B (3:17)

Of course.

A (3:18)

So it's like no treatment, you know, training. And so yeah, for months and months I worked on training this dog. Some of the trainings I learned are, are similar to how you would train a horse. I learned, yeah, I, I went on three work, three or two, three day workshops with a series of trainers to learn how to get into this dog's mind and set boundaries. And I love telling this story one, because it was such a hard time going through Covid and I think mentally I didn't realize how much I had been struggling because you know, we're just trying to keep going with our jobs and, and this gave me a moment to pause and really check in with myself on my anxiety level levels and how it was in impacting this dog that was struggling as well. So anyway, that's the fun fact. Now I do judge other people's dog walking the way they treat their dogs.

B (4:29)

But you have a right, you know.

A (4:31)

I shouldn't and I keep it to myself. But it's very interesting. It's a, I learned a lot about myself and really just, you know, how to manage my own anxiety and you know, through the psychology of dogs.

B (4:48)

That makes me want to ask you how you're translating some of those skills to your everyday, not just your personal life, but also your work life, honestly.

A (5:00)

Yeah, I mean that's a great question because I take a lot of it in my leadership role. Right. So do I project what I'm feeling, what on the people on my team and have them feel the same way or do I take a minute to go, okay, I'm going to go in with confidence, I'm going to say what I want, I'm going to be clear about what I want and I'm going to make the appropriate boundaries. And so, yeah, it translates really well into leadership roles, working with people, understanding how they learn and what works best for them. So I apply the dog training philosophy in a lot of ways across my day. It doesn't work as well with my children, but when I'm leading my team, I absolutely bring some of those principles to the table and make sure that I'm setting the example and I'm setting the rules and I'm being a true leader.

B (6:02)

I want to take a little trip down the old memory lane of your career, which has kind of been like a sandwich where it's like agencies are the bread and tech is the filling. Kind of. Because you were at Mediasmith and OMD focusing on digital media services and then in 2014 you moved over to Yahoo and you were there for more than a decade and a couple of different roles. One was director of Client solutions And then you hopped over to AWS for around three years, and you had the longest, but also the nerdiest and also kind of the coolest title so long. Head of Global Analyst Relations for Adtech and Martech, Media and entertainment, Games and sports. It's just a lot. And then in March, you moved over to IN Ocean, the SVP of media and performance. So, two questions. One, why go back to the Bread? Why go back to agency land? And also, what is in Ocean? Exactly? Because there's an interesting origin story there with the Hyundai Motor Group. And I know you have a lot of different clients today, but it was founded as an animal inhouse in house agency for. For Hyundai and for Kia.

A (7:10)

Yeah. So the bread.

B (7:12)

Yeah.

A (7:13)

So, you know, I'll start with, you know, I. I always say I'm jack of all trades, master of some, and I like that. And I think when I started my career, you know, a common interview question is, was always, well, how? Where do you see yourself in five years? And. And then where do you see yourself in 10 years? And I always felt the need to answer that question. And what I've realized is that I am happiest when my, you know, roles in my career isn't linear. I really, you know, gravitate to roles that will challenge me, make me think different, bring new perspectives to the table, and connect the dots across worlds. So when I started on the agency side, it was the media role. And then I did a stint at Berman Braun and then went to Yahoo. And the reason I stayed at Yahoo so long is because my job changed every six months. We had new CEOs. I think we went through like five different CMOs. And so my role was really rooted in strategy. How do I address the client need but be product agnostic? And so I would help the sellers help communicate how to leverage all of these things in our bag to have better business outcomes. And it gave me just a broader lens. And then when I left Yahoo, I just wanted to do something different. I wanted to do something media. And I was very lucky that I knew somebody at aws and they introduced me to analyst relations. And so that's when I really, really felt challenged because I knew nothing about cloud and it was such a great opportunity just to learn more about what the bigger tech scape out there, how is it being used, how do we communicate it to customers? And then when Inotion reached out, I was like, well, I'm happy at aws, I'm succeeding, I'm learning a ton. I'm working with such smart people and then they talked about really modernizing the agency so it's more data led and I'm like, great, nerdy stuff. I like that stuff too. So Opposed itself is an opportunity but this time I'm going to the agency that's on the account strategy and creative side and bringing my media lens to that scope. So it's a very unique organization like you said. And Enochian is really just a creative account and strategy and marketing agencies. So we support the big autos, you know, the Hyundai brand, Kia, Genesis. I'm specifically on Hyundai and Genesis. But we've been able to use all of those elements and the technology and innovation to attract those other clients. So it's become bigger than the the Hyundai family. And so I was just really interested and curious about how we take the technology that I've been learning about for the last three years and really implement that in a marketing way which is really hard to connect. It's really hard to connect the technology and the innovation to marketing outcomes because you're talking tech to marketers who are very concerned about hey, what are my outputs? How am I doing? Is my investment, you know, working for me? And so those are all the questions that I look forward to answering and you know, really work on day to day. It's all about data in marketing world now as I'm sure you're aware.

