Rewind: Let’s Talk AI Beyond the Headlines

WEBVTT
00:00:04.639 --> 00:00:10.880
Welcome to Commerce Media Matters, your go-to podcast for unpacking the world of commerce media.
00:00:10.880 --> 00:00:18.320
Here we catch up with industry leaders on news, research, insights, perspectives, and everything in between.
00:00:18.879 --> 00:00:21.120
Welcome back to Commerce Media Matters.
00:00:21.120 --> 00:00:26.000
We've had a stellar start to the year with some brilliant guests joining us across the first run of episodes.
00:00:26.000 --> 00:00:29.920
Roger Dunn, Vicky Sutton, Andreas Rythen, and Bill Fisher.
00:00:29.920 --> 00:00:34.079
All unpacking different angles of where commerce, media, and retail are heading.
00:00:34.320 --> 00:00:37.679
And of course, even if we tried, we couldn't ignore AI.
00:00:37.679 --> 00:00:41.520
It's come up in every single conversation, hasn't it, Nick?
00:00:41.520 --> 00:00:42.240
Absolutely.
00:00:42.240 --> 00:00:58.880
From Roger Dunn on how AI is shaping the shopping journey, helping people research faster, getting to decisions quicker, through to Ricky Sutton on the control and power and how publishers and brands can actually use their content and ad spend as leverage against these big tech companies.
00:00:59.119 --> 00:00:59.359
Yep.
00:00:59.359 --> 00:01:06.959
And Andreas on whether AI agents actually disrupt retail or simply become another channel retailers need to plug into.
00:01:06.959 --> 00:01:13.920
Plus, Bill on quality and trust and why they matter more now with the world flooded with AI generated content.
00:01:14.159 --> 00:01:17.680
So this episode, we're hitting rewind, we're pulling out the key themes.
00:01:17.680 --> 00:01:18.560
What's hype?
00:01:18.560 --> 00:01:19.359
What's real?
00:01:19.359 --> 00:01:21.840
And what's already starting to change?
00:01:21.840 --> 00:01:23.599
So take a listen.
00:01:23.599 --> 00:01:37.280
And in this episode, we're going to get into the topic on everyone's lips right now AI and commerce, from agentic assistance to conversational shopping, what's hype, what's real, and what's already changing customer journeys?
00:01:37.280 --> 00:01:39.439
So, Rog, welcome to the show.
00:01:39.439 --> 00:01:40.959
Thank you so much for joining us.
00:01:41.280 --> 00:01:41.920
Great to be here.
00:01:41.920 --> 00:01:42.959
Good to see you, mate.
00:01:42.959 --> 00:01:43.920
Good to see you both.
00:01:43.920 --> 00:01:49.120
Roger, you've recently launched the AI Commerce Brief newsletter, which is a cracker.
00:01:49.120 --> 00:01:51.120
Great to be dropping this at this time as well.
00:01:51.120 --> 00:01:51.840
Good read.
00:01:51.840 --> 00:01:53.359
What inspired you to start it?
00:01:53.359 --> 00:02:00.159
And what gap in the industry conversation are you hoping to fill by writing about AI's impact on shopping and commerce?
00:02:00.560 --> 00:02:01.120
Good question.
00:02:01.120 --> 00:02:03.760
And it definitely feels like the right time to me.
00:02:03.760 --> 00:02:07.040
So it feels like, as you know, like AI is everywhere.
00:02:07.040 --> 00:02:11.199
There's announcements every week, every day with new things coming down the track.
00:02:11.199 --> 00:02:12.000
Flat out.
00:02:12.000 --> 00:02:14.639
It's going to be a big impact on most industries.
00:02:14.639 --> 00:02:19.360
And obviously, my sweet spot is retail, it's e-commerce, it's retail media.
00:02:19.360 --> 00:02:26.960
So having a look at like how AI is going to be impacting that space is for me like super interesting, but also part of the job, right?
00:02:26.960 --> 00:02:29.840
We've got to understand what's happening and where it's all going.
00:02:30.319 --> 00:02:32.800
And you kicked it off with a pretty bold statement too.
00:02:32.800 --> 00:02:37.520
So you said AI isn't just a tool anymore, it's starting to shop for us.
00:02:37.520 --> 00:02:40.960
So how do you see autonomous agents changing the shopping journey?
00:02:40.960 --> 00:02:46.479
And are we heading towards consumers delegating purchasing and all decisions to AI entirely?
00:02:46.479 --> 00:02:46.719
Yeah.
00:02:46.719 --> 00:02:51.680
Or do you think it's still going to be human in the loop, or do you think the trust factor is going to be there for humans?
00:02:51.680 --> 00:02:54.240
I've got so many questions, and so is Paul.
00:02:54.479 --> 00:02:58.560
Well, I guess the first instance of AI was always the LLMs.
00:02:58.560 --> 00:03:00.400
It's that natural language interface.
00:03:00.400 --> 00:03:04.639
That was the first instance where we're we're asking questions, we're getting answers.
00:03:04.639 --> 00:03:14.479
But agentic is obviously where it's going to go, where the AI then acts autonomously and can carry out tasks and shopping being potentially one of those tasks.
00:03:14.479 --> 00:03:23.120
That's probably not here in full form today, but we're seeing you know agentique browsers, we're seeing you know instances of that coming through.
00:03:23.120 --> 00:03:27.919
So it is here and it's it's being tested, but it's not not the full story yet.
00:03:27.919 --> 00:03:34.159
But in terms of like where it's going, I mean, we have to temper it a little bit because it is going to be a percentage of a percentage.
00:03:34.159 --> 00:03:44.159
If you think about it, it's a digital activity, and we know e-commerce, depending on your category, is 10 to 20% perhaps of that category, and it's a percentage of that.
00:03:44.159 --> 00:03:48.879
But I think this is where you can get tied up a little bit in definitions.
00:03:48.879 --> 00:03:57.520
If your definition of agentic commerce is 100% automated from start to finish, then it's always going to be quite a small amount.
00:03:57.520 --> 00:04:07.840
But if you start thinking, well, I'm going to do a bit of product research at the top of the funnel and start having it go away and do things and research things, I might still go and buy it in store.
00:04:07.840 --> 00:04:10.000
I might buy it, you know, myself online.
00:04:10.000 --> 00:04:13.759
But it's shortening that route to purchase, it's shortening that time.
00:04:13.759 --> 00:04:14.319
Absolutely.
00:04:14.319 --> 00:04:17.120
It's getting you a better answer, perhaps getting you a better price.
00:04:17.120 --> 00:04:21.920
And it's all of those things which, you know, it's going to change how we shop.
00:04:22.160 --> 00:04:22.399
Yeah.
00:04:22.399 --> 00:04:26.480
I think uh the key word for me there you've used is temper, right?
