Loading summary
Transcript48 lines
- [00:09]
A
Welcome to the Big Story, a roundtable featuring members of the Ad Exchanger editorial team. Every week we bring you an in depth discussion of key developments in digital marketing and media. For the past several months our editorial team's inboxes have been flooded with AI related news. We are seeing a wave of new AI powered startups as well as products that are reimagined or bolstered by AI from many of AdTech's bold faced names. The use of AI in places ranging from creative to optimization to workflows to analytics can feel like an undifferentiated blur. So our goal today is to distill what's happening in AI for all of our listeners. To talk us through what kind of trends we are seeing in these product releases. We have our associate editor Joanna Gerber who has been taking the lead on our AI coverage. Hello Joanna.
- [01:06]
B
Hello Sarah.
- [01:07]
A
And we have our news editor Andrew Bird who recently migrated from admonsters and has been covering everything from AI adopts tools to this week how cargo built AI agents to change how buyers interacted with their products. Hi Andrew.
- [01:21]
C
Hey Sarah.
- [01:22]
A
This will be a good one. I'm Sarah Sluice, Editorial Director at Ad Exchanger and your host on the Big Story. Before we dive in, make sure you mark your calendars for May 18th to 20th in Las Vegas at the Park MGM. The conference is programmatic AI. So if you like what you listen to today, you should definitely tune into the conference. You're going to get a boost on how to make smarter decisions, be more effective and improve performance by using AI in marketing. And you can use the power pod 10 for a code for a 10 discount. That's pod 10 so you can sign up through a link on our website. So now let's dive into to AI. We're a little more than a month away from our conference, which I'm excited about, and now we can talk about AI. So Joanna, AI is your beat. So I'll, I'll start by asking you this question. You're doing so many of these product announcements, these brief briefings with startups as well as companies that are kind of releasing these AI products. What kind of themes are emerging in terms of where companies are using AI?
- [02:29]
B
That is a funny question because I think similarly, broadly I could say everywhere and in some ways that would genuinely be true. I think that the phrase from soup to nuts has been thrown around a lot, but I think one of the primary, primary places I'm seeing it is optimization and specifically a B testing. People are able to create different Variants and versions of creative of an ad of just, you know, very, very small changes to the copy. And as a result they're able to optimize and see what's performing best at a much, much, much faster rate than they did before. And I think that pretty much regardless of, of what kind of ads you're creating, what vertical you're in, what platforms you're advertising across, that is something that everybody cares about right now. And I know that, you know, you made a couple of jokey comments in the past about, oh, everyone needs 110 versions of creative these days. But honestly, like, they do, because if you test 110, you can figure out which 10 are really performing and you can do that so much faster that you're not wasting a ton of ad spend on ads that end up not performing. You get those results a lot faster and sometimes preemptively. And that is another theme. We can go into a bit more later, but I think people have been doing a lot of kind of creating audience profiles using AI and you know, like basically AI audience Personas. And by testing things against those AI Personas, they're able to determine performance without testing those on real people who might be, you know, eating up ad spend that doesn't end up working.
- [04:02]
A
It feels, feels very additive, I guess, or kind of like, oh, because we did this, now we need this and now we need this in terms of what you're describing, like, we used AI to create 200 creatives and then now we need to use AI to optimize all of those creatives and then now because of that, we need to create these AI Personas to pre test them so we don't waste too much of our budget using these 200 things. It just, I don't know, it kind of, it's the opposite of being a good thing, but it reminds me of like the Dr. Seuss where they're like cleaning up the mess that just like keeps getting bigger and bigger. The cat in the hat that.
- [04:41]
B
I was thinking more if you give a mouse a cookie, you know.
- [04:43]
A
Yeah, if you give a mouse a cookie. Yes, even better brand an ad tech vendor. We got some children's book concepts here about what's going on with AI, but it's. Yeah, it's just very interesting to see it unfold in that way where one innovation that necessitates another innovation to kind of handle that in measurement or targeting.
