The Shopper is REALLY in Control Now, with Caroline Giegerich, VP, AI and Marketing and Collin Colburn, VP, Commerce and Retail Media at IAB

Transcript

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

Lauren Livak Gilbert (00:00):

Welcome to Unpacking the Digital Shelf, where industry leaders share insights, strategies, and stories to help brands win in the ever-changing world of commerce.

 

Peter Crosby (00:22):

Hey everyone, Peter Crosby here from the Digital Shelf Institute. There is lots of consumer polling out there about why and how shoppers are using AI in their buying journeys, but you can't trust what people say. You have to know what they do. Bring on the IAB. They, along the talk shop, put together research that took direct signals from hundreds of shopping journeys to understand what is really happening out there between shoppers and their agents and their retailers, and what is not. Caroline Giegerich, VP AI and Marketing, and Collin Colburn, VP Commerce and Retail Media from IAB came to armed with the facts to help you shape your AI and agentic strategies. Caroline and Collin, welcome to the podcast. We are so delighted to have you on.

 

Caroline Giegerich (01:11):

Thanks for having

 

Collin Colburn (01:12):

Us. Yeah, thanks for having us. Excited to be here.

 

Peter Crosby (01:15):

So I have to tell you, this research has been such an inspiration to me. I'm in the midst of writing a fairly significant white paper on Agentic Commerce and the Agentic Shelf. And your research here with TalkShop just inspired so much of the way I was thinking about how to describe how the world is changing with the consumer and that they are once again more in charge than ever because of what ... So this data is really a thank you for putting it together. And the beauty of it is that it's actual shopping sessions that you follow to find out what's really happening out there rather than just a poll. It's

 

Collin Colburn (01:59):

Like

 

Peter Crosby (02:00):

Real data and it's ... I can't ... Yeah. Anyway, so I'm very excited about this. So Caroline, why don't you start out just telling us, well, our listeners particularly about this report.

 

Caroline Giegerich (02:12):

Well, I love that you mentioned we're actually watching real shopping sessions because to me, that also makes a major difference because what people say in surveys versus what they actually do in practice are quite different. So that is a very unique part of the study. We found that ... And just to quantify the timing on this, we ran the study in October of 2025, and that honestly feels like a lifetime ago at this point in March of 2026. So we're only accelerating from the data that we're about to talk about right now. Nearly 40% of US shoppers are now using AI when shopping, which I think is a huge figure. And AI was expected to influence more than 260 billion in global e-commerce over the holiday season. I think what's interesting is how much that's accelerating. So I just looked up what is the data saying now, and now it's more than half.

 

(03:09):

But I think we're going to a place where certainly consumers are using AI for awareness and consideration the first two phases of the shopping phase, I would say. We found that in this, and I know we're going to talk about this later, that while certainly shopping is being affected in the first two phases, not necessarily in the transaction phase of things, which was an interesting learning and sort of corroborated by recent signals that we've been seeing in market with OpenAI taking instant checkout out of flow. But I'll pause there and let Collin jump in with his thoughts as well.

 

Collin Colburn (03:53):

Yeah. So Carol, adding on what Caroline said there, I think one of the really interesting things to me is that I sort of went into this research thinking obviously people are going to use AI most in the consideration phase. Help me make clarity from the chaos that is online shopping today, where there's just so much choice, there's so many products out there, there are so many different places where I can buy those products. And that certainly came to life. But what I thought was even more interesting than that was that 64% of the sessions that we viewed resulted in shoppers finding a new product via AI. So it's not just for consideration, it is also helping in the early stage discovery portion where people are identifying, "Hey, I have a need or I'm looking for this sort of thing." And AI is actually helping them not only come down through the funnel, through consideration, it's also helping very, very early stage in the funnel as well.

 

(04:58):

So I thought that was really a little surprising for me, which that's one of the things I love about research is that it always kind of bucks your hypothesis in a way.

 

Peter Crosby (05:09):

That's what I think is so exciting about mission-based shopping because it's actually intended to help you discover that idea that you're going in with a problem to solve or an event or something like that, and you're kind of wide open. The intent is the mission, not a particular thing. And that's what's, I think, such an opportunity for brands that are really going to invest in conquering this new space over the next few years.

