← The Digiday Podcast
Loading summary
Transcript114 lines
- [00:00]
Host
Foreign.
- [00:11]
Kameka McCoy
Hello, hello and welcome to another episode of the Digiday Podcast, a show for anyone wondering how to employ AI while remaining employed. I'm Kameka McCoy, senior marketing reporter here at Digiday.
- [00:24]
Tim Peterson
And I'm Tim Peterson, currently employed as executive editor of video and audio at Digit Media. How's it going, Kamika? We were at the DJ Programmatic Marketing Summit last week, so we got to spend some time in person as well as with a lot of programmatic experts, especially brand and agency execs. And that the question you laid out there seemed to be a big question on a lot of those folks minds.
- [00:49]
Kameka McCoy
Yeah, I think a lot of the conversation during this event, which I want to say I'm still recovering from every time I take a trip from California back down to Atlanta and I have to recalibrate my body and my brain. But I imagine that there's a lot of marketers doing the same. After our conversations about AI and just how far it goes, you kind of had half of the folks at the summit who seem to be team, you know, AI can do a lot and we're gearing up for it to become, you know, a media buying agent. Some have already kind of started there, others have put a stake in the ground and say, absolutely not. I want humans doing my media buying. And AI can do everything else. So it was really interesting to see how the room was kind of split. What was your take?
- [01:32]
Tim Peterson
Yeah, I mean, especially considering our guest for this episode. So we did a live podcast recording during DPMS and we had Ellie Uberto, who's the director of marketing at Duluth Trading Company, to talk about how agentic AI intersects with Duluth's marketing practices. And one thing that I thought was really fascinating that she said on stage was that they don't. Duluth as a brand doesn't necessarily want to be at the cutting like the very cutting edge of using agentsic AI in their marketing, but they're using agency AI in a lot of ways that put them towards the cutting edge, including, she said they're using agentic AI as part of their bidding processes, AKA having AI agents be involved in the actual buying of ads, which based on all the other conversations we were having throughout the event, is the very bleeding edge of using AI agents in advertising. We had Scott Ensign from Butler till present a couple case studies of tests that that agency has done in December and then in February of using AI agents as part of ad buying. We also did a live game at I created as a way to like mock out. Okay, what is, what does it actually look like to prompt actual AI agents to serve as representatives for a buyer in an ad deal negotiation. And I thought it seemed like that went over pretty well and a lot of people enjoyed being able to like get hands on with, okay, yeah, how would I prompt a buyer agent to negotiate for me and how would I respond to a seller agent's counter proposal?
- [03:25]
Kameka McCoy
I think that encapsulated a lot, right. It gave you like a real life experience in a moment to look at like what exactly we were talking about here over the course of the three days of the summit. Because how many winning teams? There was one winning team, team one. And then let me back up, there were 10 teams total. And I feel like when you looked at the roster of like the teams that have gotten it right, there was obvious like team one, the win team. But like everybody else had like really major tweaks that they still needed to be able to come to a deal with these agents. Which kind of like, I guess was very reflective of how much work needs to still be done for there to be an, an understanding of just how far these agents are able to take the media buying process. Shout out to the team one, that, that one though, they got it right.
- [04:14]
Tim Peterson
Yeah. It is also like the interesting thing with that is just by nature of it being a game and us only having so much time, we had to cap how many rounds of the negotiation there were. And also I didn't have an actual human seller on the end to decide what deal to go with. We left that entirely up to. In this case, it was Gemini. Was the large language model powering this for Gemini to decide. And that was the thing a lot of the folks said is I would still, in a real world scenario, I still want a human on the other end of this to not only be making the decision, but I think it was Greg Langer from Havas Media who made the point of like, I would like to have the option of okay, if I didn't win the deal in the agency negotiation to still be able to call someone up at the publisher and say, help me out here. Can we do something? Which, yeah, I mean there was. I don't know if anyone on stage used the terms human in the loop, but human in the loop was a big thing of a lot of folks as interested as they are in bringing AI agents into their workflows and outsourcing some parts of the work to agentsic AI. Everyone seems to be drawing lines somewhere. It's just a question of where exactly those lines are being drawn at Duluth, obviously they're okay with having AI agents involved in the bidding. Butler Till is determining where are the lines when it comes to the point of transaction. But then you spoke with Glynis Richards from Bayer on stage. This was the first session of the event. She's drawing her lines very clear and very earlier in the process.
