WEBVTT

00:00.511 –> 00:05.054
[SPEAKER_02]: Two years ago, everyone on conferences said, OK, in two years, everyone will buy through a shopping agent.

00:05.174 –> 00:05.994
[SPEAKER_02]: Now, here we are.

00:06.034 –> 00:07.135
[SPEAKER_02]: Still no shopping agent.

00:07.395 –> 00:13.198
[SPEAKER_02]: Shed GPT has announced it’s like so prominently and has completely pulled back from kind of becoming a shopping agent.

00:13.298 –> 00:16.259
[SPEAKER_02]: So people still go on Google and still serve on Instagram.

00:16.279 –> 00:21.082
[SPEAKER_02]: So Google and Meta are remainder by far largest traffic channels for all online shops.

00:21.122 –> 00:21.742
[SPEAKER_02]: I know what, at least.

00:24.448 –> 00:26.851
[SPEAKER_01]: It’s the e-commerce master plan podcast.

00:27.392 –> 00:31.477
[SPEAKER_01]: It is a help you solve your marketing problems and grow your e-commerce business.

00:31.978 –> 00:39.829
[SPEAKER_01]: Cutting through the hinder to bring you inspiration and advice from the e-commerce sector and beyond, here’s your host, Chloe Thomas.

00:43.156 –> 00:45.498
[SPEAKER_00]: Hello and welcome, it’s great to have you here.

00:45.558 –> 00:49.261
[SPEAKER_00]: Thank you for hitting play and choosing to listen to one of our inspiring guests.

00:49.781 –> 00:56.006
[SPEAKER_00]: In this episode, I get to chat about a bigger picture AI with our super knowledgeable guest.

00:56.367 –> 00:58.388
[SPEAKER_00]: We’ll be getting into three key areas.

00:58.608 –> 01:02.612
[SPEAKER_00]: I think the three biggest question mark areas of AI in the commerce right now.

01:03.032 –> 01:06.114
[SPEAKER_00]: The ones you need answers to, the ones we’re all deliberating about.

01:06.675 –> 01:08.797
[SPEAKER_00]: First off, we’ll be talking about AI content.

01:09.457 –> 01:12.121
[SPEAKER_00]: what should your goal be with the AI producing content?

01:12.241 –> 01:16.928
[SPEAKER_00]: What’s possible with AI content production prepared to have your mind blown on that one?

01:17.669 –> 01:21.834
[SPEAKER_00]: We will be looking into as number two, agentic transactions.

01:21.995 –> 01:23.337
[SPEAKER_00]: Is it coming anytime soon?

01:23.397 –> 01:24.118
[SPEAKER_00]: What do we think?

01:24.578 –> 01:30.559
[SPEAKER_00]: And my guest has some really practical thoughts on that, which I have said you’ll thoroughly agree with.

01:30.919 –> 01:39.341
[SPEAKER_00]: And then our third topic is going to be internal agent consolidation by which I mean, why are we building lots of data sets and lots of separate agents?

01:39.721 –> 01:40.761
[SPEAKER_00]: We’ll be discussing that too.

01:41.341 –> 01:50.743
[SPEAKER_00]: Make sure listen to the end of the episode, because my guest has some great top tips and we get into the nitty gritty of second hand and fashion in that section as well.

01:59.883 –> 02:01.785
[SPEAKER_00]: And now to introduce our special guest.

02:02.226 –> 02:09.816
[SPEAKER_00]: Tarric Miele is the co-founder of Fashion Online Shop about you and co-founder at Digital Commerce Engine Scale.

02:10.156 –> 02:11.137
[SPEAKER_00]: Hello, Tarric.

02:11.598 –> 02:12.759
[SPEAKER_02]: Hey, Chloe, thanks very much.

02:13.460 –> 02:15.082
[SPEAKER_00]: Awesome to have you on the show.

02:15.343 –> 02:17.626
[SPEAKER_00]: How did you get into the world of e-commerce?

02:18.330 –> 02:20.911
[SPEAKER_02]: Oh, that started pretty early at the age of 15.

02:21.091 –> 02:22.791
[SPEAKER_02]: I started my first online shop.

02:22.851 –> 02:27.693
[SPEAKER_02]: At the age of 13, I was building websites and then figured that mostly online shops advertised on my websites.

02:27.733 –> 02:30.914
[SPEAKER_02]: And then started my first online shop at the age of 15.

02:30.974 –> 02:34.695
[SPEAKER_02]: I’m now 37, so 22 years of online shop experience.

02:34.835 –> 02:37.916
[SPEAKER_02]: And then, yeah, I started out with building a lot of niche online shops in different categories.

02:38.556 –> 02:41.417
[SPEAKER_02]: And then in 2014, founded about you.

02:42.307 –> 02:47.791
[SPEAKER_00]: And what is, I mean, I glossed over about you in a way that doesn’t know just this at all in that intro.

02:47.811 –> 02:51.815
[SPEAKER_00]: So can you tell the audience a bit more about you if they’re not familiar with it, please?

02:52.335 –> 03:00.301
[SPEAKER_02]: Yeah, I’m probably there are familiar if they are from Continental Europe and they are probably not familiar from the UK or US, because we’re operating 26 countries.

03:00.602 –> 03:02.683
[SPEAKER_02]: We’re covering all the Continental European countries.

03:03.304 –> 03:07.888
[SPEAKER_02]: I mean, if you’re from the UK, similar to a, so I would say, or from the US, it’s probably revolved.

03:08.628 –> 03:16.192
[SPEAKER_02]: We are selling fashion, we have our own private label roughly 10% of the revenue and 90% is third party brands.

03:16.252 –> 03:19.834
[SPEAKER_02]: So I think Nike already does kind of all the mass market brands you can imagine.

03:20.394 –> 03:25.377
[SPEAKER_02]: We sell 5,000 different brands and 1 million different SKUs that we stock.

03:25.937 –> 03:31.239
[SPEAKER_02]: And it’s a mix of private label, wholesale, retail, basically, and marketplace that we operate.

03:32.159 –> 03:34.220
[SPEAKER_02]: We had roughly 3 billion in GMV.

03:34.900 –> 03:36.081
[SPEAKER_02]: So revenue after returns.

03:36.261 –> 03:37.081
[SPEAKER_02]: Yeah, profitable.

03:37.621 –> 03:44.364
[SPEAKER_02]: We took the company public in 2021 and took it private in 2025 because we merged with with much with Zalando.

03:44.544 –> 03:49.966
[SPEAKER_02]: So technically about US, well as scale, bidong to the Zalando group now on Zalando’s the largest

03:54.773 –> 03:55.213
[SPEAKER_00]: Nice.

03:55.233 –> 03:57.635
[SPEAKER_00]: Yeah, we could do a whole episode just on that.

