Bonus: How DÔEN, Origin & Universal Ads Actually Implement AI w/o Losing Their Brand | Whalies Panel

14:34Ashley Kick:forecasting or decision-making velocity oh goodness i mean at this point i really use moby i guess the most regards to i feel like they're kind of like my analysis like helper you know whereas you know like we just launched our summer collection today shopdoen.com um and you know we had a lot of like changes in the behavior you know for this collection that we've seen in previous collections. And so I was actually over there earlier today, like, furiously asking Moby, like, what happened between the hours of 12 and, you know, 10 and 12 today versus 10 and 12, you know, on previous launch days?

15:14Ashley Kick:You know, like, that would have taken me so long to do with a bunch of, like, spreadsheets and whatnot. So is AI, like, this massive thing for Dolan at this point? And the answer is no. It's it's taking small steps that are gradually getting us there to what at the end of six months or at the end of the year will be a monumental difference or it better be year over year that's really where we're taking it i mean at this point i'm not right now going all in on anything you know like with any one thing um you know it's like back in the day everybody way back Back in the day, everybody just sent direct mail, right?

16:00Ashley Kick:And everybody's like, ah, then our physical inboxes were getting too full. So everybody then switched over to email. So, oh, wait, now our email inboxes are too full and our physical mailboxes aren't. And so, you know, it's this pendulum swift, right? Pendulum swift shift. There's my word. Swift. Taylor Swift may have worn a Doean dress last week. I'm still obsessed about it. It's swiftly shifting. It's swiftly shifting. and when it moves one place, I think there's negative space left, you know, where the pendulum used to be. And it's like, and so as everybody's moving over there and everybody's got, you know, AI people twirling around on, you know, ads, that's obviously overstating, we're gonna look different.

16:42Ashley Kick:They don't, maybe not necessarily can see exactly how, but, you know, in that space is where I want us to be, at least, you know, while we all still figure things out. So, you know, before, like now people do a little bit of email and a little bit of direct mail. And it's the blend of the two that really, I think, is the sweet spot. And like finding your sweet spot as a brand is going to be different for everybody.

17:07Eric Dyck:I love what you just said. And it's something I've been thinking a lot about, just like AI and the branding of AI. I think that actually over time that will go away. We'll stop using AI in everything. And it'll be more about what is the value that this product is giving the end user. and it'll be branded as such, right?

17:24Justin Parker:That makes sense. I've already seen that. You don't talk about the light switches anymore, right?

17:27Eric Dyck:Exactly. If you go into ads managers now, you'll see that they're slowly taking out the word AI, right? Because it's just value. It's just how auto bidding works. It's just how auto placement works. You don't need to call it AI anymore.

17:39Martha Ann Pavoni:That's interesting.

17:40Eric Dyck:So I think that's the big evolution. And it's interesting because for you, that value hasn't been shown to you yet. Whereas for me, I can't build and be competitive with the majors without having AI completely embedded in everything we're doing. So it's so crazy because that evolution will happen over time. It just hasn't given you value yet. Absolutely. Yeah.

18:00Justin Parker:It's happening very quickly. We just were reviewing a report actually we did last year with Triple Whale on the state of AI. And we were reading it and it was like, was this written in 2008? We were still talking about how your hands can't render, things like that.

18:14Ashley Kick:We still thought I Am Robot was going to be our 2026. Absolutely.

18:18Justin Parker:From our perspective, we're a newsletter and podcast. We acquire new users on Meta. And just last week, we hired our full-time agentic media buyer named Gary. And he's outproducing the lead generation agency that we were working with.

18:34Ashley Kick:But your AI is named Gary? Or is there an actual person named Gary?

18:37Justin Parker:No. Gary was actually the guy that taught me how to media buy way back in the day. But this is a robot.

18:41Ashley Kick:Can you have the robot in Superman? Like, you're number four. He's like, that's a name. He's like, so is Gary.

