstorm agent or an executor, you start building a bit of a picture of how we should start redefining briefs proces. And to go back to your point, Pete, it is about building libraries of context That is part of the process.

(11:47):
That is part of these structured SOPs that you would give to an AI because as much as a process is step one, do this, step two, do this, part of this is basically giving the context of this is who we are and this is what is our special source of doing things, what is our special process? So they are totally interconnected and I think it's so important again for all companies to start really thinking in this more structured way of this wealth of knowledge they have internally and then translating that into wealth of data that they can feed to AI to help their processes.

Lauren Livak Gilbert (12:34):
And Joe, have you seen any brands update their briefing process or automate it using AI? Because that's something that we've talked about on this podcast several times. It's something I'm very passionate about. The brief needs to be updated. It has been like a siloed process for many, many years. So I'm curious if you've seen any brands do that well and it's working.

Joanna Lambadjieva (12:56):
Well, here's again the challenges. I feel like with smaller brands, smaller teams, it's a much easier process to adopt. Some of the brands I work with, they're very, very quick to iterate their processes even like part of their team structure in order to adapt into this AI age and obviously they don't have many layers of approvals or let's call it bureaucracy that they have to go through. I think for bigger brands, it's definitely a challenge because obviously there's an established process and this process has worked and served them well probably for a long, long time, but now AI is just so disruptive and also there's just so much uncertainty like lack of knowledge, lack of understanding, like fear. And so all of these are ingredients as to why maybe in some of the larger organizations I speak with, there's more of a piecemeal adoption or change versus let's say an entire process or let's say an entire team.

(14:17):
So I think the size of teams and the maturity of organizations plays a really big role

Peter Crosby (14:26):
And I think that one of the primary things that has to happen to even get there is that there has to be adoption within the organization and that's still I think a barrier that people are trying to get over. And so I'm wondering, when you look at the brands that are doing this well or when you're working with a brand where you focus them in order to drive adoption, what's the number one thing that comes up or that needs to be instituted to get that engine going?

Joanna Lambadjieva (15:01):
I think the number one thing that I always try to focus the brands I work with is the fact that AI is not a technology problem or AI adoption in the company is not a technology problem, it's a people's problem. It's basically a psychological problem. And this is a really challenging thing for a lot of leaders to understand because I am privy to it. I am really excited about AI. I mean, obviously I am and loads of leaders go to conferences, they get all hyped up. They're like, "Oh my God, AI can do all of these things." I have to get everyone to do AI all the time twenty four seven, everyone AI right now and I get it. I mean, I feel the same, but the problem is that when you try to get your team, the wider organization to adopt AI, you then face so many psychological challenges.

(16:05):
One is there's just so much at the moment mystery around AI. Loads of people say it can do everything. You can basically press a button and sit on a beach and drink margaritas. Guess what? Not quite a reality. We

Lauren Livak Gilbert (16:22):
All wish, don't we?

Joanna Lambadjieva (16:24):
Yeah, I know, right? Maybe at some point, but not right now. And so loads of people feel like they don't necessarily understand what is and is impossible. Therefore, either you have naysayer or you have enthusiasts that are not quite getting there. I think it's a huge thing. Loads of people who feel the presure to use AI, but they don't know where to start or how to start and they fear that they will look unknowledgeable and experienced. The common fear we all have, we don't want to be the person that doesn't know the thing in the room. And then I think one of the biggest fears that people don't necessarily understand how strong it is, is the fear of replacement, the fear that the organization is here and they're pressing you to use AI so that at some point they'll replace you with AI. And so because of all of this and probably a lot of other human reasons, you have this kind of pressure, this juxtaposition, which is like the leadership and a few enthusiasts are like, "Yeah, let's do AI." And then the massive people, the middle and lower layers of organization are kind of not sure where to start, not sure what that means for their future.

(17:50):
And so there's a really big mismatch. And so what I always say to companies is before you start saying, "This is how you use AI," just first get people to really understand what AI is, what AI can and cannot do and reassure them that this is there to help them to assist their work rather than replace them. That is so important. Get them to get excited. This is really important. I always in my training at camps, I always say like, "Hey, you guys are scientists or explorers, go and build things, break things, see what is possible." And then people get excited because they start using AI in really creative ways and they see, okay, it can definitely help me with a lot of things, but it's definitely not going to replace me. And I think this is really important. When you approach AI adoption from the perspective of your team and how to get them onboard rather than something they have to do, that completely changes the mindset about using AI.