03:58.855 –> 04:13.023
[SPEAKER_00]: But we’re not going to today and tell us a little bit about scale as well because you’re not only did you find this super successful fashion business, but you’ve also been instrumental in the founding of scale, which is a very different business.

04:13.363 –> 04:38.865
[SPEAKER_02]: But very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very, very,

04:41.707 –> 04:45.749
[SPEAKER_02]: So 2018, we started with this, took us three years to really build a B2B organization.

04:46.329 –> 04:50.731
[SPEAKER_02]: And since 2021, your company scale exists and had been carved out.

04:50.872 –> 04:54.994
[SPEAKER_02]: It’s the proprietary software that was powering about you and still is powering about you.

04:55.014 –> 04:58.715
[SPEAKER_02]: So about you remains, I re-bought you about you as a scale customer now, technically.

04:59.176 –> 05:02.977
[SPEAKER_02]: But yeah, we are very fortunate to win a lot of customers worldwide.

05:03.018 –> 05:05.559
[SPEAKER_02]: So today, 300 online shops are using scale as they are

05:06.119 –> 05:23.296
[SPEAKER_02]: shop software slash commerce engine, a lot of big customers in the continental European region like Daichmann, Snipes, as Oliver, in the fashion world, but also in other Jason categories, a few months, the largest optician, for example, the baby mark, baby, but it’s the largest baby supply.

05:24.037 –> 05:26.779
[SPEAKER_02]: And then two years ago, we started three years ago, we started in UK.

05:27.500 –> 05:32.563
[SPEAKER_02]: with customers like Manchester United, for example, our heritage, complete Iran on scale and also many more.

05:33.183 –> 05:45.049
[SPEAKER_02]: And a year ago we started in the US and had very cool for signings, Levi’s globally, for example, is part of the scale bedbath in beyond group, including container store.

05:45.609 –> 05:49.031
[SPEAKER_02]: And then many more, some of them were a lot of them actually have not been announced.

05:50.071 –> 06:00.614
[SPEAKER_02]: So on average, the skate customer does roughly 250 million and you revenue and we have everything from 100 million up to multi-billion dollar revenue companies.

06:00.654 –> 06:07.735
[SPEAKER_02]: So we start, we are Shopify and the other SMB systems end and therefore playing the enterprise, SaaS field.

06:08.115 –> 06:14.457
[SPEAKER_02]: And according to God, now I think we were the fastest growing software vendor in the e-commerce space, consecutively in the last five years.

06:18.746 –> 06:24.951
[SPEAKER_02]: Yeah, we are a little bit under 100 million in AR, so which for a softer company is quite sizable, very profitable business.

06:25.772 –> 06:29.154
[SPEAKER_02]: And it’s a second to your subsidiary of Scott Elf about you.

06:29.935 –> 06:30.575
[SPEAKER_00]: Very cool.

06:30.635 –> 06:32.477
[SPEAKER_00]: It’s a pity you’ve got all of that on.

06:32.517 –> 06:44.646
[SPEAKER_00]: Thank you so much for being up for coming on the podcast to talk about, I think kind of three things which I think most people in e-commerce, they’re being brutally honest, are,

06:45.752 –> 06:50.237
[SPEAKER_00]: worried about thinking about, got playing around at the back of their brains at the moment.

06:50.317 –> 06:53.100
[SPEAKER_00]: So I’m really appreciate you for bringing this topic to us.

06:53.601 –> 06:58.106
[SPEAKER_00]: We’re going to be talking through the three critical ways AI is changing e-commerce in practice.

06:58.647 –> 07:04.353
[SPEAKER_00]: And the first of those is AI content as a growth lever.

07:05.194 –> 07:08.179
[SPEAKER_00]: what should be our goal with AI content?

07:08.219 –> 07:12.465
[SPEAKER_00]: Because I think a lot of people are using it as cost saving and efficiency rather than growth.

07:12.525 –> 07:15.810
[SPEAKER_00]: But you’re very firmly of the opinion it should be a growth leader, right?

07:16.458 –> 07:18.539
[SPEAKER_02]: Yeah, I know, I think it’s a journey.

07:18.599 –> 07:23.741
[SPEAKER_02]: So for us, the usage of AI and content started out as a cost measure to be honest.

07:23.761 –> 07:27.283
[SPEAKER_02]: So at about you, we have invested up until two years ago.

07:27.323 –> 07:31.325
[SPEAKER_02]: We have invested roughly three million euros per year in editorial content.

07:31.345 –> 07:34.126
[SPEAKER_02]: So this was basically the footage when you enter our website.

07:34.146 –> 07:37.767
[SPEAKER_02]: You see, whatever, find out the cool waiting outfits.

07:37.787 –> 07:38.528
[SPEAKER_02]: And then obviously you have,

07:39.108 –> 07:41.910
[SPEAKER_02]: wedding pictures, yeah, to kind of get the custom and the mood.

07:42.491 –> 07:47.415
[SPEAKER_02]: So this we would consider editorial content or styling recommendations, et cetera.

07:47.835 –> 07:56.402
[SPEAKER_02]: And then we had an additional roughly 7 million and that we have invested by operating our own photo studio and where we’ve produced the product detail page content.

07:57.583 –> 08:07.769
[SPEAKER_02]: And basically, when we got a product, we took a bus, pick charts, like the flat lane way, and then we also produced model pictures, because this is always a big conversion rate driver.

08:08.390 –> 08:20.657
[SPEAKER_02]: From the 1 million SKUs that we stock, roughly 200,000 per year, on new and roughly 100,000 of the 200, we have to shoot and the other half, we get the imaginary from the brands, like Nike Adidas providing us the content.

08:20.677 –> 08:22.258
[SPEAKER_02]: So, we had to shoot 100,000 new SKUs,

08:24.299 –> 08:27.401
[SPEAKER_02]: per year and that has cost us roughly 70 euro per scale.

08:27.441 –> 08:33.464
[SPEAKER_02]: So it’s up to 7 million in product-eater-page content expenses and 3 million in editorial.

08:33.864 –> 08:43.849
[SPEAKER_02]: And to be very honest, so that’s a 10 million euro cost line and to be honest two years ago my view on AI was okay cool if we can cut the 10 million by half.

08:44.569 –> 08:46.850
[SPEAKER_02]: I’m going to say five million, that’s a great thing, let’s do it.

08:47.790 –> 08:51.091
[SPEAKER_02]: And so it started out as a way to reduce cost.

08:51.751 –> 09:00.094
[SPEAKER_02]: And then, obviously, we always did a BTS kind of, so we did reword production, and we produced it with AI, and then just a BTS, the things.

09:00.774 –> 09:10.938
[SPEAKER_02]: And I think a year ago or so, maybe a 50, 50 months ago or so, we were at a point where AI and reword had the same performance, but AI at minus 90% of the cost.

09:12.158 –> 09:19.380
[SPEAKER_02]: So, this for us was the point where we said, okay, we’re going to switch to AI content and we’re going to cut the whole real-world stuff.