- [05:05]
B
So yeah, I also, sorry, I forgot to add one big trend that I think is starting to emerge in a much faster rate. Really in the past, like maybe two or three months, which I think is brand specific agents. You know, obviously Amazon has had Rufus for a long time. We've seen Walmart, Sparky, and now a lot more. Even smaller brands are building out their own AI agents that you can engage with as a customer to kind of ask specific questions. And those are trained on both like, you know, brand style and brand copy and also just specific brand information. So you might be able to say, you know, why is your product specifically good for, you know, arch support if it's a foot brand? Or why is this brand better than its competitors? And engaging with that directly on a brand's website, I think is giving customers a much more kind of direct access without having to go to an FAQ page and just hope that their question is somehow succinctly answered.
- [05:57]
A
Well, so you brought up AI agents in a B2C context, but that's actually a great segue for Andrew because you just covered this Cargo agent which is using an agent in a B2B context. And I feel like we are seeing a lot of that too as a sub trend perhaps. So maybe you can kind of walk us through this as one use case of these agents interfacing with B2B brand's customers.
- [06:24]
C
Yeah, yeah, absolutely. I think one thing we can start off is like a lot of companies are dealing with like fragmentation across like selling against like ctv, selling against display, web and different things like that. So the thing with Cargo's project Kira platform is they kind of wanted to unify all of that into one buying platform. So I know one of their clients w promote is kind of integrating some of their brands into that and they focus a lot, for example on CPU, CCTV and video. So the with SSP's project here, they kind of have a chat bot where you can like put in a brand plan, put in a certain budget. You kind of want to target ctv, maybe you want to kind of focus on pause ads and different things like that and kind of infiltrate that in a certain way. They're kind of like early into the testing, maybe like four weeks into it. So as far as like revenue lift and different things like that, it's still kind of early days for that. But they definitely have seen different ways of like integrating certain creatives at a certain pace and like putting that into certain efficiency. So I'm definitely seeing a lot in that way. But again like it's still kind of early days as far as that. But as far as like certain themes that I'm kind of seeing with AI overall, like the way I kind of went into this answer is kind of thinking from a publisher perspective because I like publisher adopts. And I think in a way like I agree with you and Joanna, it's just like, you know, we're going to sprinkle AI onto everything. But I think I do see certain patterns. I think even like I did like an AI panel maybe like a year or two ago with like hers New York Times and like one of the core things that they said is that, you know, we're not doing anything new with AI. We're just kind of streamlining some of the workflows that we're seeing. So we're like seeing people speed up internal workflows, improve ad targeting, making content easier to find, which is especially good for publishers as they're dealing with like referral traffic issues, maybe like building certain products for readers and different things like that. Like I know for example we, we had someone speak from Time, Burhan Hamid, who kind of described using a tool called Glean that you kind of put together knowledge from tools like Salesforce and Google Drive. So like you can prep for client meetings, you can do marketing presentations, you can pull up like adopt support sheets and report them to clients. One we also talked to Hearst. They kind of have like a product aura which kind of combines behavioral and contextual targeting and kind of building audience segmentation and contextual audience based on similar to ways that Joanna said. Like for example, people may be interested in test food and they say through some contextual inventory we can see how maybe that can connect to people who are into outdoor activities and family vacation and kind of blend those two audience segments together. So I'm seeing it in that way. So a lot of ways we're kind of speeding up like internal processes, streamlining things that we were already doing like within the ad tech stack. It's just like automating it in that way.