B (11:02)

I mean all about the data is.

A (11:03)

Sort of like the data.

B (11:05)

Yeah, yeah. One of your main responsibilities, and I'm just going to read this off of your LinkedIn and then we'll sort of unpack it is is to manage the successful assessment, adoption and integration of Gen AI and influence of LLMs to enhance competitive advantage and reporting. So you also have a lot of other responsibilities, more like the usual ones for someone leading media and performance at an agency or a brand. So working with the CMO to align marketing strategy with KPIs like cross functional coordination across strategy, creative media, buying, analytics. You alluded to some of it just a minute ago. But then this gen AI and LLM stuff that actually feels almost like it's its own job completely. Although obviously it cannot be. Like it has to intersect with media in a huge way. But assessing and adopting and integrating generative AI into the media function and like figuring out how to influence large language models, that really does feel like a full time job. It just reads almost like its own.

A (12:07)

And I'm not going to take credit for all of it. We have almost a tiger team and as I'm sure you realize with the development of gen AI tools is just constant LLMs, new LLMs, what you can tap into. There's new ones every day. I think one of the stats I saw for AWS is they're working with 70 different LLMs. So I can't say that I'm doing. What I'm doing is working with our team leads to figure out how Genai can solve our business problems and challenges. Is there a fit in order to do that? I can bring media lens, but the media is always connected to some other measurement. It's connected to websites, it's connected to CRM, it's connected to our digital channels. And so in order to make the best recommendation on where Genai fits into all of this and AI, I should also note it's important to work across your teams. You can't be siloed. So a lot of what I'm doing is managing the needs across the channels with my channel leads in order to make the best roadmap or understand where Genai or AI fits into these models. You know, there's always, you know, I think there's some obvious ones. We talk about productivity, you know, how can we use Copilot to make ourselves faster? And then there's the output for the marketing perspective, hey, how can Gen save me money? How can it make media more efficient? So I won't take credit for it all. It takes a, a world of different perspectives to be able to bring it all together and, and really orchestrate that so that we can make the best recommendation for our clients.

B (14:03)

Walk me through what a successful assessment really means here. Like in the context of generative AI and chatbots and LLM, like what process and what criteria help you evaluate whether a Gen AI tool or an LLM deployment is actually going to drive real business outcomes or not?

A (14:24)

Yeah, so I was, when I worked at aws, I was on the analyst relations team, as you noted, so. So I was very lucky. I got to work with the technology experts in the space and their purview is five years out, 10 years out. What can we expect? And one of the main takeaways I got from all of the analysts that I worked with was you don't do Gen AI for Gen sake. You really need to understand.

B (14:57)

I only want to stop you because you shouldn't. People do, but you shouldn't.

A (15:00)