 

Caroline Giegerich (05:41):

Just a note on that, because I actually realized I'm literally drinking something right now that was literally an AI shopping experience. And I know this is audio only, but what I'm drinking here is- It's skin. I am drinking in the office midday.

 

Lauren Livak Gilbert (05:57):

5:00 somewhere.

 

Caroline Giegerich (05:59):

That's not rare. We're in advertising after all.

 

(06:03):

Yes. It's Premier Protein. And this started as me going to GPT. No, I feel like I need more protein in my day to stop my late night snacking. And it ended with me buying this product, which I'm drinking right now during this podcast. So I think it speaks to what Carl said. It works. Affiliate fees, Premier Protein, please. Exactly. It speaks to what Collin just said though. It's like, yes, you could go looking for something, but I was just trying to figure out how to curb my late night stacking and then all of a sudden, of course, I'm buying a product.

 

Lauren Livak Gilbert (06:43):

I love it. That's such a great example.

 

Collin Colburn (06:45):

It's fascinating. I mean, it really was something that I wasn't really expecting, but where my hypothesis was proven out. So we did the shopping sessions where we watched real, I think 450 different sessions, but then we also paired it with a typical quantitative study. So we surveyed consumers about their experiences. And that actually proved out my hypothesis, which was that shoppers said, 83% of them said that the researching and comparing of products stage in their journey was the most valuable. It was the most effective, I think, was the phrase that was used. So it does both really well when it comes to discovery, but I think what shoppers use it most for, they find it most helpful for today is in that research phase of their journey. Help me make a better choice.

 

Lauren Livak Gilbert (07:46):

Were there any ways that people used AI shopping that was surprising or even any categories that were like, oh, I didn't expect them to use it for this and they did? Anything that kind of popped out?

 

Caroline Giegerich (07:59):

I mean, I don't do this, so I found it interesting, but it makes perfect sense when I say it out loud. I was like, oh no, that's obvious. I should use it for that. It was the price tracking or deal alerts. I looked to it to help me narrow down product choices, but I never realized why not just use it to track where to get the best possible pricing. I don't know why that simple fact alluded me until I saw the research, but I thought that was interesting to me.

 

Collin Colburn (08:33):

Yeah. And I think I would just add that what was kind of surprising in a way for me was I was actually a little surprised at how many shoppers say that they go to a marketplace or a retail merchant site to validate a lot of the information. Yes, it intuitively makes sense that you aren't going to just use a single source, I think, right now to validate everything, but 78%, I thought that number was just really high that said that they went to that retailer marketplace site to validate all of the information, all the details that AI was giving them. That to me, that number, the size of that number was surprising.

 

Lauren Livak Gilbert (09:18):

But I think that speaks to the trust piece, right? That's why I think the instant checkout, that's why OpenAI pulled back from it. That's why people aren't purchasing there because yes, it's very helpful and it gives you the correct answer, but how many times I've clicked on a link on ChatGPT and it goes to a 404 error and it's like a page not found, especially if you're spending money on something. I went to go buy a couch the other day. I was like, "Oh, this is the couch that I've been researching." And I clicked it and it was a 404 error. And I was like, "Oh, back

 

Caroline Giegerich (09:49):

To the research

 

Lauren Livak Gilbert (09:49):

Game."

 

Caroline Giegerich (09:50):

This data showed that as the number one friction, which was errors and citation or errors and details. And you can see how the AI companies are moving towards better data sources and these deals that they're doing, Gap with Google, Gemini, OpenAI with the Walmart, Etsy, et cetera, et cetera, et cetera. That's so that they can lower the friction that you're talking about with better data so that you're not like, "What?" Yeah.

 

Peter Crosby (10:20):

Speaking of trust, that is just essential to this journey. And one of the things that really stood out to me, and I totally stole/referenced in my white paper is what in the white paper, at least I'm calling the trust gap. And you had data on the extra steps that shoppers would use in order to verify the details. And I'm wondering, Collin, if you can share that data with us.