- [09:12]
A
Yeah. So and I, I've talked to a couple of these AI adopts companies and I think there's definitely a spectrum between like use our tool and like every single thing and adopts will be automated to companies having a slightly more limited view. But still things like they would have the maybe like record the screen of like a 30 step view process, you know where you're going between different platforms and then have AI be able to do that process. So it kind of speeds up the work, but you still kind of got the human on the controls. But they can be like do this and then it does the 30 steps versus like it just kind of doing its own thing completely. I feel like the latter feels a little bit more realistic to me or perhaps just a more granular description of whatever. The first, the first thing is, right, where you still need to, like, Like, I. I've seen people with like, Claude code, right where you kind of like, it does its thing and then it will, like, ask a question, then you have to, like, guide it, right? And the goal is to, to let it go further and further all by itself, like a, Like a little child. We're into our kid metaphors today, but, but it's really. That is interesting to. To hear from publishers that it's kind of, I guess. Did you use the word ornamenting or kind of like it's augmenting their, Their processes? Very interesting. And of course, I'm just going to throw this, this thing in here. One thing that I've been hearing a lot of people talk about is this Dunning Krueger effect in AI, which is where people who aren't experts often overestimate their abilities. And I feel like that's why, like, when I ask AI to like, you know, you know, answer questions to me, like, as if it's a doctor or something like that, I'm like, oh, my gosh, like, it did such an amazing job. And I'm sure a doctor would look at that and be like, actually worried for my safety, to use an extreme example. But, like, I'm like, oh, yeah, like adopts. Like, of course you could use AI to automate it. But I'm sure for people in the adopts world, there's so many nuanced use cases where they know that AI would not work. Do you guys want to speak to that? And yeah, Joanna, I'm just gonna say
- [11:34]
B
this feels like a bad example because I am someone who doesn't know anything about this, so I can't say any more than the anecdotal thing I'm about to say. But I've definitely, definitely heard similar arguments against vibe coding. I've never taken a computer science class in my life. I don't know anything about coding. I've never tried to even Vibe code. But I definitely have heard some coders being like, you think you can just vibe code a whole website perfectly? Like, you know, people don't understand the nuance of something, and as a result, you might be able to get an okay rough draft. But I think it's much, much harder to revise and perfect something when you don't know how to do it organically yourself.
- [12:08]
C
Yeah, I think along the lines it just kind of made me think about just how important is the data that you put into these tools. And I know like, a lot of, a lot of like ad ops companies or AI tech companies always talk about like, you know, we have like I was talking to Cargo that was like, we have 15 years worth of like ad creatives, ad placements, you know, integrations into certain publishers that they kind of built into their AI tool. Ensure that can help build like the data that you use for certain AI tools. But I think that's why people always put on the human in the loop thing that people say, always on the calls, because you kind of need to make sure that this data that you're putting into these tools is, you know, using it correctly. Because at least for me, what I've heard in my use of AI, like AI hallucinates a lot. So even if you like put certain data in, how, how do you know that the outputs that they're getting from this data is true? Because AI, I think its core is like it wants to please the person who's giving them a query. So even if they don't necessarily might not know an answer, they'll still want to give you something to do. So that's always something I'm a little bit concerned about. Even if you're thinking about like, oh, as you said, going to AI to be subject matters for doctors, of course they could pull some data from basa, but know how do you really know that they're telling you the truth? So I'm always a little concerned about that with like using AI tools, especially if you're like using it for adopt goals and revenue and putting certain data to go in certain things. It's just like, okay, how do you know they're like compartmentalizing this data out and putting it in the output? Well, I'm always concerned about that a little bit.
- [13:40]
A
Right.
- [13:40]
B
Like AI is inherently a predictive tool. So it is creating the most likely response based on, you know, the sequence of words it's generating. And the New York Times put out an article, I believe it was on April 7, if anybody listening to this wants to look it up, about the accuracy of AI overviews. And it was saying how even if there's correct information on the site that the overviews are pulling from, sometimes it still generates incorrect information based on what started off as accurate information. So it might say, you know, this person died in this year at this age, and it might be correct. But then if you look more carefully, the overview might Then try to generate more details about the death, like a specific day and that might be completely false or made up. And it's scary when it gets so close to the truth because I think when you see something that looks 90% accurate, it's very easy to overlook the 10% error.