09:20.040 –> 09:27.142
[SPEAKER_02]: That meant we completely cut it to 3 million editorial budget to basically zero, just token costs, but no real-world production.

09:27.562 –> 09:29.623
[SPEAKER_02]: And we decided to close down our photo studio.

09:29.883 –> 09:31.763
[SPEAKER_02]: And this was roughly 15 months ago.

09:32.283 –> 09:34.184
[SPEAKER_02]: Now then, what happens really interesting

09:34.944 –> 09:38.645
[SPEAKER_02]: So closing down the photo studio can’t do overnight because it’s a big operations.

09:38.665 –> 09:40.045
[SPEAKER_02]: I mean, there are a lot of people working.

09:40.065 –> 09:43.166
[SPEAKER_02]: So we had to announce it, yeah, et cetera, et cetera.

09:43.186 –> 09:45.566
[SPEAKER_02]: So it’s a process that actually took us more than 12 months.

09:45.586 –> 09:48.147
[SPEAKER_02]: So in that period, we still had the photo studio.

09:48.427 –> 09:49.447
[SPEAKER_02]: Obviously, it’s an I fix.

09:49.467 –> 09:52.708
[SPEAKER_02]: Because so we continue to produce real world imaginary.

09:53.488 –> 09:56.209
[SPEAKER_02]: But in parallel, we already also produce the AI pictures.

09:56.789 –> 10:01.450
[SPEAKER_02]: So what happened is, we constantly did an AB test between the AI and the real world stuff.

10:03.010 –> 10:08.152
[SPEAKER_02]: For me, this was mind-blowing because six months ago, when we did another really big A.B.

10:08.172 –> 10:21.257
[SPEAKER_02]: test about it, actually the AI content has performed way better than the real world content and has achieved 10% higher conversion rate, which for eCommerce is a gigantic number.

10:21.397 –> 10:22.297
[SPEAKER_02]: I’ve never seen an A.B.

10:22.337 –> 10:26.979
[SPEAKER_02]: test for a feature where the winner had 10% uplift, and we’re talking about millions of

10:29.740 –> 10:41.229
[SPEAKER_02]: So at six, so 15 months ago we were at a point where AI was minus 90% cost same performance, six months ago we were at a point where AI still minus 90% cost by 10% conversion uplift.

10:41.969 –> 10:43.390
[SPEAKER_02]: And this was mind blowing.

10:43.470 –> 10:47.233
[SPEAKER_02]: I mean, when I saw the results first, I couldn’t believe it and said, I must be a mistake in the numbers, yeah.

10:48.034 –> 10:49.635
[SPEAKER_02]: And it’s obviously only getting better, yeah.

10:49.675 –> 10:52.077
[SPEAKER_02]: So AI is only getting better and only getting cheaper.

10:52.517 –> 10:54.639
[SPEAKER_02]: So today the uplift is even higher.

10:55.625 –> 11:12.992
[SPEAKER_00]: It’s sorry to interrupt me, is that uplift because your team and your data have got better at training the AI to create the perfect graphics for your website and your user base or is it because the AI is just naturally improved with endless releases do you think?

11:14.092 –> 11:27.663
[SPEAKER_02]: Yeah, I mean, it’s very interesting because if people tell me the AI is better and decent than the reeward, I usually don’t believe it because everything you can do in the reeward, you can do with AI, but everything you can do with AI except of some crazy science fiction stuff you can also do in the reeward.

11:27.683 –> 11:30.926
[SPEAKER_02]: So, where is this beat in performance coming from, it’s a very good question.

11:31.386 –> 11:34.229
[SPEAKER_02]: So it comes from two things, first of all, it’s the model selection.

11:34.729 –> 11:37.472
[SPEAKER_02]: Now the AI has really quickly learned which models perform.

11:38.092 –> 11:53.047
[SPEAKER_02]: You can obviously have that learning mechanism in the real world as well but the problem is you cannot book the model or you cannot find the model or you know you can’t sketch a model like in the real world you have the humans that this is small catalog usually if you’re operating for the studio

11:53.427 –> 11:57.990
[SPEAKER_02]: you have like a pool of hundred models and then there’s a high fluctuation in that pool.

11:58.010 –> 12:04.454
[SPEAKER_02]: So if you found a model that performs really well, most likely have a year later she lives in Milan, then have a year later she lives at Athens.

12:04.974 –> 12:06.695
[SPEAKER_02]: That’s the reality here.

12:07.075 –> 12:09.617
[SPEAKER_02]: So you cannot optimize models and obviously you cannot

12:10.257 –> 12:26.414
[SPEAKER_02]: sketch models here so the AI has built a perfect model like that model doesn’t look like AI like I can guarantee you no customer can see the difference yeah so it’s the model and the second big lever is the outfit combination so if you want to sell a t-shirt you always take a picture of the t-shirt

12:26.934 –> 12:27.535
[SPEAKER_02]: on an outfit.

12:27.575 –> 12:29.877
[SPEAKER_02]: So you always see a pair of travelers, blah, blah, blah.

12:30.197 –> 12:31.818
[SPEAKER_02]: So the question on how you style it.

12:32.079 –> 12:44.709
[SPEAKER_02]: Now this is also limited in the river because on the photos, do you only have a limited set of products there with which you can style and you always had a human stylist, yeah, that obviously did not have the algorithmic knowledge of the AI.

12:44.970 –> 12:49.293
[SPEAKER_02]: So what the AI quickly figured is which models perform, which stylings perform.

12:49.994 –> 12:52.396
[SPEAKER_02]: And then what is even more interesting, the AI has

12:53.557 –> 12:56.359
[SPEAKER_02]: This on a global level, but even has understood it on the local level.

12:56.379 –> 13:04.045
[SPEAKER_02]: So in Scandinavia, different type of model out of combinations work well, then in Eastern Europe, then in Germany, then in Spain.

13:04.625 –> 13:12.872
[SPEAKER_02]: And you can even go once, one lever below, you can even say within Germany, within UK, person A has a different taste than person B.

13:12.892 –> 13:16.554
[SPEAKER_02]: So person A may be dressed as formally and person B dressed as casually.

13:16.935 –> 13:20.737
[SPEAKER_02]: Now you can wear a blouse, the same blouse,

13:21.701 –> 13:25.783
[SPEAKER_02]: depending on the styling, can appear formal or can appear casual.

13:26.083 –> 13:30.966
[SPEAKER_02]: Yeah, so it’s the same product, but it can appear completely different if you style it in a different way.

13:31.426 –> 13:42.072
[SPEAKER_02]: Now, AI can take all of this in consideration and really build the perfect model outfit combination for global, local, personalized.

13:42.473 –> 13:44.994
[SPEAKER_02]: And this is driving the conversion right there up.