- [06:10]
Kameka McCoy
100%. I thought it was actually really interesting that we started the conversation with Glennis. She was the first. We were the first to have a conversation, and she had a really hard line in the sand. And then we ended it with Ellie, which had pushed that line a little bit further to actually, like being embedded in the media buying process. But with Glennis, the hard land line in the sand, excuse me, was wanting to have humans in the loop just because of the limitations that exist right now around a. And her school of thought is more so around kind of what Greg was saying, where, like, a person has the nuance more than an AI machine does. Also, you've got to consider that Bayer is a pharmaceutical company. So, like, if anybody needs nuance and, like, targeting and buying and things like this, it's going to be a company that's a pharmaceutical company. So she was very, very clear on, you know, guardrails and kind of a line being drawn in the sand in terms of, like, how far AI agents are allowed to go with their media buying practice. And when it comes to putting a dollar down, Glenn is not a fan of an AI agent being the one to do it.
- [07:11]
Host
Yeah.
- [07:12]
Tim Peterson
And I think it was Emily Proctor from OMD who we were talking about, like, kind of their getting started guide for clients to be using agentic AI. And I think it was in that conversation where we got to talking about, you do need humans. Like, you can't outsource all. All of the decision making to AI because ultimately you're going to have a client who says, okay, so why did we do this? And if you say because Claude said so, that's not going to be a satisfying answer. That's going to be the kind of answer that could get you out of a job or, like, lose a client. Emily didn't say that, but that was kind of the implication of that, of why you don't want to outsource everything. And that was what to outsource and what not to outsource was a really big topic in one of the behind closed door town hall sessions that we have where brands and agency execs are able to speak freely but are kept anonymous in these sessions. So we're able to report on what gets said in there? We have an overheard article on the site recapping some of the conversation from the day one town hall, in the day two town hall. So much of the conversation was people talking about, okay, we get it, there's a lot of value, a lot of reason for me to have AI agents do some part of my work, but the importance of needing to figure out how much of the work to outsource to them and how to not like outsource all of the thinking. What was the line that someone in that room said?
- [08:49]
Kameka McCoy
Yeah, there was somebody in the room who said, quoted copy and paste will get you replaced. Which kind of goes back to what you were saying here about the nuance that's needed. If you're unable to explain to a client why an AI agent made a decision, especially when it comes to their money, you really run the risk of taking yourself out of a job. But someone else on day one had kind of put this into perspective and compared AI agents to a Roomba. At some point, if it breaks, you've got to have someone to fix it. Right? Same thing with AI agents. So it's seems like a lot of people were kind of coming to this, coming to terms, excuse me, with AI being a tool as opposed to the tool, if that makes sense.
- [09:29]
Tim Peterson
Right, yeah. And there's that like really complicated dynamic when it comes to junior level and entry level employees of. To your point with the Roomba, how do you get these people to learn how to use agentic AI as part of their workflows, but still develop the underlying skills and knowledge to be able to sufficiently supervise these AI agents? Like if you don't know, if you don't have a solid enough understanding of what the AI agent is doing, how can you check its work or the quality of its work? On the other side of it, you're going to have employees, the junior employees coming into the workforce with a better familiarity, knowledge, experience with using AI, that's kind of just part of their everyday lives. And so they may be better equipped to use AI for their work, but needing to again find that balance of them not just having outsourcing everything to AI, like you know, it's we're going to have people who come up with self driving cars or assisted driving cars, but you still want those people to be able to take over the wheel when they need to, if they need to, as opposed to them only understanding, oh, I know the right buttons to push.
- [10:58]
Kameka McCoy
Absolutely. There's still a case to be made for knowing how to do long division by hand here.