(19:00):
And it's not about leadership is forcing me to use AI because I have to. It's like, "Oh wow, AI is so cool and I can build so many things to make my life easier and I can be a better employee and this is an investment in my future." And that's a really big shift.

Peter Crosby (19:15):
Yeah. A couple of things sort of roiling around in my head as you were speaking. I think one of them is that it should never be let's use AI. It should be, what are our goals? And you should be setting goals that are only possible with AI because now there's so much more that's possible when you apply AI to a process that's going after a business goal. So always having that in mind and maybe focusing people on what is the goal we're trying to achieve and where does AI fit into making that possible. And the second thing is we were talking about this a lot at the Digital Shelf Summit is if in fact at some point AI is able to take away, I don't know, 60, 70, 80% of the drudgery of e-commerce. So you don't need to be in there changing e-commerce descriptions or trying to figure out what the changes are from retailers and what to do next.

(20:14):
The AI can recommend the next best action, but you get to focus on where you can actually have a business impact that's not simply and efficiency is important, but actually about growth because where the growth is coming from is going to be the hard not to crack in this coming five years or wherever. And I would imagine that if it can be felt, if you can do some sort of pilot project that just gives the glimpse of what that freedom could be given to you so that your job is more about whatever you want to call it, being the growth architect rather than necessarily the PXM champion or something that's what gets me excited is about the opportunity to actually have a promotion out of this rather than a replacement.

Joanna Lambadjieva (21:19):
Yeah, a hundred percent. I think what I get really excited about right now is the opportunity to really diversify each and every person's skillset and the ability to diversify the way you think. My sort of mantra for 2026 is if you can imagine it, you can probably build it and that applies both to work and non-work projects. But if you think about you can literally vibe code right now a solution to your day-to-day problems and just imagine how much that develops your way of thinking about solving a problem. And so it's really about seeing each of one of us individually, seeing what an incredible opportunity it is for our own self-development to completely diversify the way we think, diversify our data inputs in terms of how we make decisions, like build our own much more elaborate flows of gathering information, learning, and then giving superior quality of work.

(22:34):
So I think it's really exciting really.

Lauren Livak Gilbert (22:38):
I agree. The vibe coding thing is also amazing because you can go and you can create something that can solve a problem. So it is a very, very exciting time. And you've been talking a bit, Joe, about barriers that brands are seeing and a big piece is mindset and ways of thinking. Are there other barriers, whether it's executive leadership really investing in AI or change management, what is the biggest barrier outside of mindset that you're seeing from a brand perspective that maybe they can tackle to get past it?

Joanna Lambadjieva (23:10):
Yeah. I think two things is speed to change. I think reality is that if companies want to really adopt to AI, they really need to realize that AI is not a project that you can map out in the next three to five years. AI is literally moving and that's why I write a newsletter four times a week because it's moving so fast that if I don't, I will fall behind. And same thing applies to companies. If they don't realize that this is not something that they sit and they plan for ages and then they get another approval, you have to have a certain speed to market if you want to market or to team with which you test and break things and see what works. And you can probably do it in little mini test groups internally before you roll it out to the wider team, but you have to get into the mindset of a startup or build that within the team because even when we talk about, for example, to adoption, this is a common challenge, especially for larger organizations.

(24:31):
There is, let's call it a procurement process in which a tool has been selected and it's rolled out across the entire organization. But a lot of the times maybe this tool is not the best out there in the market or doesn't fulfill the needs for each and every individual team. And then you get to the point where because this tool is just not good enough or not fit for purpose, people get disillusioned or they get turned off and that poisons the well. And then you don't want that. You want people to have the opportunity to use some of the best tools. You can definitely figure out what is your AI policy. I mean, every company should have an AI policy of what data should and shouldn't be used, et cetera, but don't hamper the ability for people to really get real use out of these tools because you've just given them the tool that everybody's kind of comfortable with.

(25:33):
So I think speed and then tooling is really important. And I think just, yes, changing completely the mindset about AI. AI is really the most insane catechismic change that I think is going to happen in our lifetimes. I feel like people should really take this seriously and see it as one, a massive opportunity, but then also massive disruptor. If you don't disrupt your processes, your team, your mindset about how a company is run, how teams are structured, how processes are run, how we learn, how we train and how we motivate people, I think there will be a lot of companies that would and you'll fall behind.