13:45.723 –> 13:47.644
[SPEAKER_00]: It makes so much sense when you outline it like that.

13:47.744 –> 13:53.086
[SPEAKER_00]: So it’s like the things we already knew before AI, but now we can actually deploy them.

13:53.486 –> 13:53.927
[SPEAKER_00]: I love it.

13:54.287 –> 13:55.667
[SPEAKER_02]: Khaled percent, we always knew it.

13:55.747 –> 13:59.409
[SPEAKER_02]: We were just not able to execute on it in the real world because we were bound by everything.

13:59.489 –> 14:05.031
[SPEAKER_02]: Limited by pool of models, limited by pool of products that you can start with, limited by humans.

14:05.231 –> 14:08.713
[SPEAKER_02]: Doing all of that, limited by gathering knowledge, limited by, you know,

14:09.433 –> 14:11.976
[SPEAKER_02]: giving that intelligence to the stylist and so on and so forth.

14:11.996 –> 14:14.259
[SPEAKER_02]: Yeah, so none of this is like a rocket science.

14:14.299 –> 14:18.704
[SPEAKER_02]: Yeah, I mean, that the model counts that the way how you style a product makes a huge difference.

14:18.724 –> 14:22.869
[SPEAKER_02]: I mean, none of this is like a surprise, yeah, but finally, you can execute on it.

14:23.490 –> 14:26.274
[SPEAKER_02]: And this is, for example, a product that we have then productized.

14:26.374 –> 14:28.136
[SPEAKER_02]: So scale is our shops software.

14:28.576 –> 14:35.581
[SPEAKER_02]: The Zars call it the Shopify for the big guys and then we have productized this AI we call it AI Content engines less for the studio.

14:36.061 –> 14:41.045
[SPEAKER_02]: It also has a software as a service product because it has worked so well for us That we said okay.

14:41.065 –> 14:43.667
[SPEAKER_02]: We have to productize it and offer to our scale customers

14:44.587 –> 15:00.006
[SPEAKER_02]: which we did in the first place, but then we got so many calls from other companies that were not scared customers but using other shop systems that heard about it and that said like guys can we also use it and we said okay actually why not yeah if you pay us we make a margin so it’s a win-win situation and then we have

15:00.506 –> 15:02.467
[SPEAKER_02]: open it and it’s called SKY Studios.

15:03.128 –> 15:05.269
[SPEAKER_02]: So SKY Studios.ai is the the main.

15:05.929 –> 15:11.632
[SPEAKER_02]: You can serve registered, you know, swipe your credit card and if you are a big company, you can also go for an enterprise plan.

15:11.772 –> 15:19.137
[SPEAKER_02]: So now we have more than 200 brands already using SKY Studios that have and we’ve started SKY and open SKY Studios for non-skay customers three months ago.

15:19.477 –> 15:26.561
[SPEAKER_02]: So within three months more than 200 brands were onboarded to completely switch or to switch contact production on a

15:27.203 –> 15:30.565
[SPEAKER_00]: Wow, I suspect you’ll be getting a few more as a result of this as well.

15:30.625 –> 15:32.787
[SPEAKER_00]: I would certainly be checking out fast in that space.

15:33.107 –> 15:37.530
[SPEAKER_00]: Let’s be on to the second big AI topic.

15:37.970 –> 15:40.972
[SPEAKER_00]: The second critical way AI is changing e-commerce in practice.

15:40.992 –> 15:43.693
[SPEAKER_00]: And this is one I’m all over a lot.

15:44.114 –> 15:46.976
[SPEAKER_00]: The whole agentic transaction piece.

15:47.216 –> 15:53.119
[SPEAKER_00]: And in a couple of years time, are we going to see agents doing all the buying on our website?

15:53.139 –> 15:55.401
[SPEAKER_00]: So will it still be humans doing it?

15:55.561 –> 16:00.283
[SPEAKER_00]: The guest boils down, too, is it going to be about the transaction or the traffic?

16:00.383 –> 16:05.466
[SPEAKER_00]: And you’ve seen some fascinating days with how Google are going at this, haven’t you?

16:06.366 –> 16:06.686
[SPEAKER_02]: Yeah.

16:06.786 –> 16:07.827
[SPEAKER_02]: I mean, firstly, it’s interesting.

16:07.847 –> 16:12.409
[SPEAKER_02]: Two years ago, everyone on conferences said, OK, in two years, everyone will buy through a shopping agent.

16:12.509 –> 16:13.370
[SPEAKER_02]: Now, here we are.

16:13.390 –> 16:17.612
[SPEAKER_02]: Still no shopping agent, but basically no shopping agent exists.

16:18.472 –> 16:24.295
[SPEAKER_02]: Shed GPT has announced this like so prominently, and has completely pulled back from kind of becoming a shopping agent.

16:24.735 –> 16:26.056
[SPEAKER_02]: And I have switched an advertising model.

16:26.076 –> 16:37.527
[SPEAKER_02]: I mean, this generally I see in the LLMs, they, I mean, Google is still working on a UCP universal commerce protocol to be to build a shopping agent, but they have all prioritized advertising hire.

16:37.948 –> 16:44.674
[SPEAKER_02]: So I think the LLMs all want to be able to answer the question on what is the best television or what is the best, whatever I’m looking for and you address.

16:44.694 –> 16:45.635
[SPEAKER_02]: Can you recommend me stuff?

16:46.536 –> 16:58.641
[SPEAKER_02]: But I think most of the LLMs now think about it as an advertising model to then basically send clicks to the online shops like Google and Meta do and basically charge a click price because it’s a more common way of monetizing traffic in a way.

16:59.301 –> 17:03.863
[SPEAKER_02]: I still believe we will have shopping agents eventually some time in the future.

17:04.463 –> 17:12.809
[SPEAKER_02]: But as at least Google and CHPT are prioritizing the advertising models and CHPT has completely stopped working on the shopping agent.

17:13.350 –> 17:16.592
[SPEAKER_02]: I do believe we won’t see a big shift in this in the next one or two years.

17:17.113 –> 17:19.674
[SPEAKER_02]: And also as a traffic channel today, they are not really relevant.

17:19.694 –> 17:22.617
[SPEAKER_02]: So people still go on Google and still serve on Instagram.

17:22.657 –> 17:27.440
[SPEAKER_02]: So Google and Meta are remained by far larger traffic channels for all online shops.

17:27.500 –> 17:28.201
[SPEAKER_02]: I know at least there.

17:28.835 –> 17:36.281
[SPEAKER_00]: So do you think that’s because the consumer’s aren’t ready yet, that the tech’s not ready yet, or that the business model isn’t ready yet?

17:37.082 –> 17:41.045
[SPEAKER_02]: I think it’s a combination of all, to be honest, like I mean, look at Google shopping.

17:41.085 –> 17:45.388
[SPEAKER_02]: I mean, Google shopping could introduce a checkout as well, but still they decided to send traffic.