00:04:26.480 --> 00:04:29.680
Because I'm old now for the kids out there, listeners.
00:04:29.680 --> 00:04:30.240
I am old.
00:04:30.240 --> 00:04:31.759
Why not look at what, Paul?
00:04:31.759 --> 00:04:37.920
You look at been around long enough to see this cycles come in and out, especially with e-commerce.
00:04:37.920 --> 00:04:40.720
And sometimes you can get wrapped up in the hype.
00:04:40.720 --> 00:04:41.439
It's very easy.
00:04:41.439 --> 00:04:43.920
We know that all the headlines are grabbing attention.
00:04:43.920 --> 00:04:48.959
When you step back and you go, okay, hold on a second, let's look at traditional e-commerce versus brick and mortar.
00:04:48.959 --> 00:04:50.160
What's the market share there?
00:04:50.160 --> 00:04:52.959
And still, brick and mortar is massive compared to online.
00:04:52.959 --> 00:04:59.439
What I'm hearing from you is you need to say, you need to take the same approach here that yes, this is extremely exciting.
00:04:59.439 --> 00:05:04.160
There is massive announcements happening regularly, and there's a big opportunity here.
00:05:04.160 --> 00:05:08.879
But let's take a breather here and actually understand what it means from a reality perspective.
00:05:08.879 --> 00:05:13.040
Because it's really shaking to the core a lot of retailers and brands.
00:05:13.040 --> 00:05:14.879
What you know, is this where we're what's happening?
00:05:14.879 --> 00:05:15.519
Where are we going?
00:05:15.519 --> 00:05:16.560
Do I have a business?
00:05:16.560 --> 00:05:17.040
You know?
00:05:17.279 --> 00:05:17.600
Yeah.
00:05:17.600 --> 00:05:18.720
Totally agree.
00:05:18.720 --> 00:05:21.199
And like that, that's why I say we should temper it.
00:05:21.199 --> 00:05:37.839
But then again, if you think about the amount of time and attention it requires to get ready for something like that, even if the numbers aren't there today, then that is probably where you should be spending at least a good portion of time to understand and see where the opportunities are.
00:05:37.839 --> 00:05:46.480
Because there's only so much you can do to improve things that have been now, you know, optimized and enhanced and focused on for so many years.
00:05:46.480 --> 00:05:50.560
Like this is where you can make a difference and and obviously create new opportunity.
00:05:50.720 --> 00:05:50.959
Yeah.
00:05:50.959 --> 00:05:53.600
I also like what you said about the definition, Raj.
00:05:53.600 --> 00:06:03.439
I liked how you said if you take it at the extreme literal that it's going to go and do the shopping and all the transaction and everything for you, then it's a percentage on a percentage on a percentage.
00:06:03.439 --> 00:06:17.600
But you know, it's going to, I hope, fuel e-commerce growth because you are going to be able to interact and ask it to shop around for deals and look for choice and help you make decisions and and other bits and pieces.
00:06:17.600 --> 00:06:23.519
And I think you will see the e-commerce journey, you know, benefit, but also the bricks and mortar journey.
00:06:23.519 --> 00:06:32.079
I mean, people still will want to go in store and you know, agents will help them get to decision where they may then go, okay, well, I still want my final look in store.
00:06:32.079 --> 00:06:33.279
I mean, do you agree with that?
00:06:33.519 --> 00:06:34.480
Yeah, absolutely.
00:06:34.480 --> 00:06:43.519
Like I think, and that's why I say, yeah, it depends on your definition because there are tons of use cases that you can use it for that still end up with an in-store purchase.
00:06:43.519 --> 00:06:45.759
It's just shortening that route to purchase.
00:06:45.759 --> 00:06:55.920
I think the interesting thing about this sort of macro trend that we're seeing is that it's being released to consumers at exactly the same time as it's being released to businesses.
00:06:55.920 --> 00:07:12.399
It's almost a consumer-led trend where we've been given this ChatGPT infant interface or other you know LLMs, and we're already tinkering with it, we're already using it, we're already wondering where we're going to kind of find products and use it for these kind of purposes as a consumer.
00:07:12.399 --> 00:07:14.480
And businesses haven't had that.
00:07:14.480 --> 00:07:19.839
It's not like they've had a long life cycle of setting up some Martex solution to get ready for this.
00:07:19.839 --> 00:07:23.199
It's it's like it's here and it's now, and it's like, what do we do today?
00:07:23.839 --> 00:07:30.959
With that in mind, you know, there's been a lot of talk about the extension of these LLMs and ChatGP moving into the browsers and the browser wars.
00:07:30.959 --> 00:07:44.480
You know, with ChatGTP expanding its capabilities and rolling out features like Atlas, do you think AI tools like ChatGTP could eventually replace traditional search engines like Google, or is Google just too much of a behemoth?
00:07:44.480 --> 00:07:48.160
And it appears that Google's doing a good job at kind of keeping up with the Joneses.
00:07:48.480 --> 00:08:08.160
Well, there's a very strong argument to make that Google is the leading AI company today, or you know, at the moment, even if they're they obviously have this sort of dilemma of protecting a business model that is very lucrative in search and the Templou links and moving people over to a much quicker, shorter experience to an answer that AI can provide.
00:08:08.160 --> 00:08:14.319
And how do you manage that without you know impacting your very kind of lucrative bottom line?
00:08:14.480 --> 00:08:14.639
Yeah.
00:08:14.800 --> 00:08:19.360
So I think that they've got all the tools in the bag, they've got their own you know, chips even.
00:08:19.360 --> 00:08:23.600
So like it's they're not in a bad place in terms of like opening up this opportunity.
00:08:23.600 --> 00:08:35.360
So yeah, it's obviously interesting to see them having some challenges in the the browser space where we've got Comet from Perplexity and Atlas browser with the recent one from Chat GPT.
00:08:35.360 --> 00:08:41.039
So interesting to see how that pushes innovation and introduces you know a new way of working.
00:08:41.039 --> 00:08:47.279
But I can't imagine it'll be too long before Chrome has a few extra features coming through the pipeline.
00:08:47.279 --> 00:08:48.879
But I'm sure wait and see.
00:08:49.120 --> 00:08:49.360
Yeah.
00:08:49.360 --> 00:08:50.559
I'm building on that.
00:08:50.559 --> 00:08:54.240
Do you see, therefore, traditional SEO strategies evolving?
00:08:54.240 --> 00:09:01.440
You know, we're moving towards what we call what search everywhere optimization, where AEO and GEO become new drivers for visibility.
00:09:01.440 --> 00:09:12.399
I mean, yeah, from your perspective where you're at at the moment, Raj, how's your business sort of working towards optimizing for LLMs and Atlas and Comet and those sorts of things?