- [11:06]
Tim Peterson
That was wild. That was one of the wilder points that came up. Is like, the case for long division.
- [11:12]
Kameka McCoy
Yes. It still matters. And I think AI is really kind of revealing that in a lot of these conversations that we had in getting motivating, keeping people motivated to use the AI as a tool and not become overly reliant on it, so that they kind of handicap themselves in being able to do the work. So I think there was a lot of great conversation. I feel like we've moved past kind of the hypotheticals in AI significantly between kind of last DPMS and this one.
- [11:43]
Tim Peterson
Yeah, no, I completely agree. You know, DPMS in December was very much of, like, what understanding do people even have of, like, what agency AI is, let alone its applications? Now everyone's aware of its applications. There are people who are applying it. Again, one of the really interesting examples from the town hall was a person who talked about how they've used agency AI to automate part of their quality assurance workflows to help their junior team members. And a lot of people in the room got very interested in that use case. I'm still interested in learning so much more about how that person develop the skillset to be able to build that tool and then how they're now taking advantage of the time that frees up for them and their team. Like, okay, if AI can speed things up, can automate tasks that a person would be spending hours, if not days, if not weeks doing, then what do you do at that time? And how. So how do you justify that time? You know, something that we get into with Ali is, you know, this question of what happens to the billable hour.
- [12:58]
Kameka McCoy
Yeah.
- [12:58]
Tim Peterson
When it takes fewer hours to do work.
- [13:03]
Ellie Uberto
Absolutely.
- [13:04]
Kameka McCoy
We have a lot of good talking points. I'm excited for the conversation with Ellie. So with no further ado,
- [13:22]
Tim Peterson
So the
- [13:22]
Host
brand POV on agentic trading, I guess maybe best place we should start is kind of a baseline setting. Coming into this event with Duluth, what is the status of agentic AI being
- [13:38]
Tim Peterson
part of your programmatic advertising workflows at all?
- [13:41]
Ellie Uberto
All right. So I have to be honest, there's a lot of it that's, like, proprietary. So I can't just get up here and, like, you know, spill everything we're doing.
- [13:49]
Host
We'll get to it.
- [13:50]
Ellie Uberto
Yeah. But what I will say is, like, you know, a lot of the way that we're leveraging it is obviously to, like, you know, make our advertising more efficient and effective. And, like, so there's parts of our, you know, advertising that we're obviously willing to, like, hand over to agentic AI and then parts that we're not. Right. Like, our brand voice is incredibly important to us. The way that we talk to our customer, our sense of humor. Humor. So that's something that, like, we hold very close and you know, like, that, that would be a part of our flow that, like, you know, a lot of the conversation over the course of, you know, the last couple days has been like, what are you comfortable handing over? What are you not? And I think, you know, we've thought about that quite a bit and talked about it a lot. Like, that is one part that we've held really close. Whereas, like, you know, the way that we bid on our ads, you know, the, the versioning of creative, you know, like those sorts of things that, like, free up a lot of time. That's something that we are, you know, much more comfortable handing over to.
- [14:54]
Host
Have you handed over bidding to AI yet?
- [14:56]
Ellie Uberto
Yeah, like, that would be something that, you know, is like, we're much more comfortable doing.
- [15:00]
Host
Right, but are you already doing that?
- [15:02]
Ellie Uberto
Yeah.
- [15:03]
Host
Wow. When did you start doing that?
- [15:06]
Ellie Uberto
Great question. I mean, where I sit within our organization, like, I'm not like the person that is like, hands on doing that. Right. Like, that's something that like, lives with it with our agency partner. Right. So for like, on the brand side, like, we're much more hands on with like, for example, our creative team is leveraging AI on a day, on a daily basis as they are building out our creative. So, like, there's no more sitting down and building out like individual versions of ads. That's something that we entirely, you know, not entirely, but like, AI is much bigger part of that now. For example.
- [15:40]
Interviewer/Producer
So for you guys, agency, that's who handles the bidding. Correct?
- [15:44]
Ellie Uberto
Correct.