Right? Right. And that's what you'll see across is like, where do I, where do I get? You know, how can I use Gen AI? How can I use AI? Where, where is my Gen AI plan? And it's Great to be on the cusp and know that this is coming and it's going to have impact on your business. But my assessment always starts with, okay, what are your business challenges? Does AI solve for that or doesn't it? Okay, great. If yes, it's like a decision tree. If yes, okay, let's explore what we have already that's using Gen AI or AI, which I think coming from the back end, the technology. And I don't know that all marketers, advertisers understand how AI is already heavily influencing their media structures and performance. So if yes, okay, what are the tools in which to do that? But the most important thing I always like to say is what is your data strategy? Where is the infrastructure in which all these tools need to act on? Because they're only as good as the information that they're tapped into. So if you're in general LLM, great. Like any tool that taps into this might give similar prompts, different answers, but maybe not the answers you're looking for if you are running your business. So you want to have a data structure, you want to know where that data is housed. Is it on prem, is it in the cloud? How do you migrate that into the best infrastructure so that you can start layering tools very efficiently and then test, you know, it's, you know, what are the low risk tests you can do to prove out either the concept or the performance or whether or not it's actually saving you money in the long run. Which is I think what top concern in this, this market and going into 2026 is what are the budgetary efficiencies we can get that can be gained and it's not always a replacement of people. Right. There's always that element of human touch that I think we still need in contextualizing a lot of the data as these get built out.

B (17:26)

It's interesting because as you're speaking, I'm thinking take AI out of it and this is just a rational approach, good hygiene, the sort of way you should be thinking about pretty much anything that you need.

A (17:37)

Any technology.

B (17:38)

Yeah, Yep.

A (17:40)

And I think like Gen itself is just a subset of AI in general. And AI has been in use for what, over 20 years? 30 years probably. So it's been running in the background. All these companies, when they're analyzing data, a lot of it's ML AI. And now we have Gen AI, which is a lot more consumer facing. And so now there's a lot more awareness that's being brought to it, which opens opportunities. You can look, but I think if you root it in the business challenge and whether or not that solves for that is going to gain more efficiencies than lose, I think. And I think with the more and more companies really naturally just natively integrating gen AI into their systems, I think a lot of clients are going to be surprised that they're using it already, just by nature of using certain tools and products and strategies.

B (18:41)

Right. It's baked in and it's also becoming part of how people find information, just their everyday life as a regular consumer. But I want to talk about evaluating vendors. So putting the big guys aside, like ChatGPT, Anthropic, Microsoft, maybe even perplexity, how do you decide which opportunities are worth pursuing? And I'm asking as a poor beleaguered trade journalist who gets what feels like a bazillion pitches a day from companies that claim to have the best new AI solution out there. It's the cure for all that ails you as a marketer. And it's a little difficult for me to know what's worth writing about even. And that's on the line for me. Like it might be half an hour of my time and I'm like, I might be able to write a story here, but I mean, you might be putting money behind it.

A (19:39)

Yeah, it's, it's the wild west right now. You know, it's, it's almost, I see the journey is almost like a, you know, maybe mini is not the way to, to say, but like another little tech boom that we, that we had back in the day where digital became the thing. And so it's a little wild Westy. I can't claim to be the best assessor. What I have learned from my own tech analysts that do assess these on a daily basis is understanding use cases. If you root the technology in the use case and whether or not the vendor can actually deliver on that use case, that's a good starting point. And obviously the use case comes from your business challenge. And honestly, rates are going to come into this, costs are going to come into this, what kind of resourcing and lift so vendors can make that lift and that resourcing much more seamless. I think any marketer would appreciate that. Um, but again, it's, it's what do you need from us to make sure this tool works and provides the outputs that we are expecting. So a lot of that goes back to data strategy. Where's the housing or where is it housed? So I don't know that it's a clear cut Answer. But I always say business, business challenge, use cases and does this vendor make this? Do they deliver on what the promise is? And you know, maybe you start with a proof of concept. So a POC in order to get the vendor to show what they are capable of and that they can deliver on that promise.

B (21:34)

What breaks through the noise for you though? Because I'm asking this question again as someone who reads a lot of pitches about companies and they all sound pretty much the same, so it's hard for me to even know who is worth having a conversation with. And I assume you have the same challenge because they come to you saying we can do this and we can do that and everyone says they can do this and they can do that.

A (22:01)

Right now I will tell you, unified measurement, I've been talking to analysts about this for the three years I was at AWS and currently with, with clients now and hearing it in the market is how can we connect all of the dots. And there's a million CDPs, there's MMMs, there's MTAs. But I think being able to leverage Gen to better tell that story and fill in the gaps when there's not data available or the data needs to be built is really what attracts me to a vendor, is how can you connect everything from, from a whole consumer, consumer journey all the way from awareness media down to the lower funnel measurement. Because a lot of times it's, it's searches, getting credit for things or in market efforts or you know, those turn and burn ads I always, I always call them or the instant gratification channels usually get credit. So I think the most interesting vendors to me right now are the ones that can really connect the dots without the marketer having a heavy lift and resourcing again. Efficiencies. Efficiencies. Efficiencies.