00:09:12.720 --> 00:09:15.840
Yeah, so obviously it is a new new way of searching.
00:09:15.840 --> 00:09:24.320
And if you think about the type of search that happens, it's not just like very short keywords that are being punched in, it's like full sentences with a lot of context.
00:09:24.320 --> 00:09:26.159
That's the big difference with LLMs.
00:09:26.159 --> 00:09:31.039
It's context on what you might want to use that product for, or you know, some solution around it.
00:09:31.039 --> 00:09:37.759
You know, for us, maybe maybe how do I make this particular cocktail or what are the best cocktails to have with this particular product.
00:09:37.759 --> 00:09:40.320
So it's a lot more of a richer experience.
00:09:40.320 --> 00:09:42.399
And yeah, for us, I guess that's important.
00:09:42.399 --> 00:09:56.000
Like I think the first uh cab off the rank is always to track that and understand what people are using as tools for so that you can see where the opportunities are, see which sources are being cited and where the information is coming from.
00:09:56.000 --> 00:10:05.759
And then it's a case of yeah, seeing how you can, you know, turn up in those in those citations, in those those sources and become the recommended product.
00:10:05.759 --> 00:10:17.519
So yeah, there's lots of new disciplines, it feels like coming out and lots of few new acarims fighting around in terms of you know, answer engine optimization and general generative AI engine optimization.
00:10:17.519 --> 00:10:39.279
I think another one, even more specifically, is a gentic commerce optimization, ACO, which is another one which would be about particularly being surfaced for that instant checkout feature in things like ChatGPT, coming up in the product cards, coming up with, you know, eventually down the track, potentially the buy button next to it as well.
00:10:39.279 --> 00:10:56.000
And I guess that's the interesting thing for us is trying to see, you know, partly it's what we need to do, but there's probably some big partnerships we need to broach with the retailers because at the end of the day, that is more likely to be the place where someone buys a product and how do we surface ourselves in there?
00:10:56.720 --> 00:10:58.399
Thanks for joining us, Ricky.
00:10:58.399 --> 00:11:11.360
For the audience, you're basically the go-to guy for leaders when they want an unvarnished view of what AI and big tech are really going to do their businesses, not just their marketing deck.
00:11:11.360 --> 00:11:13.679
You're the guy, the straight talker, right?
00:11:14.000 --> 00:11:15.279
I'm happy with that description.
00:11:15.279 --> 00:11:22.320
Yeah, look, I I think that there's too much uh mollycoddling, kind of everybody trying to be a little bit too politically correct on this stuff.
00:11:22.320 --> 00:11:26.559
There's some really big issues being decided here for three-quarters of the world's population.
00:11:26.559 --> 00:11:27.600
We should probably talk about it.
00:11:27.759 --> 00:11:28.159
Absolutely.
00:11:28.159 --> 00:11:37.039
Just taking one step back to what publishers, what brands can do, uh, what even the retailers can do in this world of monopoly and AI, the rest of it.
00:11:37.039 --> 00:11:38.240
I mean, there was a lot of talk.
00:11:38.240 --> 00:11:44.399
I've just come back from NRF, and the the the NRF, the entire conference was about agentic commerce.
00:11:44.399 --> 00:11:50.960
Um, what's holding agentic back though is the ability to understand and trust the bots that are calling their site.
00:11:50.960 --> 00:11:59.600
And so there is an element of control for brands and publishers that I guess they can they can switch off the bot, they can block the bots crawling their content.
00:11:59.600 --> 00:12:01.120
Do you see a lot of that happening?
00:12:01.120 --> 00:12:05.679
Is that is that the place where you know premium publishers can still have an element of control?
00:12:05.679 --> 00:12:06.879
Do you think they'll do that?
00:12:06.879 --> 00:12:10.000
I mean, you referenced the news deal just before.
00:12:10.000 --> 00:12:14.320
Because to me, that looks like that's the last stage that they of defense for them.
00:12:14.320 --> 00:12:20.559
If they're smart around blocking these guys and and value their content enough, that's one protection.
00:12:20.559 --> 00:12:29.679
But how do they protect themselves against you know not jumping in too soon and making the right decisions and I guess going slow to go fast?
00:12:29.679 --> 00:12:30.960
They're all in a panic.
00:12:30.960 --> 00:12:34.159
They're all trying to go, hey boss, I know what out we need to do with AI.
00:12:34.159 --> 00:12:36.720
Let's jump on this new news now today.
00:12:36.720 --> 00:12:39.759
And they're they're probably not thinking about the long-term repercussions.
00:12:39.759 --> 00:12:42.399
Is that the place where they can have some control, Ricky?
00:12:42.720 --> 00:12:45.039
Well, we're we're talking about commerce here today, right?
00:12:45.039 --> 00:12:49.200
So no one makes a good purchase by rushing it, especially not a big purchase.
00:12:49.200 --> 00:12:54.879
Imagine you're making a purchase that's the entire future of your existence, like it's like buying a house.
00:12:54.879 --> 00:12:57.840
You don't have to rush to buy that house, okay?
00:12:57.840 --> 00:13:00.879
And so you need to be considered about making the right decisions.
00:13:00.879 --> 00:13:08.559
But look, I think that the industry as a whole, and when I say the industry, I mean the world basically has been captured by these monopolies.
00:13:08.559 --> 00:13:13.759
Um, and so if you're gonna start to unstitch it, there are things that can change, and you do have control.
00:13:13.759 --> 00:13:19.120
So, to your point, on publishers, which are only a tiny, tiny part of the global economy, right?
00:13:19.120 --> 00:13:27.279
But um, if you're a publisher, then absolutely you start to block the bots because the only way you're gonna get the deal you want is if you have leverage.
00:13:27.279 --> 00:13:33.919
And if you put the thing that you think is most valuable out in the world and let people take it for free, how much leverage do you have in a negotiation, right?
00:13:33.919 --> 00:13:34.960
Almost none.
00:13:34.960 --> 00:13:37.120
So you have to lock it away, right?
00:13:37.120 --> 00:14:00.000
Um, and the way I always tell this story to publishers when I sit down with them, and you know, you're dealing with some fairly serious global CEOs, you sit down with them, you go, hey, so you tell me your content is the most valuable thing that you own, but by putting it on the open web, it's like putting all your valuable money in a box and leaving it on the street, okay, outside a bank, and then being upset when you come back the next day and find people have taken the money.
00:14:00.000 --> 00:14:00.240
Yeah.
00:14:00.240 --> 00:14:02.240
You know, that's why we have banks.
00:14:02.240 --> 00:14:06.000
So what you should do is take the money and put it in the bank, and then no one can take it away.
00:14:06.000 --> 00:14:13.679
But here in the open web, what you're saying is you're going to put all your most valuable content on the open web and let people take it, but you don't lock it up.