- [15:45]
Interviewer/Producer
What was kind of the process to get comfortable with that? Because that's one thing that we talked about a lot is kind of like, where do you say, okay, yeah, I'm comfortable handing this off. What were kind of the signals that made you comfortable there?
- [15:55]
Ellie Uberto
I mean, I think actually this is a part of just one of the discussions that happened earlier today. But it's like, I think we're comfortable with it because we know that our agency knows us very well.
- [16:05]
Kameka McCoy
Right.
- [16:06]
Ellie Uberto
So there is a human element to it. You know, it's not just like handing it off without any kind of checks or balances. Like, they understand, like, what our goals and objectives are. They understand, you know, like, all the, all the different pieces that like our marketing needs to hit.
- [16:24]
Kameka McCoy
Right.
- [16:24]
Ellie Uberto
So it's not just about like, do we have to hit a sales goal. There's all sorts of other goals that our brand has. So I think that's what makes us comfortable with the portions of our business that are where we are leveraging AI right now.
- [16:39]
Host
So you all were kind of a ringer when we did the agentic game
- [16:42]
Tim Peterson
show the other day.
- [16:42]
Host
Because you were on team one, you won. You have experience with bidding using AI agents.
- [16:48]
Ellie Uberto
I mean, we're just, we know how to prompt AI.
- [16:52]
Host
What was that experience like for you? Like, how much did it match or was completely off the mark from the actual experience of using agentic AI as
- [17:01]
Tim Peterson
part of the buying process?
- [17:03]
Ellie Uberto
Like the, the game.
- [17:05]
Host
The game.
- [17:06]
Ellie Uberto
I mean, again, like, I wouldn't be the person that's actually doing that work on a daily basis, but I imagine
- [17:14]
Host
your agency may have like, walked you through it at some point just for you, like the need to know stuff.
- [17:19]
Ellie Uberto
Yeah, I mean, I think, I think the way that we, the way that we participated in that activity is like, I wouldn't say like super similar to the way that it happens at a deal. Right.
- [17:33]
Host
You can say, Tim, it was trash.
- [17:35]
Ellie Uberto
It was not trash. But no, I mean, I think like, it's more about the, the mindset of how that took place. Right. And like the way that we're thinking about how are we reaching a customer. Right. So less about like the actual like hands on of like what prompt did we give the agent and more about what does the agent need to know about our customer? How are we reaching the right person? You know, I think like one of the reasons that like we won is because we thought about like, okay, what's the household income? Like what, you know, so it's like those very specific things that they need to understand about the person that we're targeting. And I think that's really important for us as a brand because we're not out there to just reach anybody who wants to buy pants. We have a very specific customer that we want to reach so that we're, you know, getting somebody who is going to come back and purchase from us again and again. You know, somebody who's going to be like a real loyalist for our brand. And so that's, I think, like, you know, going back to that earlier question around, like, why are we comfortable handing over certain things to AI? It's because, like, we really understand our customer. We know our agency understands our customer. Yeah.
- [18:45]
Interviewer/Producer
After some of the experiences, the game, the conversations, the town halls and things like that, and you guys are leaning into agentic. How are you feeling in terms of like having to go back to work after this? Is it we're moving fast? I feel good. Or we're not moving fast enough? We've got to do more.
- [19:03]
Ellie Uberto
I mean I, I think we're moving at the right pace for our brand. Right? Like so are we, are we, you know, at the forefront right now? Like, no, but I don't necessarily think we should be. You know, there, there are parts of the way that we need to reach, for example, our current customer where AI isn't necessarily the way that we need to speak. You know what I mean? So we're doing it in the way that we need to, I guess.
- [19:37]
Interviewer/Producer
Are there any apprehensions of being a first mover?
- [19:41]
Ellie Uberto
It's not like apprehensions of being a first mover. I wouldn't say that. It's more, it's more just about like. Let me think about how to say this. It's more just about thinking through like what's the right pace for us? And again like, because we are like our brand is so like we're a brand that tells stories, right? And like that's something that we want to hold onto. We're a brand that is like, you know, we like our sense of humor is something that we need to like very much again also hold on to. So there are things like that where you know, we're still like that's something that is always going to be a really like hands on piece for us.