B (23:26)

Well, sure. And if you do that then you'll most likely run into an actually necessary use case.

A (23:33)

Yeah, yeah, yeah, well and I think there's unified measurement in unified measurement platforms is used a lot and sometimes they come with caveats. We'll sell it through a certain way and say oh actually if you want all of this, you have to spend actually more than what we talked about. So digging in more again with understanding what Gen brings to the efficiency table, especially when it comes to measurement. And yeah, so that's the big thing for me right now.

B (24:12)

I mean measurement is such an enormous challenge. We're going to take a quick break but before we do, I want to segue over to search and we're going to Talk a lot about search in the second half. It's part of your job to understand how generative AI is influencing search behavior in general. But what about personally? Are you noticing any changes in your own search behavior?

A (24:37)

Yes, definitely. I just am one that uses several gen tools and go to several sources. I don't know if it's just me knowing a lot about genot everything, but enough to know that I don't fully trust outputs when it comes to answer engines. So yeah, it's change. I might go to search if I just have a word. I might go to just like Google if I just have like a one word. But if it's a question more in depth response, then absolutely. I'm using gen AI tools, all of them, all of the time to see what kind of differences and responses I get. So that's my own personal approach to it.

B (25:27)

The differences are fascinating and also alarming because the answers are stated with such confidence and a lot of the time they are right, but then sometimes they're notably wrong. And I only know that because I know that otherwise I would just be like, oh, okay, that sounds plausible.

A (25:45)

Yeah. And that's, that's one thing that I flag actually pretty consistently is my hypothesis is they will become less trustworthy once consumers or users start to see the nuances between the different tools or the responses that they get that may not really answer the question that they were asking. I personally, before I use or reference anything, go to each source that the response is linked to to make sure I understand where it's coming from and why it's responding that way. I think all users should be doing that. I always caution that and I think there might be a trust issue, a bruin. But that's my prediction with it is because I see the same thing. I'm on the advertising marketing side and you can claim to be anything you want and say you can be anything you want, but if you don't have anything to gut check and answer against and you don't know exactly what the answer should be, then how would you know whether it's a trusted response?

B (26:56)

So it's like a very sophisticated version of that comic of the dog sitting in front of the computer being like, I could be anyone on the Internet.

A (27:05)

Right? It's true.

B (27:08)

All right. So I don't know if you could hear it, but actually my dog is growling right now.

A (27:13)

Not at all.

B (27:14)

There are dogs in other apartments that are yipping, so they're having a bit of a symphony, but I think it's not making it into the recording.

A (27:21)

No, no, I don't hear it. I will say my my dog always bark. I always tell people my dog's barking at air again.

B (27:27)

So. Oh, our joke is someone is existing in the hallway. So a dog wants us to know that. All right, we're going to take a quick break and when we're back, we're going to nerd out very deeply about generative AI search.

A (27:39)

So stick with us.

B (27:40)

Okay.

A (27:41)

Foreign.

C (27:56)

I'm Sarah Sleuths, editorial director at Ad Exchanger, and I have with me here today Ludo Devalon, the product marketing lead at Amazon Ads, our podcast sponsor this month. Hello, Ludo.

D (28:07)

Yes, hello. Thanks for having me.

C (28:09)

So to start things off, what is the biggest opportunity right now for advertisers in the streaming TV market?

D (28:16)

Well, the biggest opportunity in my view is to remove the guesswork for marketers. When you think about it, streaming TV combines the best of both worlds. It's mass rich with precision and personalization. And with Amazon Ads, advertiser can achieve this personalization at scale by serving ads to specific audience based on viewer behavior while delivering broad reach. And this powerful combination helps maximize advertising impact and remove more importantly wasted ad spend. And this is really critical because, you know, the ANA has estimated that marketer on average waste 36% of their budget through inefficient targeting, duplicative ad delivery over reliance on probabilistic audiences.

C (28:57)

So streaming delivers that same mass reach people love with TV advertising, but less waste, more, more personalization. When advertisers consolidate their streaming TV investment with Amazon ads, what happens?