Peter Crosby (26:21):
But don't be scared.

Joanna Lambadjieva (26:23):
No, everything's fine, but get your shit together.

Peter Crosby (26:30):
So to go to the tooling for a moment, and I know this probably will make this podcast dated in 60 days or something like that, or maybe even less, but what are the two tools that right now you are most focused on and you think the brand leader listener that's listening right now should be considering?

Joanna Lambadjieva (26:58):
Yeah. I mean 100% with you, this will be very outdated very soon. We'll have you

Peter Crosby (27:05):
Back.

Joanna Lambadjieva (27:06):
Yeah. Yeah. I think I just want to do a little segment to talk a little bit about why I really believe outdating of tools. Just to give you a little bit of a timeline and then I'll tell you about the two tools. About let's call it conservatively. 12 months ago, everybody was using ChatGPT and that was the only tool that anyone ever cared about and there was nothing else but ChatGPT. Then came out maybe like nine months later came out Gemini and Gemini was, oh my God, the tool that is basically changing AI forever, no one else should use anything else but Gemini. Now fast forward six months ago, then Claude, which by the way, I have been banging on about Cloud for about two and a half years, but at that point everybody caught up that Claude is the best tool and oh my God, you cannot live and breathe AI if you don't have a full Cloud subscription.

(28:11):
And so I just want to point out that again, tooling advice is a bit, it's difficult because these companies are constantly competing to figure out the best models, the best tooling, the best use cases. So on little tip before I say which tools is again, think to agnostic, build your processes, build your context libraries outside of any tool, like make sure you codify this outside in a central location in your organization. Don't make it live in the tool because then whatever tool is the best at the moment, you can switch that much easier than building everything within one tool and then be like, "Oh no, this is now crap." So to answer the question, my tool that I have been using literally religiously for two and a half years, basically since 2023, so three years now is Anthropic Scloud. I think it's a amazing tool because although it doesn't maybe do everything, though now it's getting to that point, I think what Anthropic does really well is go deep in terms of how businesses should solve problems.

(29:29):
And so for example, on great function that they built about six months ago or a little more was skills. And this is exactly what I was talking about is codifying the way you do a process and then AI reading that and executing in this way. Again, it's a concept and that concept you can apply to whether you use Copilot or Gemini or ChatGPT doesn't matter, it's the concept, but they managed to package it And build it within the functionality in a really smart way. And I also think firstly, Anthropic does things from an ethical point of view much better than I think most other companies do, which I think is important also to think about. And I also think the way that their model thinks, let's call it things and reasons through problems is far, far, far superior than all the others that I've tested and I've tested them and I test them all the time and it's very subtle and I can't explain it, but it's just much smarter.

(30:35):
For anything creative-wise, I would say right now the best model is for anything creative-wise in the space of e-commerce, let's call it, is Google Gemini. So if you want to build any flows that include any sort of visual output, I think Google's Gemini is your best bet.

Peter Crosby (30:57):
And I'm imagining particularly when you go into larger brands, they're very restricted in terms of they can't do a lot of tool hopping because if it's a Microsoft shop, I don't know. I'd love to get your perspective on that. How much flexibility is there and do you feel like larger brands will need to become more flexible to really make this happen or is that just not going to happen because it's global IT and they're a Microsoft shop or whatever? What are you finding out there?

Joanna Lambadjieva (31:32):
Well, what I'm finding is that yes, I think Microsoft Copilot is definitely the preferred solution, obviously for many security reasons, for very good reasons for a lot of large organizations. I just feel from testing it and from everything I've read about it and all of the peer feedback I've received, it may not be quite on par with some of the other players yet in terms of capability. And this is going back to what I said about risk of poisoning the well is a lot of employees just end up being frustrated by being encouraged to use AI day-to-day and not being given the right tooling in order to do their roles really well. And so I think it's a challenge because I think again, we go back to the big organization problem.

(32:37):
We have a way things have been done and it's difficult to make that change in a desirable speed. But I also feel that if employees, if the leaders of these organizations don't start thinking about competitive edge through AI and through capability, then again, there will be much more nimble agile companies who would. And so again, I'm sure that Microsoft Copilot will continue investing and developing. It's just the question is how fast and if we are talking about May 2026, is this the best tool you can use right now to build maybe more sophisticated workflows, maybe not. And so the question is, can the organization think about processes to slowly roll out more enablement? But that is-

Peter Crosby (33:47):
Yeah.