17:45.408 –> 17:49.031
[SPEAKER_02]: So it is a better business model to send traffic than transactions.

17:49.051 –> 17:55.397
[SPEAKER_02]: Because if Google would do the transactions, the merchants would only be willing to give Google their margin as a commission.

17:55.977 –> 18:02.680
[SPEAKER_02]: Now, the thing about sending clicks is most online shops are paying more than their margin for the click because they bet on a customer lifetime value.

18:03.180 –> 18:08.963
[SPEAKER_02]: So today, Google charges more than the margin on a click basis.

18:09.223 –> 18:14.886
[SPEAKER_02]: The online shops able to generate new customers that they then monetize in the future.

18:15.586 –> 18:16.607
[SPEAKER_02]: And they are able to upsell.

18:17.127 –> 18:23.610
[SPEAKER_02]: Kind of if you look on monetization per user for Google, it’s better to send clicks than to proceed transactions.

18:24.370 –> 18:38.610
[SPEAKER_02]: And this is also what Chechibdi has figured, and that’s why they went for an advertising model, so it’s a better business model, but also I would say from the user side, at least as it now uses enjoy being in an online shop, it’s a little bit the same like, you know, offline has still a share of 75%

18:40.392 –> 18:58.544
[SPEAKER_02]: and not everyone wants to buy stuff online, people want to experience things, and even online people want to experience the proposition of an online shop, rather than going through a chat with each checkout, I mean nobody knows what the future will bring and I would also differentiate between use cases, so if it’s purely the mantra of use case, so I just want to

18:59.444 –> 19:00.664
[SPEAKER_02]: by these headphones, yeah.

19:01.024 –> 19:04.365
[SPEAKER_02]: So I know what I want to buy, it’s really, really like transactional.

19:05.105 –> 19:10.067
[SPEAKER_02]: This I think in the future will happen through some kind of shopping agent, because it’s a it’s a binary thing, yeah.

19:10.187 –> 19:16.288
[SPEAKER_02]: And know what I want, I just wanted the cheapest fastest way, so it’s very binary, there’s no fun in getting it, yeah.

19:16.708 –> 19:21.989
[SPEAKER_02]: So this I think we go through shopping agents in a near future, or at least in the mid to long term future.

19:22.529 –> 19:27.731
[SPEAKER_02]: You know, when it comes to the these lives that are into things where I’m really wanting to freshen up my wardrobe, or I want to

19:29.231 –> 19:31.194
[SPEAKER_02]: These are things where people actually enjoy.

19:31.534 –> 19:34.798
[SPEAKER_02]: For most people, it’s not a pain, but it’s actually an enjoyment.

19:35.019 –> 19:41.367
[SPEAKER_02]: And this I believe people like to go into an online shop where the environment is, you know, fun and inspirational.

19:42.068 –> 19:43.029
[SPEAKER_02]: And I mean, chat to you, P.D.

19:43.049 –> 19:43.610
[SPEAKER_02]: and Gemini.

19:44.311 –> 19:46.012
[SPEAKER_02]: their LM’s large language models.

19:46.032 –> 19:48.412
[SPEAKER_02]: They are basically translating knowledge.

19:49.053 –> 19:52.974
[SPEAKER_02]: It’s not the fun environment and the inspiration and the environment that you’re seeking for.

19:52.994 –> 19:57.735
[SPEAKER_02]: It’s more of a knowledge text, a factual environment.

19:57.795 –> 20:05.958
[SPEAKER_02]: So this is very good when it comes to who can give me these headphones here, the fastest cheapest, but it’s not very good when it comes to who can fresh not my wardrobe.

20:06.710 –> 20:09.632
[SPEAKER_00]: I cannot say how much I agree with you on all of that.

20:09.812 –> 20:12.615
[SPEAKER_00]: Oh, so rather me, just repeat it all back to you.

20:12.655 –> 20:23.523
[SPEAKER_00]: Let’s go on to number three, which is the internal agent consolidation, which is something I must admit I’ve been mulling over as well, but you are much more knowledgeable on it than I am so.

20:24.143 –> 20:32.109
[SPEAKER_00]: The thing is, at the moment, it feels like there’s a lot of deep-licated effort for the brand with the kind of behind-the-scenes agents or their customer service agents.

20:32.529 –> 20:43.637
[SPEAKER_00]: They’re building a data set and an AI agent on Shopify, then they’re building another one in the CRM, and they’re building another one on the personalization platform, being more customer service, and pulling all the same data together to speed their work flow up.

20:44.097 –> 20:44.438
[SPEAKER_00]: It’s like

20:45.278 –> 21:01.249
[SPEAKER_00]: huge duplication of work on the the back end of all of this given we all know that if you want an AI agent to work well you have to train it and train it and train it and get the right data it seems kind of crazy that we’re doing this multiple times does it not?

21:02.014 –> 21:03.575
[SPEAKER_02]: Yeah, I completely agree.

21:03.675 –> 21:10.498
[SPEAKER_02]: I mean, I think there was a trend like five, six, seven years ago in tech where it was all about composability and modularity.

21:11.158 –> 21:16.501
[SPEAKER_02]: And you had a module and you know, composable vendor, a composable tech infrastructure and a vendor for everything.

21:17.081 –> 21:27.426
[SPEAKER_02]: Because the back then, let’s say the common sense was, if you wanna be the best, you have to take the best in every area and then stitch it together somehow.

21:28.206 –> 21:31.668
[SPEAKER_02]: And I would say that Viper has changed massively in the last three, four years.

21:31.688 –> 21:39.852
[SPEAKER_02]: So now often, companies want to have more from one vendor and also kind of consolidate their tech landscape, I would say.

21:40.192 –> 21:45.235
[SPEAKER_02]: Still doesn’t mean that we are back into the monolithic world where you basically one tech system that’s everything.

21:45.695 –> 21:52.599
[SPEAKER_02]: Yeah, it’s also, it doesn’t go that extreme, but like, there’s definitely a consolidation happening in terms of how many in which tech systems you have.

21:53.039 –> 21:55.601
[SPEAKER_02]: And I would say agents doing work for you will

21:56.461 –> 22:11.888
[SPEAKER_02]: rather accelerate the consolidation because you also don’t want to have like you cannot have like 10 agents helping you to manage your online shop yeah that just doesn’t make sense or at least they have to be really well connected so what we are seeing is that

22:13.199 –> 22:13.459
[SPEAKER_02]: Yeah.

22:14.080 –> 22:29.830
[SPEAKER_02]: Agents and AI is taking over tasks and it’s becoming way more smart and we at scale for example we have implemented also a co-pilot now a backend but we are very lucky and fortunate with this game because we are covering a lot of the landscape while customers technologically we are kind of a full-fledged.

22:30.907 –> 22:40.051
[SPEAKER_02]: Suite that comes with an auto management system, product management system, check out front-hand shop management, promo engine, and so on and so forth.