D (29:12)

Well, I think there are two advantages to work with Amazon ads. I mean, first of all, Amazon ad is the only DSP that has all premium streaming inventory under one roof. So of course we have prime video ads, which is our own property. But advertisers also have access now to all premium publisher including Netflix, Disney, Disney, Roku and more. And the second advantage is that we power our advertising solutions through the Amazon Ads authenticated graph. This is a unique graph which is built on verified relationship and not model data. And so in the US we can reach 90% of household to help advertiser manage through unduplicated reach and frequency and it's delivering great performance. So for instance, with the same budget, advertiser can see on average 42% improvement in unique reach for their campaign with a reduction of 27% of frequency.

C (30:01)

So we've got inventory and identity as the two unique pieces. So let's close with looking ahead. Where do you see advertising on streaming TV heading in the next few years.

D (30:14)

So I think streaming TV is really democratizing access to TV advertising. The barrier to entry are coming down with more self service options without minimum budget or year long commitment. So for instance, at Amazon we have Sponsored tv, which is our self service streaming TV solution for businesses of any size without any commitments in terms of minimum budget. And the second driver is AI tools that make video advertising creation both accessible and also very affordable. And I think this means that small businesses who could never afford TV before will join the game. And all in all, I think we could go from thousands of advertisers to potentially millions of TV advertisers in the next few years, which will unleash a new golden era for creativity with more choice and more entertainment for consumer.

C (31:02)

So we will be seeing more small advertisers entering the TV market using AI to create their ads. Totally agree with that prediction. Thank you Ludo. And thank you to Amazon Ads for supporting our podcasts.

D (31:15)

Thank you for having me.

B (31:30)

All right, we're back and I want to talk about an article that my colleague Joanna Gerber, our associate editor, published pretty recently. It's. And we don't have to talk about the tool specifically. It happens to be about this tool called the Innovid Orchestrator, which is this orchestration layer for advertising. And it coordinates and manages the interactions between different AI tools related to the advertising process. So like ad creation and delivery and measurement and optimization, all of that stuff. But I think we're going to see more tools like that coming out because I hear from the marketers that I have the opportunity to talk to that it is a problem for them that there's this fragmentation between AI tools and it feels like we could end up in this weird. You were alluding to it during the first half, this weird catch 22 situation where people bring on AI agents to help them with things, but then the AI agents don't talk to each other. And so that requires a ton of manual oversight. So how big of a problem is that fragmentation? And are you seeing any progress with orchestration?

A (32:38)

No, I see this right now. From what I've seen, it is more about how I do this. So we had the curiosity, what, two years ago, what is this? Then it was like, okay, how do I use this? And now it's too like, I need this. But they don't know why. So I think to your point, you're going to see AI branded everything. I think from all the advertising and marketing, it's something that, again, this is just my opinion in My prediction, knowing that AI is so important to CEOs, especially of companies, you know, how are we using AI? And they're asking that of their CMOs. So the CMO is going to their team saying how are we using AI? And I think a lot of these vendors are responding to that because they're getting asked. So I think you're right to assume that the AI and everything is happening and that's a selling point. Very fragmented. I wouldn't be surprised if we go through the buying up game where bigger companies are buying the smaller gen AI companies that are, you know, the companies that are very focused on gen AI and maybe don't have the orchestration, but it is fragmented. I don't see that changing anytime soon. And I think like again, if you go, if you're a marketer or your business owner or CEO is always rooting your decisions and your needs into those use cases and understanding where your data is because that is unique to each brand or company or retailer, that that first party layer is what's going to differentiate what you're getting out of these tools than anything else. And so understanding where that data lies. And I know I'm really repetitive when it comes to this, but I think the infrastructure is equally as important as the tool because the tool is going to have the fun UI and it's going to look really pretty and it's going to have really beautiful charts that could be very appealing to a marketer, an advertiser. I'm very drawn into it myself. I'm like, wow, this looks so nice and easy. But what feeds into those pretty charts is the really crucial thing to give you the best business out, you know, outputs and outcomes and, and the ones that you can make decisions on. So the fragmentation is real. I think as a company you can sync that yourself by having that data infrastructure really built out and in one place.