- [14:30]
C
Yeah.
- [14:32]
A
So I want to talk about the AI startup realm as well. I think we're kind of alluding to that with some of the AI adopts companies. I know yesterday I was talking about to a company that's trying to work on measuring for brands how they show up in AI search results. That's kind of one category of company. When you, when you look at companies that are kind of being founded by using AI, kind of selling their, their tech to brands or agencies or publishers, are there any groupings that you, that you see? Joanna?
- [15:10]
B
I think there's several categories. I think probably the main one I'm talking about these days is the type of platform that likes to call itself end to end in that it really can generate almost an entire campaign life cycle. So there are a lot of companies where the advertiser will submit a brief kind of, you know, traditionally the style they might, and from there an AI tool is able to, you know, look over that brief, synthesize the core information from it, pull out a suggested media plan, optimize based on, you know, what channels it thinks will perform best, generate audiences, run the campaign against those channels and those audiences, and then do the full reporting. So yeah, it sounds like, you know, you might say what kind of ad AI platform are you seeing? And I think there are types of platforms, AI companies that, you know, focus on one chapter of that. But I do think that kind of the From A to Z is really, really starting to come out. And a lot of those are coming from larger companies that already exist. It's maybe new tools than pre existing companies. Some of the smaller tools, I think, or not smaller tools, smaller, newer AI companies tend to be focusing more specifically on dashboards and reporting. But I would say a lot of the new tools are kind of a everything in one combo platter.
- [16:30]
A
And I feel like that seems like maybe a trend not just in startups, but the cargo thing you described, Andrew, that feels very end to end. Certainly all the big tech platforms are end to end. Like put this in, we'll make a campaign and optimize it for better or worse. So that's really interesting. And I think there's also the companies that are trying to understand the, you know, the AI and the Geo or
- [16:59]
B
of the Smaller startups. The two biggest trends I've been seeing besides, you know, end to end are probably reporting and analytics startups and then geo startups, everybody. Well, there's GEO for Generative Engine Optimization. Then some people are calling it aeo, which is answer Engine optimization, or aio, which is AI Artificial Intelligence Optimization. All of which are the same thing. Which are basically will help your brand understand how it's showing up in AI search and we will help you optimize so it shows up more frequently and more highly recommended. Which I think is an extra interesting conversation now that a lot more chatbots are offering advertisers. Like, do you want to be advertising in a chatbot or do you want to be optimizing so that you don't even need to advertise in a chatbot because you're showing up in the organic search results? I think that's a really interesting kind of two sides of the coin right now.
- [17:48]
A
You know you're in a happening category when there's three different three letter acronyms.
- [17:56]
B
Every time I write an article I'm like, oh, which acronym will I use today?
- [17:59]
C
Literally. Yeah. Honestly, on the line of geo, it's something I've heard like publishers talk about, like something that they think about. But at least to the publisher that I've talked to recently, I haven't heard them use any GEO tool. So maybe something that I'm like going to be thinking about as I talk to more publishers is maybe what geo specific tools or AEO specific tools that they're looking to use. Because of course that's something they're thinking of because they're having an issue with audiences coming to their site. But I would be interested to see like if they're kind of building their own internal tools or using stuff like that or you've even heard some.
- [18:33]
A
And I think ultimately though there's going to need to be a company that is tracking like, you know, what publisher content is powering what results. So it's less about you know, clicking through that fine print of like citation, like how many people read footnotes. Right. Versus giving credit. And I think that will be a really interesting company that solves that problem.
- [18:59]
B
So, Sarah, be so wary of saying that on the podcast because I'm going to get five pitches next week. I was just going to say we are the first company to ever do this.