 

Collin Colburn (10:50):

Yeah. So on the trust piece, that's obviously where a lot of this friction happens. 46% of shoppers saying they don't fully trust or only 46% saying that they fully trust AI's recommendations and the information that's being given to them. So that means a majority don't yet fully trust AI for shopping. One of the things that we didn't cover in our research, but that this makes me think of, and Caroline brought it up before, we did this research in the fall, so part of me's kind of already like, oh, as soon as it hits the presses, it's already old.

 

(11:32):

But one of the things that I think is in the zeitgeist at the moment, and at least in the retail commerce space when it comes to Agentech commerce, is not only is there consumer trust that needs to be bridged, there's also trust between the agent and the retailer that I don't think has been talked about enough, but is starting to be talked about. So the question is, what is the agent allowed to do? So in a normal human-driven shopping world, the retailer knows someone has come onto target.com, maybe they're logged in, they're a circle member, Target knows who you are. In this agentic driven world, they don't know who you are. They might be able to know that you're an agent, but most importantly, they don't know what you are authorized, what the human, what the consumer has authorized that agent to do. So has the user given them authorization to browse and search on their behalf?

 

(12:33):

Have they given them the ability to add an item to a cart? Have they given them the ability to purchase? All this stuff matters, and it's a trust issue between agent and retailer because you have all these downstream retail use cases that are really dicey. If an agent buys a product that is a scam, which I think probably ... I know I have in my digital career of buying something-

 

Lauren Livak Gilbert (13:06):

Instagram ads.

 

Collin Colburn (13:08):

A shady website. So who's at fault for that? Is it the agent? Is it the human? And then the same thing for when it comes to reading return policies. I know for me, there are certain things, and definitely my wife with apparel that she or I buy or won't buy because we know that you have to get it back pretty quickly for you to get the full ... So there's all these little ... It's the nuances of retail that I think that is so important when it comes to trust on both sides of the equation. So consumers don't fully trust what AI gives them, but then also retailers, I know, have these concerns with trust too.

 

Peter Crosby (13:49):

Well, also the LLMs right now do not really give users great ways to express their preferences clearly like, "I don't want you to do this on my behalf. I do want you to do this. " At least in certainly in OpenAI that I tend to use a lot for this, there's just no way to set preferences in a clear way. And- So

 

Caroline Giegerich (14:11):

That's

 

Peter Crosby (14:11):

Where- Yeah, please,

 

Caroline Giegerich (14:13):

Carolyn. I know this only because I was just on a panel with MasterCard at South by Southwest. That's where MasterCard Visa, PayPal are going to come in to bridge the gap, Collin, a little bit. They've come out with something called verifiable intent, which is a cryptographic stamp of what you are authorizing, which not just to keep the consumer secure, but also the retailer. So that, wait a second, what have you authorized to do here when you get to the retailer? They're trying to bridge that gap by saying, "Okay, we are the MasterCard agent that are now loaded with all the things, Peter, to your point about, well, there's not enough detail here to know. Is it this threshold of payment? Is it this specification of category?" They're looking to fill that as the payment intermediary, if you will, which I think is really interesting. They also are looking to fill the gap of, okay, let's say Lauren has an agent, she only authorized $100, but it went out and spent $500 on her behalf and she's now pissed.

 

(15:28):

She's making a face very funny. I'm very

 

Lauren Livak Gilbert (15:29):

Pissed.

 

Caroline Giegerich (15:32):

Who's responsible? Exactly. The security is supposed to be baked into what they're building, but it'll be very interesting to see in practice how that works out.

 

Peter Crosby (15:42):

I mean,

 

Collin Colburn (15:43):

The banks

 

Peter Crosby (15:43):

Will not be interested in handing you back the $400. They're like, "What the agent does on your behalf is your problem, not ours, I

 

Collin Colburn (15:50):

Would imagine." Yeah. And then building on that, who is going to take the liability when something goes wrong? So when the wrong item is purchased, or if the agent decides, "Oh, I'm going to buy multiples of these." And the consumer's like, "No, I actually don't need multiples of those." Or if you go back to my scam example, who's accountable? Is it a consumer? Is it the agent? Is it this intermediary that could be a MasterCard or a PayPal? So many questions there.

 

Peter Crosby (16:20):

Chaos. It's chaos, but that's part of the exciting part about being in the middle of this right now is just sort of making all of this unfold.

 

Lauren Livak Gilbert (16:30):

Figuring it out.