B (35:31)

And it's not as if marketers aren't used to fragmentation or as if fragmentation has been solved in other places.

A (35:38)

It's a cobbling of things. We often caveat, you know, our insights. Well, this is only measuring this or we can only capture this for this reason. And so omnichannel measurement or unified measurement is always going to be a very challenging thing. But the contextualization of that data is really important. What decisions can be made. We don't have data for data sake either, is that, you know, data, hygiene, all of this is connected. So when I think about the marketing tools out there, it's not like 15 years ago when we're talking Gen AI where, you know, maybe there was five to choose from. Now not only do you have to choose from all of them, you have to decide whether or not the companies you're already working with have solved the need with their Gen AI offering. And then on top of that there's a tech layer, know a data layer, a tech layer, an implementation, implementation layer that requires more than the marketer to be able to utilize all of this together. It's very complex.

B (36:54)

Yeah, it's just very complex.

A (36:56)

Yeah. And it's, it's not, it's not an easy solve and I think that's what I'm getting at is it's not very easy to decide like, hey, if I, if I had it my way, if the vendors that I'm already using and have access to my data and know how to use it and we're already getting outputs, have a gen layer that will build in more efficiencies and deeper insights, great. To me that seems like the obvious way to go, but.

B (37:24)

It'S just a lot of work.

A (37:26)

It's a lot of work. Yes. It's actually the best conclusion you can make is a lot of work. And it goes beyond Gen AI, you.

B (37:34)

Know, although I want to shift back over to Genai Search and what brands can do like actually do to influence LLMs. We touched on it briefly. So yeah, what, what can you actually do to influence.

A (37:54)

Those are protected right now I think every advertiser out there is just waiting for, you know, a chatgpt or perplexity to offer ads. We haven't seen that yet. We're seeing new browsers.

B (38:13)

So.

A (38:15)

For. Well, first I'll say I just learned a new term, geo.

B (38:21)

Well, I was going to say I hear a lot of people talking about GEO Generative Engine Optimization. I mean the way I understand that though, it's just about getting your content included and cited in AI generated answers, not just listed as a link. But you don't necessarily have like control about how that content is used or how it's reconstituted.

A (38:40)

Right. 100%. So some of the things that we've learned, so one, I'll step back a moment is there are now AI measurement tools. Where's your rank? There's a few out there. There's semrush, there's profound. I think there's, you know, there's others that are coming out or have recently come out where you can measure your rank based on certain prompts or you can see what prompts are most related to your product or the messaging you want to create awareness around. One of the ways they measure you and the things you can see is the referring source. What are they combing? What are the LLMs combing? How are they getting this stat? That's why those links that I was talking about earlier, when I don't trust all the sources, I click into all of those. Understanding how LLMs pick up their responses is really important. And some of these tools can help with that. You will see association with certain prompts on, you know, for auto, we see a lot of endemic based on the question of the prompt, we'll see a lot of the endemic partners get ranked higher up. Wikipedia ranks really high. Retailer sites rank higher as far as resources. So one of the ways to help with ranking is generating content that's relevant to the prompt that's being given by users. And being as specific as you can to answer that question is pretty important. And as you can imagine, there's a bunch of tools, gen tools that consumers have. So any one person could ask the same prompt to five different tools and get five different answers. And so that's very complex for an advertiser to influence. But building content that answers prompts, the top prompts is one way to influence the responses. The Fed is something that needs to be tested, proven out. Who knows if that's going to evolve. But, but, you know, that's, that's one of the ways advertisers are exploring what they can do when advertise. Traditional advertising isn't available.

B (41:14)

What do you look. Okay, I had to ask this. What do you do if one of your competitors is just dragging you? Right? I mean, and that gets picked up and it gets put into an AI answer and presented as truth. Of course, anyone can post anything, but when you have blue links, at least there's the semblance of doing research. Because Tracy, like, you are very unique. You're not. Well, I was gonna say you're, you're. You're probably one of just a few people that click into everything to try and understand what is being, you know, fed into an answer. And I think most people, and myself included, because I'm working fast, will just read it and if it's wrong, it's wrong. But I just wonder about manipulating it. It seems like there's a challenge there.