22:40.652 –> 22:50.716
[SPEAKER_02]: So our core pilot has access to a lot and the most critical data and can therefore also derive informations and give the right recommendations and also execute under recommendations.

22:51.617 –> 22:56.459
[SPEAKER_02]: And yeah, we see that this performs way better than companies that are trying to, you know,

22:58.022 –> 23:10.049
[SPEAKER_02]: somehow get 10, a CRM agent, and a CMS agent, and a pricing agent, and a promo agent, and a shop management agent, and an Ole Miss agent, and I don’t know agent, and it’s because of my service agent.

23:10.269 –> 23:12.891
[SPEAKER_02]: I don’t know, I mean, I’ve never seen that work.

23:14.091 –> 23:22.696
[SPEAKER_02]: All ends in the chaos, and basically no more humans needing, rather than less humans, and also less good decisions.

23:22.716 –> 23:24.898
[SPEAKER_02]: This kind of very fragmented, cluttered world

23:25.778 –> 23:34.008
[SPEAKER_02]: dozens of vendors and dozens of agents trying to solve stuff with all of them having to sign up information, then silo execution capability side.

23:34.088 –> 23:35.249
[SPEAKER_02]: I don’t think that works very well.

23:36.270 –> 23:41.537
[SPEAKER_00]: Yeah, I think it’s going to be very interesting to see who wins on that front over the coming years.

23:41.577 –> 23:45.401
[SPEAKER_00]: We know which model kind of comes to the four.

23:49.100 –> 23:53.882
[SPEAKER_01]: E-commerce most at land is supported by some of the greatest companies in the E-commerce sector.

23:54.022 –> 23:55.463
[SPEAKER_01]: Here’s a reminder of who they are.

24:06.878 –> 24:13.781
[SPEAKER_00]: Okay, I love the top tip section because it gives me and I’ll list us some really quick ideas for taking our businesses to the next level of Terrac.

24:13.841 –> 24:15.721
[SPEAKER_00]: Are you ready for the top tips?

24:16.102 –> 24:16.402
[SPEAKER_02]: I am.

24:16.902 –> 24:22.224
[SPEAKER_00]: Okay, not that we haven’t given people loads of advice to get working on already, but here we go.

24:22.784 –> 24:30.427
[SPEAKER_00]: The book, top tip, if everyone listening to this podcast agreed to take Friday off and read a book to make their business better, which book would you recommend?

24:31.030 –> 24:34.353
[SPEAKER_02]: I’m the last person that should be asked for books and I’m not a book reader.

24:34.493 –> 24:39.037
[SPEAKER_02]: I read a lot, but I read short texts and yeah.

24:39.677 –> 24:43.020
[SPEAKER_02]: I listen to podcasts a lot, so I’m more of a podcast guy than a book reader.

24:43.440 –> 24:46.403
[SPEAKER_00]: So, favorite podcasts you’d recommend to listen to on a Friday.

24:47.326 –> 24:52.528
[SPEAKER_02]: Yeah, mostly it is to some kind of industry nerdy podcast some in German.

24:52.588 –> 24:55.509
[SPEAKER_02]: Yeah, so it’s it’s both each German podcast.

24:55.549 –> 25:11.835
[SPEAKER_02]: You’re honest like And then some general in I like to I like to also like to listen to interview podcasts like like treatment for example But they are really kind of don’t listen to one podcast regularly, but more like I guess I if I find guess interesting Then then I would listen to it.

25:11.855 –> 25:14.896
[SPEAKER_02]: Yeah, but it’s it’s mostly e-commerce related niche podcasts

25:15.486 –> 25:15.807
[SPEAKER_00]: Nice.

25:16.047 –> 25:16.408
[SPEAKER_00]: Okay.

25:17.209 –> 25:23.019
[SPEAKER_00]: The traffic top tip then which marketing method do you either price above all others or think doesn’t get the press it deserves.

25:23.779 –> 25:26.541
[SPEAKER_02]: I mean, for us, the most important marketing channels are Google AdMitter.

25:26.721 –> 25:36.668
[SPEAKER_02]: And here, I believe it’s really about granular optimization, kind of giving Google AdMitter the right information spec, our companies that still have no multi-touched Jewish system running.

25:36.708 –> 25:43.392
[SPEAKER_02]: And still, whatever, I do know customer lifetime value forecasts on the autosuccess page and give both information spec to the systems.

25:43.452 –> 25:46.715
[SPEAKER_02]: I think losing out a lot of efficiency in performance marketing.

25:47.335 –> 26:04.027
[SPEAKER_02]: So yeah, I would say for all online shops I know optimizing Google at meters, the largest level and there’s a lot a lot to do still, especially in meter obviously the whole AI development has helped to upload way more creatives and test way more creatives and the more creatives you test them the higher the likelihood is to hit.

26:05.047 –> 26:32.622
[SPEAKER_02]: Yeah, and the most underrated channel I would say it’s with an obvious case because I’m really in digital person, but like it’s We are sending physical postcards to an active customer’s because it doesn’t require to develop in more or less all countries in the world And yeah, it costs like a euro so and the reaction quarters are actually quite high depends on the country in some countries it’s too expensive, but in most countries that actually works really well as a Reactivation measure and really offline measure

26:33.583 –> 26:45.151
[SPEAKER_00]: I started by career in direct mail, so I always love a direct mail tip, but yeah, and you’re so right about the potential now with with meta with the AI and the data and all the rest of it is just mind blowing.

26:45.751 –> 26:56.799
[SPEAKER_00]: The tall top tip, maybe a collaboration tool, a social media plugin, a phone app or just a way of working, is there a cool little tool you use that makes you and your team more efficient from day to day?

26:56.819 –> 27:01.622
[SPEAKER_02]: To be honest, the highest efficiency heck in terms of tools

27:03.303 –> 27:06.708
[SPEAKER_02]: unfolded in last 18 months was deleting Instagram.

27:08.650 –> 27:11.493
[SPEAKER_02]: I deleted my Instagram account, 18 months ago.

27:11.633 –> 27:21.623
[SPEAKER_02]: I actually opened it again four weeks ago, but I had a screen time of 45 minutes with Instagram, deleted it so I saved me 45 minutes of doing scrolling per day.

27:22.303 –> 27:32.773
[SPEAKER_02]: Unfortunately, I installed it again, but now I deleted it again, I was like, well, it’s a bit of a funny, not funny, but I mean, to be honest, like, this is really, it’s had beaten everything to get 45 minutes.

27:34.597 –> 27:37.078
[SPEAKER_02]: that we’re uselessly used before.

27:37.318 –> 27:38.198
[SPEAKER_02]: Yeah, this was big.