 

(16:32):

So different retailers have taken different approaches to this. So we've seen Amazon has kind of the walled garden, but no other LMs can scrape Amazon. They just have Rufus. Walmart is doing both. They have Sparky and they're partnering with OpenAI and every day a new retailer comes out that says, "Hey, we're partnering with OpenAI and we're allowing different features." So as you've done this research, as you've talked to brands and to retailers, how do you think about both of these approaches? Do you think there's a better option? Do you think that both of them make sense? How are you thinking about the overall approaches by most of these retailers? And I don't know, Collin, do you want to start with

 

Collin Colburn (17:14):

That one? I'll start. I know Caroline has perspective here too. I was excited to hear what she was going to start with, but I'll start. I think both approaches are valid right now. I think every retailer should be figuring out their own approach to AI and agent commerce. Walmart is obviously doing a lot in ... They're trying to zig when Amazon is zagging, which I think is not atypical in any way. They've been doing that for a while. And then you'll have others that are going to say, "You know what? We'll piggyback on what Amazon is doing, or maybe we'll piggyback and see if what Walmart does is successful." So there's those folks out there that are doing the wait and see approach. To me, I think the biggest risk is not doing anything or not thinking about what your approach is. So I think that's the one that's not the good approach.

 

(18:14):

I don't think that we've identified what the best approach is yet because things are so fluid.

 

Caroline Giegerich (18:24):

And I would agree with that. It does remind me of this point in time because I worked both at HBO and Showtime, HBO being primary, Showtime being Challenger brand. If you remember the days in which cable television went to streaming, HBO was pretty dug in on, we're not going to be first to market and offering our service through Amazon, blah, blah, blah. They were like, "This is our house." We command our house where Showtime's like, "Sign me up." And Amazon has a $56 billion advertising industry to protect. And so its thought process is we're going to block crawlers from our site. We're going to keep everything in- house. We're going to protect our advertising business by doing so. And I understand that position completely because they also don't want to become the ... I'm using air quotes for everyone listening, dumb wires of the business.

 

(19:24):

They don't just want to be the end of the line. Whereas Walmart, not exactly Challenger, but let's just put them in the Challenger bucket. They're like, "Listen, at this point, I think the last stat I read is that ChatGPT accounts for 20% of their referral traffic." So that's what is potentially being given up. And Walmart, my impression, this is no inside information, but my impression is they're like, "Let's hedge our bets in both avenues to understand the data that we're going to get here. And we can always pull back if we want to, if we're not seeing that this is going in the right direction, but we're going to be learning and testing the whole time." So I'm sorry that might be wholly unsatisfying from Collin and I because we're like, we understand both positions, but I do.

 

Lauren Livak Gilbert (20:15):

I mean, I agree. It makes sense. The advertising piece is a huge part. The Retail Media 3.0 and what does that look like in the world of agent commerce is a massive question. I know you tackle that, you're going to be doing even more research on that kind of coming up, but it's a big, big question because it's a massive profit center for these retailers. And there's a big question mark around how it's actually going to work if people just go to ChatGPT, type something in and go straight to the PDP.

 

Collin Colburn (20:44):

The really interesting thing for me, and I know Caroline knows this is a passion point of mine, is you have all these apps within an app that's the current state that we're in right now. And the big question for me is, are the LLMs going to be able to buy across those apps? Because to me, the holy grail right now, when I think about using an LLM for shopping purposes is find me not only the best product, not only the one that's at the lowest price or the price point that I'm looking for, and not only the one that's most convenient and is going to get to me when I want it, it's all those things regardless of who or where it's being bought, merchant agnostic. Are they going to be able to orchestrate that? Right now, we know that that is not possible.

 

Caroline Giegerich (21:40):

Yeah, the one product at a time suggests no.

 

Collin Colburn (21:44):

Yeah. So that to me is really, that's the holy grail. Will we be able to get there? I think how quickly? That's the big question. I don't know how quickly, because I think my suspicion is that it's a lot harder than it sounds.

 

Peter Crosby (22:03):

Yeah, which anyone that's been in retail or certainly digital commerce for the last hundred years or 20 years, depending upon which one you're talking about, knows that this stuff is hard.