A (42:03)

I wouldn't call it manipulating it as much as, you know, every LLMs and gen AI tool are going to have their own methodology and reasons and algorithms for delivering on the answer, but it really depends on the prompt. And this is where we now have to work with, you know, we have to work as marketers. We have to work much closer with PR and their response or lack thereof. Or do we bad press? Do we give it air? Does it affect anything? And it may not, unless somebody's actually prompting the tool to say, hey, why is that retailer lipstick smear all over the place? I mean, they get pretty specific. So you may not have that big of an issue when it comes to negative news. And it really just depends on how big it is. Is it going across the news channels or is it just a blip on, you know, one property's editorial site? So I don't think it's easy to manipulate any of these tools. I think it's just getting as much factual content out there from trusted sources is the most crucial thing.

B (43:31)

Can you make it real for me? Like, just think of a time when you were able to guide an LLM, like an example of adjusting an underlying data input or even working directly with a tech partner to. Yeah, to make an LLM cough up what you want it to cough up.

A (43:51)

So I. My short was like, because it's so new and this is something that as an agency, we've really prioritized, really, since I got here. And we've learned along the way. What I can't address is the tactical part of what we're advising our clients at this time. So fortunately, I can. Fortunately, I can say, I know we know content drives responses, the sources in which the content lives is equally as important. And then whatever the prompt or the algorithm that the tool uses or the user prompts it to use will deliver on that. But it's. You know, I'd hate to give you a real world example, and I was very smart that you don't have to.

B (44:51)

Give me a jar of your secret sauce. I guess a different spin on that question. What is content going to look like going forward? Because I think journalism will still exist, hopefully in some form if it can monetize and journalists have an opportunity to eat and therefore be alive to write.

A (45:13)

Um, I will always hear you, I promise you so much.

B (45:18)

But I. I've heard people refer to the need to write for LLMs to really be cognizant of how a piece of content or whatever it is is ingested. So you're not just writing something that has the information that you want an LLM to pull in, but you're writing it in a certain way. So how does that impact what you do? And I assume it speaks to that closeness that you were talking about. Between comms and PR and marketing, which they've always existed in a similar sphere, but it'll probably get closer than ever.

A (45:55)

I think what we're learning, especially now that we have LLM measurement tools and ranking tools available to us as an agency, is being specific and answering a true question. So instead of going into search and saying Tulum, Mexico, you can now go into Genai and say, what can I do in Tulum, Mexico that I can't do in Puerto Vallarta? Why would I go to Tulum? So you can get more methodical with what you're asking so that you're getting more specific answers. And so I think content is going to have to. If LLM rankings and prompts is important and is a business goal or a business objective for a journalist, answering those questions is really important instead of just talking about something very generally. So that's what we've been seeing is that it doesn't have to be so specific. But, but if you're answering a question, solving a need, I think that's where content is really going to change. You know, I can't speak to PR or journalism because that's not my area of expertise, but I think as an advertiser, when you're building that content, those are the things that you should be thinking about. And it would have been really nice to have Genai when I was developing custom content ideas back in my Yahoo days because I could have come up with creative ideas much, much quicker and then, and then, you know, suggested the content themes would that would answer the questions related to whatever product or brand objective is.

B (47:43)

I have prompted LLMs for help with editing a sentence here and there or generating a headline too. And what's interesting about that experience is I know what's good and I know what's not good. And so I'll say give me multiple options. Give me 10 options for how to fix this sentence to make it sound less clunky. 10 Interesting Headlight Options for a story about this. And it might be that 9 out of 10 are not good or the beginning of one can be kind of jammed on to the end of another and then that actually is good. So I do feel needed at this point because at least I can for my own output QA, the output of the LLMs that I'm prompting.

A (48:30)

Oh yeah, I use it to refine emails so that I can be more concise so I'm not rambling. I use it to gut check. I use it to kick off creative ideas or thinking again, you know, I. It's you know, I use it with a grain of salt. You know, there's you know, journalism, you know, a headline so you can automatically instinctually gut check what it's delivering. And I can do that for my areas of expertise and what I need, but general users may not. So I'm always a glass half full kind of person. I do always raise the alarm sometimes with Gen AI and the responses you get and the ability to gut check, especially when you're looking for backups factual information, hey, what percent growth did this company earn? To decide whether you invest in it, you might get several answers, several different ways. And quite frankly I wouldn't trust a financial anything coming out of Geni at this point.