Joanna Lambadjieva (33:48):
And

Peter Crosby (33:48):
Sorry, I didn't mean to put you on the spot necessarily specifically about Microsoft, but they are the one that comes to mind

Joanna Lambadjieva (33:54):
When you think of

Peter Crosby (33:55):
Big brand IT. And when I think about, this is a moment in some ways like no other, but when you think about the big technological shifts over the last several decades, at some point the business requirements overwhelm the conservative nature of things. So as always, our community are the advocates for change in their organizations and it's a hard job, but I think anchoring it to outcomes and being diligent about proving that and I think that will gain power over time. It's just will it gain power in time enough. So I guess that's what we're going to find out.

Joanna Lambadjieva (34:50):
Yeah. I guess it's really about trying to build small case studies internally and then the proof is in the pudding. It's always about that. So building enough proof to, I guess, let's say ask for a change.

Peter Crosby (35:14):
Prove the value. Yeah, yeah. Yeah. It's one of my most favorite things to do is to try similar things across a couple of options and just see what the difference is in terms of it being a thinking partner and also is it easy to store context in the way that you're talking about? It's really hard when I feel like I need to remind the AI every time I start a new chat and I feel like I'm starting from scratch because there isn't that infrastructure for locking all of that in. Or maybe I don't know about it in some of the tools that I use, but it does feel like Cloud has that kind of sorted out, which is pretty cool.

Joanna Lambadjieva (35:53):
Yeah. But again, context always should really live outside of the tool. The chat. But here's one little trick or tip that I can think- Oh yeah, let's

Peter Crosby (36:07):
Close with a trick. I love it.

Joanna Lambadjieva (36:08):
Yeah, let's close with a trick. So on tool that I really love, and this is not an AI tool, is actually like I would call it more of a productivity tool is called WhisperFlow and WhisperFlow is really fantastic, fantastic tool. It's very simple. It's basically a voice transcription tool, but the trick here is that it would transcribe what you are saying in any environment, whether it's a Slack or WhatsApp or an email or an AI model. And the beauty here is that when we, for example, sit down and type something, that takes time. And if we have a lot to think, to say to type, sometimes you get lazy. And specifically with AI, there's this itch to press the enter button just to get the thing done. And that is obviously where the lack of context gives you mediocre results. But if you talk and if you just go through with your flow of thoughts and you have an interrupted flow of thoughts and that is the context you give, you immediately alleviate that barrier through the typing and you improve so much the results.

(37:26):
And so I've found that just generally not just by using AI, but just even typing a WhatsApp message to my partner, I'm so much better at giving enough context so that the communication flows well.

Peter Crosby (37:43):
That is amazing. Whisper flow.

Lauren Livak Gilbert (37:45):
Applies to all elements of our lives.

Joanna Lambadjieva (37:48):
100%. Communication is the pivotal thing of everything that we do, I think. So

Peter Crosby (37:54):
Joe, I have to ask as I let you go, anyone who's listened to this podcast now is going to want to subscribe to your newsletter. So what's the best way to do that?

Joanna Lambadjieva (38:04):
Sure. Well, firstly, you can go and find my very long story name. I'm the only one to run a Lamborghini on LinkedIn and you can connect with me. Yes, it's not a very cashy name. No,

Peter Crosby (38:17):
It's a beautiful name. It's just hard

Joanna Lambadjieva (38:19):
To find. Thank you. Thank you. Of course though.

Peter Crosby (38:22):
Yes.

Joanna Lambadjieva (38:24):
And you can connect with me on LinkedIn and there's direct link to my newsletter there or you can literally type in aiforcommercenewsletter.com and that is where you can also find some of the previous issues. So again, don't subscribe if you don't like it. First have a read and then subscribe.

Peter Crosby (38:42):
Love it. Well, Joe, we are so grateful to have your brain and your four times a week newsletter and everything sort of working on behalf of our industry and also, of course, for coming and sharing it with our community. We're really grateful.

Joanna Lambadjieva (38:56):
Thank you so much. And it was such a nice chat. I really enjoyed the guys. Thank you

Lauren Livak Gilbert (39:00):
So much for having me. Thanks so much, Joe.

Peter Crosby (39:02):
Thanks to Joe for all the insights and for building really strong stools. Great advice as always on hand at digitalshelfinstitute.org. Make sure you remember. Thanks for being part of our community.