- [20:30]
Host
What are the parts that you don't want to be hands on about? Like with having AI agents in the mix, like obviously everyone's effectively outsourcing some part of their work. But also like as the brand you don't need to be involved in all of the nitty gritty anyways. And so what are the things that you don't care to know as a brand in terms of how your agency or the DSPs you work with or anyone like down the supply chain from you are doing when it comes to agentic AI as part of their processes?
- [21:03]
Ellie Uberto
I mean, I guess there's like, it's not like there's, I mean I want to know everything but like, I mean, I guess you know, kind of going back to like what is the exact prompt that you're giving, you know, the agent? Like that's probably not something I need to know, but I need to know like the spaces where we are using it, you know. So yeah, but I think where we're comfortable handing it over again, it's like where can we be more efficient, where can we, you know, be quicker? Like, those are the areas where I were comfortable doing it, and, like, you know, again, the areas where we're not. It's like our brand ethos is what we need to hold on to. You know, like, AI might be able to, like, understand, like, the syntax of how we speak, but they can't understand the ethos.
- [21:46]
Host
What are the areas in which you want to be quicker?
- [21:49]
Ellie Uberto
Again, it's like, you know, it's. It's the bidding, it's the, you know, burgeoning of creative. It's, you know, like, all those things that, like, help us get our. Our ads in market faster and getting them in front of the right people faster. But, yeah, why are you going through this process?
- [22:09]
Kameka McCoy
Like, you know, as we prep for
- [22:11]
Interviewer/Producer
this conversation, it was a lot of unknowns, but given we're on the last day of this. Right. Is there anything that you feel, anything that you feel you've got the answer to now or something that you kind of want to press on when you go back?
- [22:27]
Ellie Uberto
I mean, I actually think one of the things that I feel like after the last couple days is like, I do actually feel like we're doing a lot of things really. Right. You know, I do. A lot of the conversations that I've overheard, I'm like, again, going back to, like, I think we're doing things at the right pace for us. You know, I think I feel really confident in the relationship that we have with our agency partner and how they understand our brand. And, you know, so it's like. It's almost like a renewed confidence in the way that we're working right now.
- [23:00]
Interviewer/Producer
How did you guys talk to your agency about wanting to experiment with some of these agentic processes? So, for instance, when I talked to Glenna, she said her and the Bayer team raised their hand to start experimenting. That is in there early. What was the process there for you guys?
- [23:16]
Ellie Uberto
I mean, it was sort of the same thing. It was a little bit of a hand raise on our side. And also for us, how can we bring it into our organization as well?
- [23:28]
Host
How have you brought it into your organization?
- [23:30]
Ellie Uberto
So that's another area where I have to be a little bit careful about how much I share out publicly. But it is bringing it in in a way that helps us to be more efficient in the way that we're doing our business.
- [23:42]
Host
And what's been the process in getting yourself, but also your teams up to speed on what's possible, how to be using these tools for us?
- [23:52]
Ellie Uberto
It's it's really looking at again, it's like, our workflow and where can we free up time? So I think that's been another theme that I've heard a lot, is like, is AI going to replace us? Right. And the way that. That I've always viewed it is like, AI, like, is there to, like, get you to, like. Like to get you to that place where you. The 10%, right. That you. You're spending before A.I. you spent all your time getting to that place where, like, there was 10% left to do. And then you're, like, burnt and you can't get that, like, final piece done. And now I can kind of get you to the finish line and you get to spend all your energy crossing the finish line. Right. And so it's like, for our team, now that you have AI to kind of do a lot of that, like, base work, how are you reinvesting your time?
- [24:39]
Interviewer/Producer
How are you guys reinvesting your time? What now becomes freed up to do now that AI has done the.
- [24:46]
Ellie Uberto
I mean, it's a lot more. We get to spend a lot more time on strategy. You know, it's a lot more. We get to spend a lot more time thinking and doing strategy, and it's a lot less time, like, building the deck and, you know, like, doing, like, the reporting process. Right. Like, we get to just digest the data and figure out what to do with it next. So, like, those are the areas where it's really freed us up.