18:46Justin Parker:It's just me. I just put on glasses. No. It's a full-out robot doing all our media buying. Now, we also have a creative strategist named Blanche, and she's not as good yet. She's not as kicked in. But Justin, I think one of the things, the promise of AI is being able to scale without proportional headcount. It's a big promise. What roles or workflows are changing the fastest inside your organization?

19:07Martha Ann Pavoni:For us, it's the analysis and reporting, right? Like, we don't have a BI team anymore. We simply go to Mobi and just ask questions. And I think that from there, the natural next step is going to be like marketing briefs, creative briefs. It's going to be probably some creative generation, most likely copy first for us. I could see some like, again, for us specifically, PDP imagery, that sort of thing, like things that's literally just like switching a color on something just because of the nature of our business.

19:41Justin Parker:Yeah, that makes sense. In an organization that has fully gone in on Mobi, do people still use chat and Claude? Or is that sort of like an organizational risk in a way because those programs don't have the context that Mobi will have?

19:55Martha Ann Pavoni:Yeah. So outside of the front end, the marketing side, people still do use the other programs. We're trying to roll it out a little bit wider in the company. The thing with Mobi is exactly what you said. It has all of the context. It has all of the definitions. And to the earlier point a couple panels ago about it being a harness, right? Like you're getting the newest version of whatever's out there. And you're also getting the talent of an awesome team that's working for you to pull together a product. They're the ones that are living on the bleeding edge of what's possible with AI. You know, as an operator, I don't need to and shouldn't be spending my time learning what's the most like current thing with AI.

20:39Martha Ann Pavoni:I need to be making the company better. And if that includes AI, great, but that's not my primary role. My primary role is selling product. And so me spending my time on the things that the team here at Triple Whale is doing, it's duplicative. Like I just rely on Triple Whale for that stuff.

20:58Justin Parker:Yeah. Martha, back to TV a little bit here. I think companies like Universal Ads have been really, again, democratizing and the ability to buy TV. I think for a long time, people thought there was a big barrier to that where you'd have to call up a station or buy huge blocks of TV, right? But in the age of connected TV, it's made it much easier to do. Maybe a little future casting. How do you see TV buying evolving in this AI world where you could just sort of have things bought programmatically maybe for the right campaign to reach people on the good screen?

21:32Eric Dyck:I think it really comes down to who you're partnering with, right? if you're using a triple whale, what your objectives are. So this is by no means a one-fits-all, but you should absolutely be able to buy that through a centralized, like going back to the MCP connector I was saying, if that's hooked up to our campaign management and multiple platforms are doing that, again, Mobi could be that connector, right? The connector definition is different for everyone in this room. But yes, that should be able to be done through one interface, and that interface does not have to be an ads made in a journal future state.

22:07Eric Dyck:That's what I think. If you read meta blogs, if you read our blog, that's kind of what we are future casting right now. And that's okay. That's what we expect.

22:17Justin Parker:I'm interested, Justin, back to the agentic media buying thing, because the way that we did it with Claude was essentially we had a long conversation with Claude about how we think about media buying. And sort of we imparted, okay, here's how we think about testing. Here's how we think about scaling. and that was our input based on what we think are best practices. I'm also like, AI can also read the new product releases on Gem and Lattice and MCP. So it knows the technicalities of maybe how you should media buy. How much like human input have you had on how your agents actually buy?

22:51Martha Ann Pavoni:A tremendous amount, especially in the beginning. We have a director of paid media at Origin and I actually took over paid media while we were doing the implementation so that I knew exactly what was going on. And it's one of those things where the amount of human in the loop has got to be proportional to the risk that's involved with the process, right? If it's something that's customer facing, if it has the potential to cost a lot of money, if it's not easily reversed, you've got to have a human in the loop. And you also have to have the right human. AI is great at making good operators even better.

23:29Martha Ann Pavoni:But it's also great at making poor operators.

23:32Justin Parker:It will gaslight the hell out of you, that's for sure.