 

Collin Colburn (22:15):

We have a lot of case studies that show that there are a lot of graves out there. Yeah.

 

Caroline Giegerich (22:20):

Well, and Peter, that point, I imagine there were a whole bunch of people when that news about OpenAI pulling back on instant checkout because they noticed that people were actually going to the retailer to complete the purchase. A whole bunch of retailers were like, "Told you so. "

 

Peter Crosby (22:39):

Yeah.

 

Caroline Giegerich (22:39):

"Told you so. " It's hard.

 

Lauren Livak Gilbert (22:41):

It's hard. It's hard to build commerce infrastructure. It's not like you can just be like, "I did it. " There's a lot that goes into it, so I agree.

 

Peter Crosby (22:49):

It was a relief when I saw it because I was like, great.

 

Lauren Livak Gilbert (22:54):

Slow down.

 

Peter Crosby (22:54):

First of all, LLMs, focus on some stuff, focus on some fundamentals here and don't get fancy, which is easy for me to say from the cheap seats. But I think that the seeing more focus in part driven by companies that now want to start having IPOs and things like that, you've got to have a clear understanding of where is your revenue going to come from. They

 

Caroline Giegerich (23:19):

Might've gotten your memo, it seems, from the press releases

 

Peter Crosby (23:22):

We see. Yes, exactly. Which is terrific.

 

Caroline Giegerich (23:25):

Although with Fiji Simo over there, don't hold your breath that they don't figure out how to do the whole shebang at some point.

 

Peter Crosby (23:33):

Sorry, what was the term you used? CG-

 

Caroline Giegerich (23:36):

BGCMO is the former CEO of-

 

Peter Crosby (23:38):

Oh, right, sorry. Yes, they're head of ...

 

Caroline Giegerich (23:41):

Applications, I think it's called.

 

Peter Crosby (23:43):

Yes. Very

 

Caroline Giegerich (23:44):

Strong first. But if anyone can help them figure out the whole shebang, it will be her. I just don't think ... I think they realize this isn't fully baked and we're going to pull back for now and do something better.

 

Peter Crosby (23:55):

Yeah. And I mean, I hope they do. I would love to see the full options for the consumer that they want. The question is, how baked do the initial parts of it, how baked should they get before you're putting limited resources on already trying to close the sale when the consumer ... It's always that thing,

 

(24:18):

If you only gave consumers what they wanted, you wouldn't give them much. But then at some point, the imagineeers need to think, "Oh, what really would change the game?" And so that's always the little edge of the sword here. So gosh, just this whole conversation leads to the fog of war for brands. How do they make sense of for them where to focus? When you folks look at this and think about what's important for you to start getting right in your testing and learning, in where you're leaning in. And Caroline, I'd love to start that with you. What is your advice as you wander the halls of various shows and all the great research you do there?

 

Caroline Giegerich (25:08):

Yeah, I mean, I think brands are really afraid right now because they don't control the message anymore. The LLMs may disintermediate their advertising and all of their own services, their website, their app. Bless you to Lauren. And that's scary. So I right before IAB worked at Warner Music Group, and part of my job was to work on the side of the business, which covered artist merchandise. So think Ed Sheeran's merch, Gorilla's merch. And so if all of that site information is now being disintermediated by someone going into an LLM and potentially getting wrong product specs, wrong photo, the image doesn't even match, that's kind of paralyzing for the brand. So that's a challenge. The solution though is what I'm sure your listeners have heard of is structured data, thinking about how does the machine want to understand more information about your product, and that requires a full audit of the systems to make sure that you're providing the data the way the machine wants it and not just the way the human wants it.

 

(26:29):

However, it also requires a rethinking of marketing in full. We would think like website, app, ads. Now you have to think about things. I literally heard someone at ALM say Wikipedia page. And I'm like, "Holy holy." Brands are thinking about creating Wikipedia pages now. This is really spun on its access, but they have because the control is now out here because LLMs, of course they know the brand is going to say great things about the brand where they're trying to get the trust signals from is external to the brand, but that's a rethinking about how brands show up completely and they're all trying to get there.

 

Peter Crosby (27:11):

Yes.