- [25:09]
Host
How are you thinking about compensation model for your agency? Like, the billable hour doesn't really fit in an AI landscape because workflows are a lot different there. So how do you change the agency compensation model? Moving away from billable hours, or do
- [25:26]
Tim Peterson
you move away from billable hours?
- [25:29]
Ellie Uberto
I mean, that's not really. It's not really something that we've had to, like. That's not something we've had to deal with yet, I guess. Yeah.
- [25:39]
Host
Is it something you have to be thinking ahead to, though?
- [25:42]
Ellie Uberto
Not right now, no. I mean, perhaps the future, but not now.
- [25:48]
Interviewer/Producer
I'd be curious also kind of about the guardrails. As much as we talked about efficiencies and things like that, we've also talked about, like, hallucinations and, like, data safety and things like that. What guardrails do you guys kind of have in place to ensure that those things are protected and doesn't go off the rails?
- [26:04]
Ellie Uberto
So I think the way that I think about it is it's like. It's like you. You used to just like, hand your data over. Right. And like, now it's more like we're lending our data and AI is giving us the result. Right. So it. Yeah.
- [26:24]
Interviewer/Producer
How do you lend your data?
- [26:25]
Kameka McCoy
I know.
- [26:26]
Interviewer/Producer
I think it was Henry who was up here talking about kind of like a librarian that exists that taps. Is that similar for you guys?
- [26:31]
Ellie Uberto
Yeah. Yep.
- [26:33]
Kameka McCoy
Nice.
- [26:33]
Host
What data do you not hand over, lend?
- [26:38]
Ellie Uberto
I mean, we actually like our. Because our customer is like. Our customer is so, like, our current customer is so specific. And like, we have. We've really, like, built this, like, very strong customer base over like, many, many, many years. And like, we just know who they are. So, like, there's not really like a fear of, like, hand there. There never has been really a fear of like, handing that data over, like, whether it's AI or, you know, to the algorithm to go out and find a lookalike model. You know what I mean? Like, that's never really been a fear for us because it's such. It's just like very solid data that brings back a very solid customer. And we're also in a period of time of like, really trying to go out and find a net new customer. So that's another piece of it too, where it's like, yes, we have this current customer, and yes, finding more customers that look like them is very, very valuable. And also we're building a new customer profile.
- [27:44]
Host
What about the concern? Because I feel like this came up in the town hall yesterday early in the discussion around sensitive information or just information and the level of comfort of letting a large language model access it. I think the fear there was if the large language model has access to it, it's conceivable that anyone could, with the right prompt, then get that data through the large language model. What are the guardrails? Is it just like some data just does not touch an AI model? I know there are AI models you can run locally on a computer that is never. That doesn't touch the Internet, like an air gap computer. And that can be a way to localize it. There's retrieval, augmented generation systems. Is that something you all are having to think about? Like, what's the data? That it's solid data. It's good data. And maybe that's the reason why you wouldn't want an AI model to touch it.
- [28:41]
Ellie Uberto
Yeah, I mean, yes, it's something that we're having to think about. Yes. Yeah. But like, I mean, it's a push and a pull. Right? Like, it's a push to pull. So it's something that we need to think about for sure. And there's, there's a benefit to like allowing that data to you know, like bring in again like these like very qualified like buyers. Right. So it's, it's a little bit of a push to pull and you know, we also feel again it's like this confidence in like how it is being leveraged by our, our partners. Right. So like just knowing that like the proper precautions are being, you know, taken.
- [29:29]
Interviewer/Producer
There's a lot of trust happening here.
- [29:30]
Ellie Uberto
Yeah, well,
- [29:33]
Kameka McCoy
I guess that's the foundation of an agency, right?