00:14:13.679 --> 00:14:18.799
So these blocking deals that you can do is a bit like putting your content in a bank.
00:14:18.799 --> 00:14:24.559
And then once you put it in a bank, you can apply a price to that and you can have some leverage in this negotiation.
00:14:24.559 --> 00:14:28.879
So the first thing is there are thousands and thousands and thousands of publishers.
00:14:28.879 --> 00:14:36.159
Some I think a number quoted is about 55-56% of all global publishers are now blocking these bots.
00:14:36.159 --> 00:14:48.799
But the AI companies, Google and others, are refusing to change the bot that they have that searches for content for training and making it the same bot that searches for search.
00:14:48.799 --> 00:14:49.200
Right.
00:14:49.200 --> 00:14:53.279
So you if you want to block it for AI, then you can't be on search.
00:14:53.279 --> 00:14:56.720
And so there's a bit of a sound happening here.
00:14:56.720 --> 00:14:58.879
Um, but look, this will resolve itself.
00:14:59.200 --> 00:15:01.600
That doesn't sound like these guys, Ricky.
00:15:01.600 --> 00:15:04.720
That doesn't sound like a trick that Google would pull on anybody, does it?
00:15:04.960 --> 00:15:05.919
That is exactly the deal.
00:15:05.919 --> 00:15:10.720
Uh deal, it's in it's in their paperwork on their websites, in their terms and conditions, so it's not even secret.
00:15:10.720 --> 00:15:14.080
But um, you know, that will change in time, okay?
00:15:14.080 --> 00:15:18.879
So we're starting to see marketplaces emerge where content is being licensed.
00:15:18.879 --> 00:15:24.000
We've got some companies, Microsoft is spending real money to buy access to content.
00:15:24.000 --> 00:15:25.919
So that market is starting to emerge.
00:15:25.919 --> 00:15:37.440
But I do want to go back to one thing that I know is very close to your heart, you know, which is that the only reason that these tech companies are as powerful as they are is because of how much money they make.
00:15:37.440 --> 00:15:38.000
Okay.
00:15:38.000 --> 00:15:43.679
Um, and because they make that much money, then they have incredible power and they have incredible value.
00:15:43.679 --> 00:15:48.559
And it means that they can buy other companies and pretty much buy off governments and do whatever they want to do.
00:15:48.559 --> 00:15:51.120
You know, here's a new data center for you, Denmark.
00:15:51.120 --> 00:15:52.320
Like, please get off our back.
00:15:52.320 --> 00:15:54.879
These kinds of things are happening all over the world.
00:15:54.879 --> 00:15:55.120
Yeah.
00:15:55.120 --> 00:16:03.120
But the thing that again gets missed by lots of people is just how reliant most of these tech companies are on advertising.
00:16:03.120 --> 00:16:09.200
And so um, if you have a look at Google, 78% of Google's earnings comes from advertising.
00:16:09.200 --> 00:16:10.399
Okay, 78%.
00:16:10.399 --> 00:16:11.840
Astonishing amount.
00:16:11.840 --> 00:16:15.279
At Google, sorry, at Meta, it's more than 99%.
00:16:15.279 --> 00:16:16.240
Okay.
00:16:16.240 --> 00:16:24.559
Now, all of that money that's being spent with advertising with those companies is a specific decision by a CMO somewhere.
00:16:24.559 --> 00:16:30.080
And so when those CMOs are ringing me and saying, Ricky, how on earth do we have control and leverage over these companies?
00:16:30.080 --> 00:16:33.840
I'm like, you're giving them all the power by giving them all your advertising.
00:16:33.840 --> 00:16:42.799
And so it would be strongly advisable, I think, for the marketing industry to think about how it shares its money and where it puts it and to use that as leverage.
00:16:42.799 --> 00:16:59.519
Because at the end of the day, it doesn't really make a world of sense to kill every competitor in the market by denying them access to advertising and marketing funds uh just to pay the five tech companies, because then that just means that they become even more powerful and you have even less say.
00:16:59.919 --> 00:17:08.000
Yeah, surely that's a narrative that exists with certain CMOs and that surely that's have the conversation that's happening in the back room in the bars and cafes and cannes.
00:17:08.000 --> 00:17:12.640
How do we get together and kind of l and create some control again?
00:17:12.960 --> 00:17:14.880
Yeah, so some countries have done this.
00:17:14.880 --> 00:17:26.160
So in Korea, um, they have a law that all the advertising that goes to US tech companies has to be traded through a government-run ad agency, right?
00:17:26.160 --> 00:17:37.359
So they limit the amount of money that goes to Google and Meta in Korea and make sure that the money is shared across all of its domestic publishers and companies as well.
00:17:37.359 --> 00:17:45.599
So they have a thriving outdoor industry, they have a thriving media industry, they have a thriving radio industry because it's still being funded by this money.
00:17:45.599 --> 00:17:49.279
So look, I think uh you asked me what's the most popular misconception.
00:17:49.279 --> 00:17:55.759
I think the most popular misconception is that we're all powerless in this, but we're not, because we decide what we buy and where our money goes.
00:17:55.759 --> 00:17:57.759
And ultimately, this is all about money.
00:17:58.480 --> 00:18:06.480
It's interesting you say about the you know the control with the content, that piece around publishers, because it looks like the big tech aren't immune to this either.
00:18:06.480 --> 00:18:22.240
We've seen Amazon sue perplexity over access to you know Amazon's infrastructure because they're going to lose the the front door to the customer and that customer relationship if they come through perplexity and they're allowed to leverage off Amazon's thing.
00:18:22.240 --> 00:18:26.079
So the big tech are playing that same playbook when they're affected by AI.
00:18:26.079 --> 00:18:38.720
So I think your advice holds true is that you know hold hold that important asset really close to you and limit the access to it and make sure you can trade it on your terms because that's the real value here, isn't it?
00:18:39.039 --> 00:18:40.799
Prepare for a bloodbath, actually.
00:18:40.799 --> 00:18:42.559
Uh I think is the way to think about this.
00:18:42.559 --> 00:18:49.119
You've got, if if you put five sharks in a tank and you don't feed them, they're gonna start to fight pretty aggressively.
00:18:49.119 --> 00:18:55.200
Okay, and so you know, these tech companies, as we talked earlier on, there's no more money or customers left in America.
00:18:55.200 --> 00:19:00.160
So they have to grow outside America, which means the second largest market in the world is Europe.
00:19:00.160 --> 00:19:03.599
And Europe is busily building laws that these guys don't like.
00:19:03.599 --> 00:19:05.759
And if you then think, okay, well, where do I go from there?
00:19:05.759 --> 00:19:06.480
I'll go to China.