B (49:40)

You should definitely not ask questions that you will use the answers to make investment decisions or, or whatever. But that makes me, well, penultimate question, it's kind of a leading question, but it's related. How vital is it for brands to show up in the right way, whatever that means? Because like in a generative search experience, because the way that information is presented is definitive. So if you're the kind of person that would say well what stocks should I invest in? And then you get a list of stocks, then they might just go and invest in those stocks. Oh yeah, I think those people do exist. But it might not always be right even though the chatbot is making this definitive statement. So you might get, I don't know, perplexity telling you that this, that or the other brand is the best brand. And that might just be the end of research of the whole research process for some people, which is very different from traditional search.

A (50:31)

Yeah. So we, you know, I've tracked some of the trends with this and, and you know, it's the zero. We're entering the zero click search world. Right. So why do you have to click on anything if you can just get your AI overview right then and there or you're using other tools to get those answers that's affecting publishers site traffic and so they have to revisit the content that they're developing. So that's just one example of. It's crucial. It's crucial consumer behavior is changing rapidly while they're trusting the results of, of the prompts, wherever they're prompting. You know, I think the awareness of the results, like the validity of the results are going to come into question more. I think it's already starting to happen where trust, trust in the results could be a little shaky. Which is, which is, I mean it's. Right, but as a brand that's where media marketing advertisers all need to be working much more closely with their PR teams so that the right messages are getting out at a higher frequency. And whether you do that through custom content or you work with trusted editorial sources, update your Wikipedia, make sure your YouTube content has strong messaging. And the right messaging, I always say it's not just one piece, it's the sum of the parts that will help you with the LLM result and making sure that it's delivering from trusted sources. Your websites is a trusted source, your own OEM or your own retailer site is going to be a trusted source by an LLM. And so making sure the right content on there is really important. So while you may be getting less traffic, the quality of the content and the questions it answers on your site is going to be essential to making sure the right information gets delivered.

B (52:51)

Last question, different flavor. Are we heading for a future? And I know this is kind of facetious, but like a future where media buyers can make coffee or green tea or whatever in the morning, check that their agents are on track and then clock out for a while, do it all again the next day.

A (53:11)

That would be my dream. Yeah, that would be my dream.

B (53:15)

I mean, is that a version of the future that you can envision or. Nah.

A (53:22)

Yes and no. I think there's always this question of whether or not the technology is going to replace people, people. And I think in some ways it will because coding is an obvious example, like it's hours for somebody. But if you can replicate code in a much more efficient way, then you probably need less coders, but you still need that human element to be able to bring context. And you might get numbers in a chart and you might get that contextualization, but there's always going to have to be somebody there to gut check it. So maybe it's not the media planners sitting there, you know, hands on keyboard. It's going to be a different level of expertise that's going to be required than we're used to seeing. So I think there's going to be some automation, but I think with more data and more insights, it triggers more questions and the need for more efficiencies. So I think there's going to be a little bit of rebalancing. I don't think you'll ever have a role where you could just drink coffee and set it and forget it. And I think the demand for something else is going to have to fill in those gaps. But I think there's trade offs at the end of the day. There's trade offs and there's. If I was going to advise college students that are looking for their path, I would say get in deep with Genai so that you can bring that expertise to the table where it doesn't necessarily always exist in every organization.

B (54:56)

Well, set it and forget it would be pretty boring, right?

A (55:00)

Who wants to do that anyway?

B (55:02)

Think of one job that you can't automate, which is dog training, right?

A (55:07)

So way to bring it back. Thank you so much.

B (55:10)

I'm really proud of myself with that one. This episode is sponsored by Amazon Ads. Amazon Ads offers a range of products and solutions that can help businesses achieve their advertising goals. Advertising needs a world where marketers no longer have to choose between building their brand and driving results. Amazon Ads helps marketers prioritize solutions that break down silos and simplify campaign management, enabling the orchestration, execution and measurement of holistic campaigns that achieve both objectives. We remove the guesswork for advertisers by making it simple to manage all of their TV planning and buying with Amazon Ads.