 

Lauren Livak Gilbert (27:13):

I

 

Peter Crosby (27:13):

Agree. Can

 

Lauren Livak Gilbert (27:14):

We have your snacks collapse of the plus 1000 to that 100%?

 

Collin Colburn (27:21):

I'll zero in on the retailers here. I would do two things if I was at a retail organization today. One is being really serious for the first time finally about what's known as zero party data and using that with your first party data to build a preference signal for your consumer. So not just using behavioral signals that we've always been really good at being able to capture like, what is my consumer searching for online? What are they purchasing? All that sort of stuff. We've been great at that. More of the like, how does this consumer want, how do they want to be treated by us? What do they want us and maybe their agent in the future to know about them and how they shop and what their preferences are? That's really all zero-party data. So I would really be looking to build that out in a much more significant way and think about that really as first-party data.

 

(28:24):

It's just first-party plus preference.

 

(28:28):

And then the second thing, and Caroline referenced this a little bit before, getting ahead of the attribution challenges that are going to exist. And I think we all know what those challenges are because if you have AI that's doing a lot of the shopping, searching, consideration, potentially even buying, then what is attribution? I think a retailer really needs to start defining what AgenteCommerce is to them. I mean, Lauren, we were at Shoptalk last week. There was a whole debate on what is Agent Ecommerce. There's clearly no consensus that exists in our industry for what is Agentic commerce. So guess what? If I'm a retailer, let's get ahead of it and just define what it is for us right now. And knowing that we might have to come back and revisit that definition, but that's okay because then that will start to frame how you think about attribution today and how it needs to evolve rather than just passively waiting for self-promotional here, passively waiting for someone like the IAB to come out with what it should be.

 

Peter Crosby (29:39):

Tap, tap, tap. We're all waiting, Collin. Yeah.

 

Collin Colburn (29:42):

We're

 

Caroline Giegerich (29:43):

Working on it.

 

Peter Crosby (29:45):

And the two concepts that spring out at me from this conversation are, what do you control in this new environment, and then what must you influence? And I feel Like those two pillars have to be really well understood and that will be a shifting thing for the next 10 years. Well, five. In five, I won't care anymore.

 

Caroline Giegerich (30:13):

We all won't have jobs in five, so you definitely won't

 

Peter Crosby (30:16):

Care. Exactly. Yeah. The agent will be doing all this for me. But in the next five years, that's going to be a completely moving target. But I think capturing both those, and particularly when you think about influence, that part you were saying like Wikipedia, I think what this means is a renewed and stronger relationship between collaboration between the e-commerce side of the house and the brand marketing side of the house. For sure. It's not only what's being said about you out there, it's what forces are you going to deploy, whether it's paid influencers or a response from your brand or whatever that is, an article in the Wall Street Journal or whatever, whatever that might be, all of a sudden these sort of traditional brand and influencer marketing things become really important to making agentic commerce work the best you can for your brand. So who are the people involved in making this whole equation work?

 

(31:15):

And it's so cool to talk about, but it's-

 

Caroline Giegerich (31:18):

What you said is so wise, and I hazard a guess that it's very limited being done. I'll give you an example, right? Yeah. Creator, where does that live in an org? Typically digital. Do you think the digital folks and the e-commerce folks are discussing how the impact of their creator marketing is actually encouraging trust signals with the LLMs? You would've hoped that someone somewhere would be doing it, but I do think what you just said is so wise in terms of this is the gold standard to head for. Even if you're not doing it today, just start thinking about how you could create a better bridge between those arenas to have those discussions.

 

Peter Crosby (32:03):

You just gave me chills because it's in my white paper. And I'm making it up. Go, Peter. Well, it hasn't come out yet, so no one would know. But thinking through all those links, well, it's been sort of fun for me and Rob to do, but also now to start bringing it to the world end of this month. I'm sorry, I shouldn't be plugging my future research on your current research thing. But you're so

 

Lauren Livak Gilbert (32:28):

Wise, Peter.

 

Peter Crosby (32:29):

I know. Well, I

 

Caroline Giegerich (32:30):

Clearly just gave you an objective read, my friend.

 

Peter Crosby (32:33):

Thank you. I'll put that on the ... If only we were doing books still, I would put it on the jacket, the blurb.