23:34Ashley Kick:It will lie to you. It will straight up lie to you sometimes. You want to know something funny? Yes. A while ago, it wasn't Mobi, but I was asking an AI a question about the cannibalization of my retargeting to my email. And it was like, oh, they're cannibalizing. You really need to consider cutting back on your retargeting spend. I didn't. And the next week I asked it again. And it was like, oh, they're working so well together. They're really doing a great job of being additional touch points and lifting each other up. And I was like, thanks, AI.

24:07Martha Ann Pavoni:Yeah. And that's the advantage of having the guardrails, right? Like if you take the time to set it up in the beginning, you can actually, that human in the loop isn't encountering the same mistakes, the same learnings that you went through earlier on. And even part of that is like, we think to the phrase, the exact phrasing of your question, we think we know what's best, but we're very often just as you would with any employee that you're doing a one-on-one with every week. We asked Moby, like, what are we missing? How would you set this up if you didn't have your guardrails, right? Like having genuine, like as genuine as you can have with an AI, having a genuine conversation and being open to the fact that you might be missing something is super important in actually like making progress, iterative progress in the A and B.

24:54Justin Parker:So you got to know when to impart what you think on AI, but also when to, you know, be judged by it. For sure. Interesting. And it sounds like your situation, Ashley, is less human in the loop, more AI in the loop. You seem a lot more guarded, right?

25:07Ashley Kick:I just don't want to destroy it all, you know? Yeah.

25:09Justin Parker:And I guess your example there is how you've seen that, how it can get carried.

25:13Ashley Kick:I could have like shut down all of my earliest part of my retargeting. And then, you know, that's super high Rojas traffic, you know? Yeah. And so it's, it's, you know, it's, what is it, you know, measure, measure twice, cut once sort of thing.

25:28Justin Parker:Part of that problem, I think in theory with Mobi, that it should do that less because it has access to so much historical data.

25:34Ashley Kick:Yeah, I'm a big fan of Moby moving things into SQL tables. I'm like, okay, thanks, Moby. Pop that over. And then, you know, then you kind of get to, like, test it. You kind of get to poke around with it, spend time with it, maybe update different variables within the SQL and see if it's holding up. Like, you know, that's like, you know, it's just taking, having it get me to the right data and information more quickly is how I'm utilizing it right now. But like I said, we're taking baby steps in it, but we're taking larger steps in how it's disseminating. And so that's, you know, we're not 100 % in the dark ages, but, you know, we are certainly being much more cautious.

26:16Ashley Kick:And I think, yeah, it's just that's who the brand is. We're very transportative, like in our photography and like who we are. We're very, it's very grounded. And so for us to start making moves that I think would counteract what we are and who we are as a brand and what we value as a brand could be problematic for sure.

26:40Eric Dyck:I was just going to say, I think AI can be value negative when it's siloed and not connected to the data foundational source of truth. and so as companies are getting into this space it's hard to do this especially when you're a long tenured company but you really have to start with the health of the signals that are informing the ai engine and that's where listen i'm the first one if you come talk to me afterward i'll say one of our biggest strengths right now is we're last to market we're getting to build everything zero to one in a world that's ai native and that is a freaking power play yeah right you don't have to carry over an entire ad stack to the new AI world.

27:24Eric Dyck:I worked at LinkedIn, Snapchat, Twitter. It's really hard at the same time. Wow, like meta, Google, everyone's crushing it. It's just how fast they're going. But it is pretty incredible when you can start day one with a clean data slate where everything's pumping into that same AI engine.

27:42Justin Parker:Justin, on your AI forward approach. Does your team, have you had to be really explicit with your team about what they are empowered to do and not when it comes to using various different tools? Or is it sort of like, does the best bubble to the top? Or have you had to come out and say, guys, hey, we're using Mobi only and then maybe Claude. Like, have you had to be explicit?