27:38.498 –> 27:49.123
[SPEAKER_02]: I mean, other than that, like, seriously, like in the past, I would say, man, years, like, Claude and kind of Claude co-work and connecting Claude to Google Drive, and everything this for me has been a game changer to be honest there.

27:49.723 –> 27:57.146
[SPEAKER_00]: No, I like both those recommendations, the practice of one, and the take things off your phone, one, carbon-top tip to finish off them.

27:57.166 –> 28:00.687
[SPEAKER_00]: What’s your favorite way to reduce the carbon footprint of an e-commerce store?

28:01.653 –> 28:05.156
[SPEAKER_02]: Second hand people want to freshen up their wardrobes.

28:05.176 –> 28:07.518
[SPEAKER_02]: So they don’t want most people.

28:07.578 –> 28:15.925
[SPEAKER_02]: I’m not one of them I’m more the person that wears the same t-shirt every day, but because I have it in 20, 20 times But like most people actually want to wear

28:16.946 –> 28:34.417
[SPEAKER_02]: different things regularly and the only way to reduce carbon footprint and fashion that really has a lever when you can obviously produce more sustainable and this is always a good thing but like the best way is just to utilize the already produced products more and the biggest lever here second hand.

28:35.027 –> 28:46.904
[SPEAKER_00]: Yeah, if you think if it was a sea saw in fashion where the where the biggest bit of the carbon is used up, it’s not the beginning of the product’s life, even though there’s lots of room from there, it’s that end of the product’s life.

28:47.545 –> 28:50.489
[SPEAKER_00]: How do we get it get more wears per item and all the rest.

28:50.509 –> 28:50.970
[SPEAKER_00]: Oh, yeah.

28:51.270 –> 28:56.135
[SPEAKER_02]: I mean, if the, I mean, on average in Europe, a fashion product is 1, 7 times.

28:56.536 –> 28:57.817
[SPEAKER_00]: Is that it?

28:57.857 –> 29:04.844
[SPEAKER_02]: 25% of products produce a not one once, not even one once, on average 7 times.

29:05.345 –> 29:07.907
[SPEAKER_02]: So if you imagine how much resources and carbon,

29:08.688 –> 29:12.269
[SPEAKER_02]: utilization goes to producing the product and then it’s one seven times.

29:12.850 –> 29:19.992
[SPEAKER_02]: So if you just make sure it’s fine, people just want to wear it seven times, but then give it to another person that can also wear it seven times.

29:20.252 –> 29:24.174
[SPEAKER_02]: Boom, you have half the number of produced items.

29:24.194 –> 29:29.776
[SPEAKER_02]: This beats every measure in more sustainable production, which doesn’t mean more sustainable production is useless.

29:29.816 –> 29:37.799
[SPEAKER_02]: Like it is obviously makes sense, but like the higher utilization of already produced product beats every efficiency gain in production.

29:38.682 –> 29:47.908
[SPEAKER_00]: So how, as the co-founder of a mega fashion retailer, are you working towards helping consumers do that?

29:48.528 –> 29:49.989
[SPEAKER_00]: If it’s not too geeky a question.

29:50.535 –> 29:54.217
[SPEAKER_02]: Yeah, it’s a very good question and we have not been very successful in that so far.

29:54.778 –> 30:05.244
[SPEAKER_02]: We have that on C2B2C model, so trade and systems, so our customers can send in their products and then get money for it and then we take a picture and then resell it.

30:05.444 –> 30:13.169
[SPEAKER_02]: So it’s called C2B2C because a consumer sends it to a business, being also being whatever any other trade and partner and then we again sell it to a consumer.

30:13.729 –> 30:18.612
[SPEAKER_02]: But in second hand, the winning model has been C2C, like the wintered basically example,

30:19.773 –> 30:22.215
[SPEAKER_02]: The C2B2C unfortunately lives in a niche.

30:22.835 –> 30:38.246
[SPEAKER_02]: So as far as I’m convinced about second hand, unfortunately we have not found, we have not contributed in the most meaningful way to improve that other than motivating customers to basically not throw away their products, but to resell it.

30:39.026 –> 30:43.190
[SPEAKER_02]: And yeah, if they don’t want to use C2C platforms, they can do it with us.

30:43.290 –> 30:50.316
[SPEAKER_02]: But C2B2C is roughly 15% of the overall market, 85% of C2C, which is not a field where we are playing at the moment.

30:50.880 –> 31:13.404
[SPEAKER_00]: It’s an interesting one, isn’t it, to see how the industry will manage to monitor the unique product industry will manage to monetize it and encourage it when platforms like vintage and eBay, and depop, et cetera, are just going so mad with consumers just being really into doing the consumer to consumer peace.

31:13.902 –> 31:15.083
[SPEAKER_02]: But I think it’s a very good development.

31:15.103 –> 31:28.976
[SPEAKER_02]: Yeah, I think that it was less a point on about you, but on general, you know, I think embracing, recomers in general, I would say is from my point of view for most consumer companies,

31:29.930 –> 31:36.937
[SPEAKER_02]: The biggest contribution you can make to the planet in a way to reduce carbon footprint because it means less produced products.

31:37.237 –> 31:41.381
[SPEAKER_02]: That obviously comes at the expense of then at some point also shrinking market volume.

31:41.401 –> 31:43.263
[SPEAKER_02]: But I would say this is how it is.

31:43.283 –> 31:45.285
[SPEAKER_02]: Yeah, then you have to make sure you have a higher market share.

31:45.405 –> 31:51.131
[SPEAKER_02]: And I think by delivering a great experience to customers and by doing the right thing, you will also increase your market share.

31:51.371 –> 31:52.332
[SPEAKER_02]: And this is your way to grow.

31:52.923 –> 31:54.944
[SPEAKER_00]: completely, completely agree with you.

31:55.304 –> 32:05.106
[SPEAKER_00]: Tarek, before we say goodbye, could you please let the listeners know when they can find you, scale, scale studios, et cetera, if they want to get in touch and find out more?

32:05.666 –> 32:08.607
[SPEAKER_01]: Yeah, so scale.com, scale is written with an AY.

32:08.627 –> 32:14.929
[SPEAKER_02]: It’s kind of the hint to the module, background to scale, but wrongly written with AY in the middle.

32:15.529 –> 32:17.930
[SPEAKER_02]: And then it’s scale.com or scalestudios.au.

32:18.998 –> 32:19.778
[SPEAKER_00]: Simple as that, everyone.

32:19.858 –> 32:27.941
[SPEAKER_00]: And wherever you are watching this, listening to this, et cetera, somewhere near you, you will find links to both scale and scale studio.

32:28.141 –> 32:29.942
[SPEAKER_00]: So it’s very easy to get there.

32:30.062 –> 32:32.022
[SPEAKER_00]: No excuses, people, no excuses.

32:32.682 –> 32:33.243
[SPEAKER_00]: Tarek, thank you.

32:33.263 –> 32:33.763
[SPEAKER_00]: No, no, no, no.