 

Lauren Livak Gilbert (32:39):

One thing I want to add to that too though is I think the corporate comms PR team too is a function that has never really been talked to by the e-commerce and digital team, but I keep flagging as like, you need to talk to them. You need to figure out who they are. You need to get them on board. They need to help with citations. They need to help with all of these things that they're already doing. They are a big part of this. So go make sure that e-comm and brand marketing's working together, find your PR and corporate comms team. They're all critical in being successful here.

 

Caroline Giegerich (33:11):

So

 

Collin Colburn (33:12):

True.That's one of the things that's so ... So Lauren, we talk about this organizational transformation stuff all the time. The mix of retail media has scaled and has really changed the relationship between brands and retailers. And then you add in this agentic commerce bit and it is the perfect storm for, I think, what is going to happen over the next three to five years in retail, which is total transformation of the way in which brands operate and these silos that exist between sales teams and marketing and trade and shopper. And then on the retail side, the way in which merchants interact with their own marketing teams and their media arm, their media network as well. I think we're going to have totally different looking organizations three years from now than what we do today.

 

Peter Crosby (34:06):

Buckle up.

 

Collin Colburn (34:07):

100%.

 

Peter Crosby (34:08):

Yeah. So the IAB is a garden of fresh data and you both have reports that are, all right, now I have to continue the metaphor, are being seeded and about to rise out from the ground. So you've got ... Well, first of all, for the research that we've been talking about today, if you just go to Google and put in IAB shopping research listeners, you will find this research at the top of your search bar. That's the easiest way to do it. Caroline, you have some new research coming out. Tease

 

Caroline Giegerich (34:43):

It. So we will be doing a further look into AI shopping data. We're going to go even further in the behaviors that we're going to be looking at. We're also going to be looking into the effect on the retailers as well. And the purpose of that is to be complimentary to what Collin described in terms of the attribution challenges. You have a world in which it's very difficult to figure out who gets credit for a sale. And the quick example I'll give is if someone has served an ad in OpenAI, for example, and the audience is matched by Criteo, for example, and then they go to the retailer site to purchase, well, who gets credit for that purchase? It is a very murky territory. And so alongside my commerce partner in crime here, we're going to be putting some serious thinking power behind how we think about that as an organization later this year.

 

Peter Crosby (35:43):

And Collin, one to never be upped by your colleague. You also have some research that is coming out.

 

Collin Colburn (35:51):

I'll be up by Caroline any day. I don't mind that at all. It's probably

 

Peter Crosby (35:55):

Inevitable.

 

Collin Colburn (35:57):

We have our Connected Commerce Summit coming up April 14th in New York. Lauren is going to be speaking there on some of the- Super excited. ... transformational organizational operating model stuff that we need to start getting ready for. And we're going to cover a lot of this. Caroline's leading two different sessions on agent commerce, two very different ones. One that's going to be a debate on is agent commerce going to kill retail media, which is going to be a really fun one. And then a breakout session on the challenges of doing agentic commerce in LLMs. A lot of what we talked about here is going to be discussed in that breakout. So yeah, if anyone wants to join us, we would love to have you in New York on April 14th.

 

Peter Crosby (36:47):

I think it's amazing that you could afford to have Lauren at your conference. She is notoriously expensive in her whole list that she needs in the decimal. Her

 

Lauren Livak Gilbert (36:56):

Entourage is my rider is free everything.

 

Collin Colburn (37:01):

I had to take a pay cut just to get her on here this year.

 

Peter Crosby (37:04):

And you are already not making a lot as it is. Well, team, I can't tell you just how delighted we are that we keep somehow luring you back to share the great work that you do that really helps all of the communities and commerce, but certainly the DSI community gets smart and keep up on what's happening. So thank you as always for sharing your data and your wisdom with us. We really appreciate it. It's a

 

Collin Colburn (37:35):

Pleasure. Thank you, Peter. Thank you, Lauren.

 

Lauren Livak Gilbert (37:37):

Thank you so much. This was a lot of fun.

 

Peter Crosby (37:39):

Thanks again to Caroline and Collin for bringing the facts and wisdom to us. So much of all that will be on stage and in the halls at the Digital Shelf Summit starting May 4th. Don't miss it, digitalshelfsummit.com. Thanks for being part of our community.