28:02Martha Ann Pavoni:With my team specifically, we are Mobi only because it has OpenAI and Claude in it. So there's no, with the context that it has, there's no real advantage to going outside of the mutually owned system of Mobi. So the other parts of our organization, they'll use other AIs and stuff, but we're specifically Mobi with my team. Yeah.

28:24Justin Parker:And then how would you answer that? Oh. Sounds like there might be some edicts.

28:27Ashley Kick:I mean, I'm just certainly not as forward. And right now, a lot of my team, we try to keep it all like one in, one out right now. And so I don't have too many people in the kitchen at this point. Basically, what I have come to rely on are my Mobi support staff. Because right now, we're very slow in rolling this out. And so as I started to build out things, I have weekly calls with either my account manager or the poor guy that got assigned to me to teach me MOBI too. And so they're incredible. So if you ever have any problems or issues or definitely everything that you've been talking about, not necessarily me, they're definitely here to help and they will build reports for you.

29:20Ashley Kick:I know that my triple L team has been.

29:23Justin Parker:Nice. Most important question, are we still using MDashes? They're such a useful piece of punctuation, but you can't use them. you got to use semicolons. No one knows how to use them. It's one of the biggest problems in our space right now is MDash. If I was MDash's agent, I'd be very upset.

29:39Martha Ann Pavoni:I pull them out of everything.

29:41Justin Parker:Yeah.

29:41Martha Ann Pavoni:Just because it looks like it's AI created.

29:43Ashley Kick:I know, and I used to love them.

29:45Justin Parker:They're such a good piece of punctuation.

29:46Ashley Kick:Is it because I'm old?

29:48Justin Parker:Well, if you're that old, you'd like semicolons like me.

29:50Ashley Kick:Exactly.

29:51Justin Parker:So maybe to leave everyone here, if people out here are using disparate tool sets for their AI, maybe in a siloed fashion, what's the first step they could take after this talk today to get unsiloed, get on the same page and get building recursive loops with great AI?

30:07Martha Ann Pavoni:I think the people in this room are the, they're the closest to being able to make change in their organization. You guys have the best proximity to the data. You have the tools to actually make change happen. And some of the biggest wins that we've had have actually been cross-functional wins. Like we've pulled in data from other parts of the business and married it with the e-com and marketing data to actually come up with like changes that need to be made to our products. So we're handing that off to the product team. So I would encourage you like, be the evangelist for your company. Go out there and like, get the data, put it in, ask Moby a simple question, like find me the one product that needs to be fixed.

30:50Martha Ann Pavoni:And then you can bring that to the product team if they're not AI native. I think in most organizations, it's a bit of a, you have to do the legwork. They're not going to come to you. They're not going to adopt it natively. So you have to be the one that goes to the other branches of your organization.

31:10Justin Parker:Sets the expectation.

31:11Martha Ann Pavoni:Yeah.

31:11Justin Parker:Anything to add?

31:13Eric Dyck:Yeah, I think the sexiest AI is the AI you can't see. It's the AI that's driving ROI. So I would say, start with why am I building this? What am I trying to improve? And is it going to improve the outcome I'm expecting it to drive? ROI, start with that. But yes, I would say the best applications of AI are not visual. They're ROI driving.

31:41Justin Parker:Data-based, right? Based on how you use the data. And any final word from our AI pessimist?

31:46Ashley Kick:Oh my gosh, yeah, right? Talk to your developers. If you have a dev agency, they've probably broken this on somebody else's site. And yeah, I would say just try not to go it alone. Talk to your industry peers. Most likely you're not the first person down whatever path you're on. And always more voices are better than just yours probably.

32:10Justin Parker:Beautiful. Well, special thanks to our panelists for this awesome Whaley's panel. My first Whaley's. I had a great time. Thanks for everyone for showing up.

32:24Justin Parker:Thanks so much for listening to today's episode. If you're not a subscriber to our newsletter, you can do that right now at directtoconsumer, all one word, dot co. I'm Eric Dick, and this has been the D2C Podcast. We'll see you next time.