00:19:06.480 --> 00:19:08.160
Well, it's even more tricky in China.
00:19:08.160 --> 00:19:09.279
So I'll go to India.
00:19:09.279 --> 00:19:10.400
Oh, well, hang on a minute.
00:19:10.400 --> 00:19:11.680
India's part of BRICS.
00:19:11.680 --> 00:19:18.480
So that means basically Brazil, Russia, India, China, you know, South Africa, massive markets.
00:19:18.480 --> 00:19:21.680
They're all partnering together and saying, hey, we don't like America very much.
00:19:21.680 --> 00:19:27.920
So those five companies are going to get start to get really, really rough and tumble with each other.
00:19:27.920 --> 00:19:32.240
Right now, they're partnering up with each other and putting their arms around each other because it makes sense.
00:19:32.240 --> 00:19:42.720
That's why OpenAI is doing deal with this with this tech company, and this tech company is doing deal with this tech company, and these guys are all sharing because they're sharing the last blood they're extracting from the stone.
00:19:42.720 --> 00:19:46.160
But when that stone's dry, this is going to turn really nasty.
00:19:46.160 --> 00:19:49.200
Um, and they're gonna really, really, really fight with each other.
00:19:49.200 --> 00:19:51.519
And Europe is well ahead of the game here.
00:19:51.519 --> 00:19:54.000
It's already thinking about that future.
00:19:54.000 --> 00:20:07.119
So they passed a law in Europe that basically says we are just going to assume that all of these tech companies that we named are basically gatekeepers, which means they're basically monopolies.
00:20:07.119 --> 00:20:11.920
So we're not gonna need to prosecute them for being a monopoly, we're just gonna accept that's the case.
00:20:11.920 --> 00:20:17.599
Now, because of that, every company that we've decided is a gatekeeper, and it's all the ones you'd know, okay?
00:20:17.599 --> 00:20:21.200
Then we are gonna pass laws and we're just gonna impose those laws on them.
00:20:21.200 --> 00:20:29.359
So we're not gonna bother to have a two-year inquiry or a five-year investigation or a court case that's gonna take up tons of time and lawyers.
00:20:29.359 --> 00:20:30.640
We've already decided.
00:20:30.640 --> 00:20:33.440
And what we're gonna do is we're just gonna tell them how they play.
00:20:33.440 --> 00:20:38.640
So in the European bloc, 27 countries, they're gonna tell them what they can do and what they can't do.
00:20:38.640 --> 00:20:41.279
Now, until now, the tech companies have gone, so what?
00:20:41.279 --> 00:20:41.680
Who cares?
00:20:41.680 --> 00:20:43.759
We'll go and tell the White House and they'll fix it.
00:20:43.759 --> 00:20:47.039
But that's gonna go away when the money's not in America anymore.
00:20:47.039 --> 00:20:56.079
And when they need to get into more of that European money, then they're gonna start to fight tooth and nail with each other, and it's gonna be quite messy.
00:20:56.079 --> 00:21:03.440
So, a really clever somebody at the DOJ told me this a long time ago, the Department of Justice in the US that brought the antitrust trust.
00:21:03.440 --> 00:21:07.839
They basically said we can fight the fight, but the best thing we can do is to set them to fight each other.
00:21:07.839 --> 00:21:11.440
And I think the economics of the world are leading to that.
00:21:12.400 --> 00:21:18.319
We haven't mentioned the bingo word AI or its new cousin agency yet, or there I've said it.
00:21:18.319 --> 00:21:22.559
Look, agentic commerce is obviously the phrase on everyone's lips.
00:21:22.559 --> 00:21:29.680
I cannot wait to go to a few of the conferences coming up, and it's gonna be the bingo word it can for sure.
00:21:29.680 --> 00:21:35.920
But you know, AI isn't new to retail media, it's been powering bidding and targeting and optimization for years.
00:21:35.920 --> 00:21:38.480
What feels different now?
00:21:38.480 --> 00:21:44.960
What are you seeing with the the releases and UCP and what's happening with Shopify?
00:21:44.960 --> 00:21:45.759
What's going on?
00:21:45.759 --> 00:21:50.720
What are you seeing that's different with the agentic commerce space, which is really now kind of blowing up?
00:21:50.720 --> 00:21:52.079
What does it mean to retail media?
00:21:52.079 --> 00:21:53.200
What are you seeing?
00:21:53.599 --> 00:21:56.400
So I would always differentiate between two models.
00:21:56.400 --> 00:22:13.519
If you have the Amazon Rufus model, that is your Owned agent that is proprietary and basically helps users that hit your website to better find the products they are looking for or inspire them to buy something.
00:22:13.519 --> 00:22:17.200
That's one thing because that's what I would still call on site.
00:22:17.200 --> 00:22:21.680
So you can monetize that in exactly the same way.
00:22:21.680 --> 00:22:24.799
It's actually something I we just analyzed.
00:22:24.799 --> 00:22:29.759
Like what where do they place, where does Amazon place ads in Rufus?
00:22:29.759 --> 00:22:33.519
How they have sponsored questions, for example.
00:22:33.519 --> 00:22:41.359
And there are lots of ways how you can monetize this, in my opinion, equally well, if not eventually even better than the normal website.
00:22:41.359 --> 00:22:51.680
But then we have those third-party agents that is a um ChatGPT and Gemini and uh and Claude and so on.
00:22:51.680 --> 00:22:59.279
That's always like framed as a potential retail media killer.
00:22:59.279 --> 00:23:07.759
I think that's very short-sighted to say because it's it's a threat to the retail model itself.
00:23:07.759 --> 00:23:16.400
Because retailers like what is a retailer that no longer owns the access to customers?
00:23:16.400 --> 00:23:25.440
If people no longer hit websites but buy from agents, retailers become fulfillment companies in the end.
00:23:25.440 --> 00:23:25.839
Yeah.
00:23:25.839 --> 00:23:31.440
And that's clearly nothing they they they will ever want.
00:23:31.440 --> 00:23:40.559
So the the question is now I mean, clearly there is a new channel emerging, I would say.
00:23:40.559 --> 00:23:44.079
So many years ago, and you guys have seen that, right?
00:23:44.079 --> 00:23:46.960
Like, how often have we heard retail is dead?
00:23:46.960 --> 00:23:51.119
Because first it was social commerce that would kill um retail.
00:23:51.119 --> 00:24:00.000
And we had like waves, and every time this was the big revolution, and ultimately um they are still they are still around.
00:24:00.000 --> 00:24:17.119
So, what I expect is that sales coming through AI agents, similarly to marketplaces where you also don't own the transaction and the customer data, they will just be become part of the mix.
00:24:17.119 --> 00:24:25.279
And um they will grow slowly, maybe contribute to 5% of the volume at some point, maybe 10%.