- [29:35]
Ellie Uberto
I mean that it is, it is the foundation of the relationship, right. Like, you know what I mean? Like that is something that's really important to us and like, I don't think that like we would, like that's why we're in the relationship that we're in. Right. You know, I mean that's a part of the process like in selecting the partners that we have is having that trust there and making sure that like when that selection takes place that like we're having these discussions in the upfront so that we can feel that confidence. Right. And it's like continually having those discussions as we're going as. Because things are changing very, very quickly, right. So you can't just kind of like sit on it and not be continually talking about these things. So I think that's where the trust comes from. You know, I think the open conversation, doing all this hands on work, right. Like so, you know, having workshops where they're coming to our offices and talking to our teams about, you know, how we can better be leveraging AI, how they're, you know, so it's just continual cycle and I think, you know, I know there's a lot of agencies here and so I would encourage them to be like doing the same things with their clients because then yeah, I think it, it like I heard all the concern that was taking place in, in these rooms and that's why I was saying earlier, it's like I kind of felt like, oh, like we're kind of doing it, right. Like I don't have those concerns, I don't have those fears. So.
- [31:01]
Tim Peterson
Must be nice.
- [31:02]
Ellie Uberto
Yeah, I, that's honestly how I felt. I was like, I feel pretty good about shit.
- [31:07]
Host
So how often are you having them come in to do those workshops?
- [31:11]
Ellie Uberto
It's on a relatively regular basis. I mean I think the last time there was like one couple of months ago.
- [31:19]
Host
You know, so it's so it's like once a quarter.
- [31:23]
Ellie Uberto
I mean it's not like it's not scheduled like that regular, but like, I want to say it was like a couple of months ago was the last time they were in. And, you know, so it's like, it's a. It's a regular conversation. Right.
- [31:32]
Host
How do you decide or like, talk with them about deciding what the scope
- [31:36]
Tim Peterson
of each workshop should be?
- [31:37]
Ellie Uberto
So we have a task force in our office, and so it's like the, you know, we have a lead at our task force that's working with them, you know, so it's like a conversation there and they'll come in and.
- [31:48]
Interviewer/Producer
Yeah, yeah, I see how you have an agency partner and you trust them. They're humans. These are people's faces you're looking at. But is there, like, how does that trust then extend to the limbs that they're using? Is there a halo effect that's happening here? Or did they say, you know, this is my approval list of alums that can be used that type of deal?
- [32:09]
Ellie Uberto
Good question. I mean, I think again, like, that would be that, like, falls a little bit out of my, like, purview, so probably not the question for me, but.
- [32:23]
Host
So that's one of the things, like, you don't have to. You don't need to be involved in
- [32:27]
Tim Peterson
picking the AI models.
- [32:29]
Ellie Uberto
No.
- [32:30]
Tim Peterson
God, please, never.
- [32:32]
Ellie Uberto
Yeah.
- [32:34]
Host
So coming out of this event, anything that you're going to be going back to your team or to your agency of. Hey, I'm really thinking about this in terms of how we want to evolve our agentic AI strategy going forward.
- [32:48]
Ellie Uberto
Yeah, I think. Yes, I think so. I think there are some areas where, to your point, while I do think we're moving at a pace that is appropriate for us right now, I think especially coming out of this, there's probably some discussions that we could be having around, like, what is next for us. Right. Like, are there areas like, do we want to move a little bit more quickly in some, in some spaces, are there areas where we could, particularly as we're thinking about, like this net new customer that we're. That we're targeting, could we be leveraging it a little bit differently there? Right. So I think those are some spaces where there's probably some additional conversations that we could be having.
- [33:35]
Interviewer/Producer
Absolutely. I really appreciate you spending some time with us today to walk us through what this looks like for Duth and kind of how you are thinking about it individually, but it's been a fantastic conversation. Thanks for having it with us.
- [33:46]
Ellie Uberto
Thanks for having me.
- [33:50]
Host
Thanks.
- [33:57]
Kameka McCoy
Well, that brings us to the end of this episode. Of the Digiday Podcast. Thank you to everyone for listening. And please don't forget to share this episode with someone who you think would enjoy it. You can even rate us and leave us a comment on Apple Podcasts. We'll be back next week with another episode of the Digiday Podcast. Thank you so much for joining us.