32:34.283 –> 32:36.245
[SPEAKER_02]: It’s just one more remark on Sky Studio.

32:36.265 –> 32:40.370
[SPEAKER_02]: So if you’re registering on Sky Studio at the moment, you get 50 credits free.

32:41.352 –> 32:43.855
[SPEAKER_02]: So this allows you to generate 50 images.

32:43.975 –> 32:45.957
[SPEAKER_02]: AI images out of your product catalog.

32:46.578 –> 32:52.305
[SPEAKER_02]: That is a really cool offer I would say, because it means you can play around without having cost in the beginning.

32:52.795 –> 32:53.515
[SPEAKER_00]: very nice.

32:53.575 –> 32:56.757
[SPEAKER_00]: I’m glad you mentioned that because I’m sure that will get our world with the audience.

32:57.097 –> 32:58.758
[SPEAKER_00]: But look, thank you so much for coming on.

32:58.878 –> 33:04.441
[SPEAKER_00]: I could chat to you for about 20 more hours about all of this, but it’s been a real, so it’s been a real, real pleasure to have you on the show.

33:04.481 –> 33:05.082
[SPEAKER_00]: Thanks, Haric.

33:05.602 –> 33:06.462
[SPEAKER_02]: Thanks for having me.

33:06.542 –> 33:07.403
[SPEAKER_02]: We’ll speak fun to be here.

33:07.503 –> 33:07.783
[SPEAKER_02]: Thank you.

33:13.576 –> 33:16.258
[SPEAKER_00]: Well, I thoroughly enjoyed that chat with Tara.

33:16.278 –> 33:18.079
[SPEAKER_00]: Can I hope you all did as well?

33:18.580 –> 33:22.443
[SPEAKER_00]: Such a, I mean, clearly he really knows what he’s talking about.

33:22.523 –> 33:29.388
[SPEAKER_00]: He and his businesses are working through this stuff in large scale day to day, all these kind of AI areas.

33:29.668 –> 33:29.788
[SPEAKER_00]: And

33:31.501 –> 33:38.324
[SPEAKER_00]: I’ve seen it does feel very pragmatic and very straightforward what he’s recommending on the the AI content side.

33:38.824 –> 33:46.507
[SPEAKER_00]: Clearly, if we can give the AI the data it needs to be able to optimize and produce a better content, then we’re going to be seeing growth coming from it.

33:47.227 –> 34:05.812
[SPEAKER_00]: And I’m, am I the only one who’s kind of glad it took them a while to close down the Fogphoto Graphics Studio, so they got that long-range testing of AB comparison between the two, because that’s some phenomenal numbers seeing that how it’s now producing greater impact on sales than the human created, I guess, photography was.

34:06.772 –> 34:13.654
[SPEAKER_00]: And then going on to talk about transactional versus ad driven monetization for the AI platforms.

34:13.694 –> 34:15.595
[SPEAKER_00]: I mean, when you look at it the way he was

34:16.985 –> 34:30.614
[SPEAKER_00]: Why would they go after transactional agents any time soon, much, much, much more complicated to build and to manage in terms of integrations and payments and all the rest of it, or they can just slam at more ad fees on the rest of us.

34:32.275 –> 34:35.978
[SPEAKER_00]: Google and Meta have already got workable ad models.

34:36.039 –> 34:42.004
[SPEAKER_00]: So why would they be looking at transactional income rather than growing out the ad piece?

34:42.224 –> 34:43.926
[SPEAKER_00]: It’s the tried and tested business model after all.

34:44.546 –> 34:48.830
[SPEAKER_00]: So I guess we can park worrying about transacting agents anytime seen.

34:49.391 –> 34:50.291
[SPEAKER_00]: And then finally,

34:51.172 –> 34:59.995
[SPEAKER_00]: that data pot consolidation, which I think I’ve talked about before, you know, why are we building multiple agents in multiple places that are learning off the same data sets?

35:00.035 –> 35:05.737
[SPEAKER_00]: We need that consolidation, but what’s going to be fascinating to see is where that consolidation happens.

35:05.857 –> 35:08.337
[SPEAKER_00]: Will it be on the web platform as they’re seeing at scale?

35:08.397 –> 35:13.939
[SPEAKER_00]: Will it be in the personalization platforms of the CRM platforms or a standalone data warehousing system?

35:14.079 –> 35:14.779
[SPEAKER_00]: We shall see.

35:15.300 –> 35:15.940
[SPEAKER_00]: Time will tell.

35:16.580 –> 35:18.541
[SPEAKER_00]: Well, that certainly got you thinking about AI.

35:18.561 –> 35:29.187
[SPEAKER_00]: You can get your hands on our notes from this episode, including those top tips, links to what we mentioned, including both scale and scale studio by heading over to ecommercemasterplanned.com.

35:29.808 –> 35:38.673
[SPEAKER_00]: Use the short link ECMP.info forward slash 619 to take you to the correct page on the site immediately, if you want to.

35:39.313 –> 35:42.175
[SPEAKER_00]: And when you get to the website, get yourself on our email list.

35:42.215 –> 35:43.696
[SPEAKER_00]: So you know more about the stuff we’re sharing.

35:45.737 –> 35:54.305
[SPEAKER_00]: This episode of the previous two put out three massive episodes about practical usage of AI and e-commerce, which include lots of thoughts about where it’s all headed.

35:54.325 –> 36:01.732
[SPEAKER_00]: So I highly, highly recommend you listen to the two episodes before this one too, if this one has really got you thinking.

36:02.092 –> 36:08.056
[SPEAKER_00]: And if you’re feeling AI fatigued, then you’ll be glad to know we have something completely different for you next time.

36:08.356 –> 36:15.601
[SPEAKER_00]: We’re going to be talking to a highly successful luxury jewelry firm in the U.S. about how they’ve grown the business and what’s working for them at the moment.

36:15.641 –> 36:20.365
[SPEAKER_00]: So a nice step change for our AI focus content over the last three.

36:21.025 –> 36:23.186
[SPEAKER_00]: and thank you so much for tuning in to all these episodes.

36:23.746 –> 36:32.851
[SPEAKER_00]: We produce the new interviews every week because we want to inspire and help you to grow and thrive with your business, including encouraging you to make more sustainable buying decisions.

36:32.931 –> 36:44.236
[SPEAKER_00]: So if you know someone this show can help someone who’s struggling with the topics we’ve been discussing then please do ask them, tell them, suggest that they listen to the e-commerce master plan podcast.

36:44.557 –> 36:46.758
[SPEAKER_00]: I hope you have a great week and don’t forget to keep

36:50.598 –> 36:53.940
[SPEAKER_01]: Thank you for listening to the e-commerce Master Plan podcast.

36:54.361 –> 36:59.244
[SPEAKER_01]: Find out more at e-commercemasterplanned.com Slash Podcast.