00:24:25.279 --> 00:24:27.759
So they were part of that mix.
00:24:27.759 --> 00:24:35.279
I don't believe that um suddenly retail will want to go, uh will go away.
00:24:35.279 --> 00:24:43.359
Retailers don't want AI, third-party AI agents, so they will try to pull traffic um to their site.
00:24:43.359 --> 00:24:49.440
Users sometimes also like the usability, look, and feel of the websites they are visiting.
00:24:49.440 --> 00:24:54.079
Reason why they also don't all buy through marketplaces that are around.
00:24:54.079 --> 00:25:00.240
So I don't think that we will see an immediate disruption of the whole retail media game.
00:25:00.240 --> 00:25:03.200
Um, I think it's gonna be just an additional channel.
00:25:03.519 --> 00:25:07.759
Well, I mean, we're seeing that a little bit, aren't we, with the OpenAI ad experience?
00:25:07.759 --> 00:25:16.079
And we're seeing the big retailers buy space uh in there to bring them back into their systems and to power their RMNs.
00:25:16.079 --> 00:25:21.440
We've seen the announcement with Critio also working in that space uh of recent days.
00:25:21.440 --> 00:25:27.519
So, yeah, is it just another surface now to help the retail media networks and get them back to the retail?
00:25:27.519 --> 00:25:36.480
And is Agentic just a new way of saying that they're going to be able to connect up APIs or you know, like they they would have had to to hardwire it?
00:25:36.480 --> 00:25:38.240
I mean, that's what I think.
00:25:38.240 --> 00:25:51.279
Uh, it at NRF it was interesting to hear OpenAI say that you know, we can't let agents go out in the wild because you know, a retailer can't determine between a nefarious bot and uh and a good bot.
00:25:51.279 --> 00:25:59.759
So, you know, they were encouraging people to set up brand experiences and and use those ad services within within their interface.
00:25:59.759 --> 00:26:02.240
So I I I think I agree with you, Andreas.
00:26:02.240 --> 00:26:11.680
I think that this is another channel, at least for the short term, uh, that retail media networks, if they're really clever, can can harness and bring back into their world and their environments.
00:26:12.160 --> 00:26:16.319
A question on um on the functionality of those Chat GPT ads.
00:26:16.319 --> 00:26:23.359
Are they supposed to link traffic back to the websites and the transaction happens on the on the website?
00:26:23.359 --> 00:26:29.200
Or is this a shoppable ad that transacts on the um chat GPT side?
00:26:29.599 --> 00:26:36.640
Well, I mean, I um uh haven't seen much of the test being outside of the US, but I think the idea is that there's going to be both of those things.
00:26:36.640 --> 00:26:45.680
Certainly from uh uh vested interest, Voodoo is very interested in seeing the shoppable ad, um seeing the the actual conversion happen within the surface.
00:26:45.680 --> 00:26:50.960
And that's not just within the LLM or the or the chat surface, but also within ad surfaces.
00:26:50.960 --> 00:27:08.160
But as I understand it, and uh from the commentary I've been reading, what we're seeing is more the just the same sort of thing as sponsored search, whereby they're using that sort of those questions and add words type approaches to get more traffic back into the into their environment.
00:27:08.160 --> 00:27:21.599
So just filling up the top of their funnel that they're losing from traditional search models that have been disrupted because people are starting their, I guess, product discovery journeys out by going into an LN LLM and asking a question.
00:27:21.599 --> 00:27:22.960
That's what I've seen.
00:27:22.960 --> 00:27:24.079
I you might know more.
00:27:24.079 --> 00:27:26.880
You might be, yeah, it's a it's an interesting and evolving space.
00:27:26.880 --> 00:27:28.960
Obviously, it's only a test from OpenAI.
00:27:30.000 --> 00:27:39.119
Yeah, I believe as long as things would link back, that's not a threat at all because you you still get the traffic on the websites all all nice, you own the transaction.
00:27:39.119 --> 00:27:56.160
If people at a much larger scale than we will see it uh in the um in the next couple of months and maybe even years, if people transact to a great extent um from those AI agents and you don't see the transaction, I think that's the real threat.
00:27:56.160 --> 00:28:02.160
And this is the other question is how much um to what extent will people continue to visit websites?
00:28:02.480 --> 00:28:11.759
Yeah, look, you're asking questions of of uh of a company that has a kind of firm hold in this space for a little while.
00:28:11.759 --> 00:28:26.480
What we have seen, and we can't obviously go into detail, but from some very big activity with a number of retailers, that the strategy that OpenAI is employing, where they're going to be driving traffic right now.
00:28:26.480 --> 00:28:52.079
So they're creating obviously there's the intent, they're they're leveraging the intent, and then at the moment they're driving that traffic to said retailers, that will start to shift and change over time as they see a need and and a commercial, a simply a commercial need or commercial outcome to leverage, which is they can participate in a transaction fee if they can hold that shopping behavior there.
00:28:52.240 --> 00:29:00.720
Yeah, there's no reason why the retailer would be blind to the transaction with some of the technologies that are being proposed because it's still they'll be participating.
00:29:00.720 --> 00:29:03.759
Yeah, yeah, and they still need the pipes to flow.
00:29:03.759 --> 00:29:07.039
So they're still gonna see that Stripe is the gateway.
00:29:07.039 --> 00:29:16.480
That's gonna be a Stripe gateway into said retailers, Stripe, and so they're still, yes, they're probably sharing the data a little bit, but yeah, it'd be interesting to see.
00:29:16.480 --> 00:29:20.960
Anyway, that is a world that we're gonna see unfold rapidly.
00:29:21.920 --> 00:29:31.759
Bill has spent years digging into the data behind how people actually discover brands, consume media, and shop, separating the real signals from the industry hype.
00:29:31.759 --> 00:29:36.559
So today we're getting inside his brain to unpack what that numbers actually say.
00:29:36.559 --> 00:29:38.720
So, Bill, welcome to the podcast.
00:29:39.039 --> 00:29:39.920
Welcome aboard.
00:29:39.920 --> 00:29:40.960
Thanks for having me.
00:29:40.960 --> 00:29:42.160
It's an absolute pleasure.
00:29:42.400 --> 00:29:45.039
So, Bill, you know, a bit about your background.
00:29:45.039 --> 00:29:54.240
You produce a huge body of research spending consumer behavior, ad spend, as we said, retail, digital trends, you know, across Europe and globally.
00:29:54.240 --> 00:30:00.079
When you step back from it all, what's the big headline story you've been seeing this year?
00:30:00.079 --> 00:30:03.200
What's what's the most relevant thing you're seeing right now?
00:30:03.440 --> 00:30:07.920
Uh it would be pretty easy to say AI is a headline, wouldn't it?
00:30:07.920 --> 00:30:10.000
That would be a little bit a little bit boring.
00:30:10.000 --> 00:30:10.720
Um, yeah.
00:30:10.880 --> 00:30:12.240
It's a bingo word, yep.
00:30:12.480 --> 00:30:17.839
But you know, clearly it's reshaping everything, media, retail, marketing, everything that I I cover.
00:30:17.839 --> 00:30:26.720
But for me, I think the whole AI conversation leads to a a trend that will become increasingly important this year.
00:30:26.720 --> 00:30:33.519
And I'm going to use two words, and those words are trust and quality, because they they kind of go hand in hand, right?
00:30:33.519 --> 00:30:42.240
And and what I think we're seeing is this is the threat that AI will accelerate all these different ecosystems faster than trust can keep up.
00:30:42.240 --> 00:30:46.480
The example I'll use here is is in the social media space, one of the areas that I cover.
00:30:46.480 --> 00:31:02.000
And this is where consumers say they're seeing the most AI-generated content and this synthetic content that is flooding lots of social channels, to the extent that we're seeing synthetic influencers.
00:31:02.000 --> 00:31:05.119
They're not real people, they're entirely AI produced.
00:31:05.119 --> 00:31:07.440
And they're optimized for the algorithm.
00:31:07.440 --> 00:31:11.119
You know, they're they're there for a specific reason.
00:31:11.119 --> 00:31:16.079
AIs are pending other things as well, you know, recommendation engine search, lots of examples.
00:31:16.079 --> 00:31:26.400
But but what this is all doing is it's creating attention and it's leading consumers particularly, but marketers as well, to ask lots of questions around trust.
00:31:26.400 --> 00:31:31.599
So consumers are asking, to my example, are these influencers real people?
00:31:31.599 --> 00:31:33.039
Is is it authentic?
00:31:33.039 --> 00:31:36.079
Um, or is is this product review real?
00:31:36.079 --> 00:31:47.039
Marketers, meanwhile, are asking questions like, is this inventory uh uh brand safe or or are these impressions valid?
00:31:47.039 --> 00:31:57.200
I I'm always wary of using personal examples when talking about big big trends, but I think this one uh stands stands out quite quite well.
00:31:57.200 --> 00:32:04.319
At eMarketer, we are encouraged to use AI in our content production.
00:32:04.319 --> 00:32:05.920
Uh we use it quite heavily.
00:32:05.920 --> 00:32:18.799
I I use it relatively heavily, but nothing, absolutely nothing makes it out of the out of the door without being intensely uh interrogated, double checked, triple checked, stress tested.
00:32:18.799 --> 00:32:26.720
Because we we have to be a quality product, that's how we're having to position ourselves because folks can just go away and use AI now.
00:32:26.720 --> 00:32:32.319
So we have to be seen as a quality, trusted source.
00:32:32.319 --> 00:32:39.839
And and we go and use AI, but we we don't trust it yet to to publish without any input from us.
00:32:39.839 --> 00:32:56.079
So th this might be a little bit biased and protectionist in case my paymasters are listening, but quality, trusted products and environments, I think are gonna they're gonna gain real strong currency this year and next year and for years to come, I think.
00:32:56.400 --> 00:32:57.599
Absolutely, yeah.
00:32:57.599 --> 00:33:09.279
You know, in a in a world full of fake news where we saw, you know, social media heading, AI generated content has only proliferated that sort of approach.
00:33:09.279 --> 00:33:26.799
And so, you know, there is trust issues, um, and people are going to want to know exactly the content that these models are trained on, and they've seen enough hallucinations and enough errors in the results that they've got for their own experiences now to really want to interrogate those sorts of things.
00:33:26.799 --> 00:33:30.960
So we we we couldn't agree more that it's uh there'll be a return to trust.
00:33:30.960 --> 00:33:37.680
And, you know, we've seen some announcements from um OpenAI around agentic commerce.
00:33:37.680 --> 00:33:43.039
Um, and not that we agree with everything that they've said there, but I think that that's because of that trust.
00:33:43.039 --> 00:33:46.640
That I think the human in the loop component is critical.
00:33:46.640 --> 00:33:49.200
They are very powerful, very interesting tools.
00:33:49.200 --> 00:33:52.240
And like you said, they're great for helping you produce content.
00:33:52.240 --> 00:33:59.119
But that human in the loop, that creativity, that reasoning that you've got to actually look is does that look right to me?
00:33:59.119 --> 00:34:00.960
It's a it's an important factor.
00:34:00.960 --> 00:34:02.480
So I we couldn't agree with you more.
00:34:02.720 --> 00:34:05.359
What I think is really interesting is the trust and quality component.
00:34:05.359 --> 00:34:12.559
For me, trust and quality really has always been some of the pillars of a of a solid brand, right?
00:34:12.559 --> 00:34:36.800
But so trust and quality now is really really a brand really needs to lean on that to give that differentiation and to allow them to kind of be, you know, relevant in a world of just so much and so many knockoffs and copies and dilution and you know that trust and that quality, like you just said, we might eMarket has got to stand for that because yeah, like hey, we can go and pull that from AI now and go and pull that.
00:34:36.800 --> 00:34:38.000
Why would I come to e-market?
00:34:38.000 --> 00:34:44.719
Well, e-market as brand is now more important than anything because that brand really is the IP.
00:34:44.719 --> 00:34:54.960
So the trust and the and the quality that that needs to be instilled at everything you do, I think is it is a lesson for all businesses operating in a world of AI now.
00:34:55.280 --> 00:35:10.800
Yeah, and and and from a marketer standpoint as well, they're increasingly asking the questions, as I mentioned, do I want my brand in this environment that is going to be increasingly questioned if it's if it's a low-quality environment?
00:35:10.800 --> 00:35:13.840
Brand safety is an issue that comes up quite a lot in this as well.
00:35:13.840 --> 00:35:27.360
And brand safety is a double-edged sword when it comes to AI, because AI can help brand safety, it can help spot AI generated content and things like that, but also it increases the amount of risk as well.
00:35:27.360 --> 00:35:34.719
So AI is it's a risk manager, but it's also a risk multiplier, I would say.
00:35:35.599 --> 00:35:35.920
Yeah.
00:35:35.920 --> 00:35:55.039
Do you do you think brands are also um on that, you know, worried about how they can actually show up within AI ecosystems as a brand, and that they're worried about, you know, losing the premium of what they've created and that relationship that they have as a brand with the consumer.
00:35:55.039 --> 00:36:06.159
I mean, it's it's a very different proposition than what brand marketers are used to, being able to communicate what they stand for, you know, in a language, large language model.
00:36:06.159 --> 00:36:07.119
It's just text.
00:36:07.119 --> 00:36:08.239
That's what we're hearing.
00:36:08.239 --> 00:36:13.360
Are you hearing the same thing that they're that markets are worried about how they show up as a brand within these large language models?
00:36:13.599 --> 00:36:14.079
Yeah, yeah.
00:36:14.079 --> 00:36:24.400
And we're and we're we're I mean, we're still very early stages, right, of of advertising appearing in in these environments, but but that is something that they're absolutely concerned about.
00:36:24.400 --> 00:36:26.880
But they want to be there because they have to be there.
00:36:26.880 --> 00:36:41.599
And as we talk about uh, you know, as we get into the conversation today about how consumer behaviours are changing, if that's where consumers are, the the dollars follow the eyeballs, you know, it's a it's a hackneyed uh phrase that we always use, but you've got to be there.
00:36:41.599 --> 00:36:51.199
It's how how you can guard against some of these issues that these AI generated responses throw up.
00:36:51.199 --> 00:36:53.440
So again, it's this double-edged sword thing.
00:36:53.440 --> 00:36:58.960
Um, and I'm not gonna give you any answers today because I don't have any because we are so early in the process.
00:36:59.199 --> 00:37:01.280
Oh, I think I think you're spot on is so early.
00:37:01.519 --> 00:37:02.559
Yeah, absolutely.
00:37:02.559 --> 00:37:07.360
This is something that that marketers really want to get a handle on as early as they can.
00:37:07.360 --> 00:37:20.800
And and this is why hopefully a quality, trusted environment like eMarketer can help clients begin to understand this rather than just listening to the AI that's sticking up for itself, right?
00:37:20.800 --> 00:37:22.639
Oh, absolutely, absolutely.
00:37:22.719 --> 00:37:26.880
Like I I I Paul and I uh have are lovers of AI.
00:37:26.880 --> 00:37:30.719
I was talking about it at length with uh one of a uh a team.
00:37:30.719 --> 00:37:37.519
I feel like I've spent a lifetime of briefing, so I feel like I'm now briefing an engine to be able to produce the content that I need.
00:37:37.519 --> 00:37:43.920
But at the same time, you've got to be really careful about that information that's coming to you through that and the brand safety comes through, right?
00:37:43.920 --> 00:37:55.760
Which I we've talked about another part podcast with another guest about triple threat redundancy where we're taking the content out of that AI, we're running it through another one, another one to give us some kind of level of clarity and kind of cut away the hallucinations.
00:37:55.760 --> 00:37:59.280
But you talk about this is something that's very apparent now.
00:37:59.280 --> 00:38:04.880
Um we in our industry, and I think I'm what the kids call an old man now.
00:38:04.880 --> 00:38:22.000
Um I think anybody calls it an old man, but I'm thinking back in my career, and we always become in this industry in advertising and marketing, we become hyper-fixated, not just on what's on and happening now, but on what's coming and new, and we've got to be on top of it and we've got to be aware of it, and we've got the next big thing.
00:38:22.000 --> 00:38:26.880
You know, two questions, I guess, and one of them dovetail out of what we just talked about then.
00:38:26.880 --> 00:38:36.960
Firstly, is there something that's just way, and it could be something really kind of media, you know, kind of you know, minor that we're really overhyping at the moment.
00:38:36.960 --> 00:38:42.559
And conversely, is there something that deserves a little bit more focus?
00:38:42.559 --> 00:38:45.440
Like, is brand safety something we should really get into?
00:38:45.679 --> 00:38:46.400
Uh yeah.
00:38:46.400 --> 00:38:51.280
Let's stick with AI for a minute because I don't want to sound like a broken record.
00:38:51.280 --> 00:38:53.440
But there are three camps when it comes to AI.
00:38:53.440 --> 00:39:01.760
Um, there are those that think AI is going to fix everything, there are those that think it's the devil incarnate, and then there's this big middle where most of the sensible people sit.
00:39:01.760 --> 00:39:07.280
But I I think that idea that AI will do everything is the bit that's overhyped.
00:39:07.280 --> 00:39:09.840
So let's look at retail for a moment.
00:39:09.840 --> 00:39:20.000
So we know that that AI is having a significant impact on consumer shopping journeys, but checkout on AI platforms is really, really small.
00:39:20.000 --> 00:39:34.480
So we we produced a forecast in December last year that suggested that checkout on AI platforms would account for less than 2% of total retail e-commerce sales in so this is a US number by 2029 as well.
00:39:34.480 --> 00:39:44.400
So even a few years of this being around, so you know there's this narrative that AI improves everything, creative, targeting, personalisation, commerce outcomes.
00:39:44.400 --> 00:40:04.639
But in reality, to your earlier point, but we like AI, it does lots of good stuff, it certainly increases efficiency, lowers production costs, it can improve speed of optimization, but it doesn't inherently improve uh brand equity or or quality, and as we've already mentioned, doesn't breed consumer trust necessarily.
00:40:04.639 --> 00:40:09.760
So AI is going to continue to be transformative, but it's far from being the absolute answer.
00:40:09.760 --> 00:40:15.599
So it deserves a hype, but I I wouldn't go so far as to say it's the answer to everything.
00:40:15.599 --> 00:40:17.840
That's the bit that's probably overhyped.
00:40:18.559 --> 00:40:19.519
Yeah, I would agree.
00:40:19.519 --> 00:40:23.519
You still would you say that ChatGPT was launched in November 2022?
00:40:23.519 --> 00:40:25.519
We're only a few years into this.
00:40:25.519 --> 00:40:28.239
It's akin to being 1996, right?
00:40:28.239 --> 00:40:30.400
Google launched really in 98.
00:40:30.400 --> 00:40:33.760
It really didn't start to hit the straps 2000.
00:40:33.760 --> 00:40:36.800
And look where we are today and what's come from all that.
00:40:36.800 --> 00:40:47.119
You were really just so early, and we've got to be really careful that we don't get just swept away up in this just this hype that can kind of become misguided.
00:40:48.719 --> 00:40:51.920
Thank you for listening to the Commerce Media Matters podcast.
00:40:51.920 --> 00:40:55.599
If you enjoyed this episode, make sure to follow and subscribe.
00:40:55.599 --> 00:41:04.239
We're always on the pulse of the ever-changing landscape of commerce media, ready to share the newest trends, insightful perspectives, and expert opinions with you.
00:41:04.239 --> 00:41:09.599
Until then, stay curious, stay informed, and keep making media that matters.
00:41:09.599 --> 00:41:17.360
This podcast is executively produced and edited by Perea Tahazada with co-producers Sandy Ho and me, Rachel Fang.
00:41:17.360 --> 00:41:25.760
This podcast is recorded on the lands of the Bidigal, Gadigal, and Warrenjo people, and we wish to acknowledge them as the traditional owners of the land.