- [19:10]
C
I'm about to say you're giving people business ideas on the podcast
- [19:15]
A
while someone, someone's got it. Okay, so let's talk a little bit about. I Guess the, maybe the, the reason for being, for why these companies are doing this. So, you know, is it just that, you know, their CEO said that they need to release an AI product and people are kind of slapping AI on it and that ultimately these products won't be more efficient maybe than just calling up a company and saying, hey, this is my brief. You do this or whatever. Is it saving time or is it improving marketing performance? What kind of things do you hear from these companies you're interviewing that either underscores like, wow, this really just saved a ton of time or boosted performance versus it being kind of that AI washing that we, that we saw definitely, you know, 18 months to two years ago, I feel like that was more of the dominant AI. Has that changed?
- [20:16]
C
I mean, I think just to kind of start off, I do think we're getting a little bit of that, like, oh, we need to put out an AI product because that's the kind of the thing to do right now. So I don't want to like disengage that. But I think a lot of times, a lot of it is to kind of save like workflow time in general. Because I mean, a lot of people are talking about like putting campaign briefs into chatbots and like streamlining how fast we can make creative, how fast we can put these through the bid stream, how fast we can sell these ads, how fast we can kind of connect at least from my side, the sell side to the buy side and different things like that. Even if you're thinking about like the sell side agent talking to a buy side agent. I know, I just kind of did the article talking to Optimal Weather Company News Core about like them developing their own sell side agents. So a lot of that is kind of like adding a conversation layer into like the auction in the bid stream to kind of like streamline some of these kind of contextual signals that they kind of say get lost in the bid stream. So I do think a lot of times they kind of want to streamline and make a lot of the operations that they do faster so they can think more strategically. Like a lot of the, like, like beginning adopts level stuff like manual data and, and putting in stuff like that can kind of go to the AI and they can kind of think a lot more like strategically. That's kind of one thing that I've definitely been hearing a lot over time for sure.
- [21:34]
A
And Joanna, thoughts on AI vaporware and the prevalence of it.
- [21:40]
B
I'm thinking a lot about AI for the sake of AI versus AI for, you know, Actual effectiveness, which is obviously kind of the same thing that seems to be spinning through your mind as well. And I think that for a lot of companies it really is somewhere in between.
- [21:57]
A
Right.
- [21:57]
B
Like I do think that for a lot of companies they are creating tools that are genuine necessities or if not necessities, at least benefits. But at the speed at which they need to evolve right now to stand out in the marketplace, I think people might be rushing them and maybe not creating the best version of a tool because they want to get there first rather than best. And I don't know if anyone's seen that new documentary called like the AI Doc or, or why I became an apocalyptimist. And they interviewed Sam Altman in the documentary and said something very similar. It's like everyone is currently kind of racing to be the AI company that does things first. Obviously this is like the big players. We're talking Anthropic, OpenAI, et cetera, not like small ad tech startups. But the idea was like we need to trust that the people who are on top are also the ones that we can trust to do it best and most safely. And you can't just be doing it to get to the finish line because if you are, a lot of privacy might be sacrificed and you know, a lot of errors might be made in terms of data and analytics and processing. And if that's on a small scale it might mean your media isn't optimized. And if that's on a large scale it might mean that, you know, your data is going to the US government in like what it shouldn't be.
- [23:10]
A
So I think, and there's no, I think there's no question that AI is very powerful. I think it's what things can it be very powerful over within the advertising and marketing space. That's still a question. Thinking of the headline, I think from earlier this week, Anthropic has a mythos release that it's actually not releasing because it finds zero day vulnerabilities and it's, it could be so, you know, misused by hackers that they're like, we're not going to release it at all because. But actually if you are like a cybersecurity researcher, we're going to let you use it, but like everyone else, no, because it could just like break the world if hackers get ahold of it. And then so we'll see what happens. You know what, what new programmatic ad fraud will be created by the AI adpocalypse? I Don't know where. Is that the name of the documentary?
- [24:08]
B
No, the. The documentary was about apocalyptimism, which is trying to find the balance between the apocalypse and optimism. It was actually pretty interesting. Yeah. Showing in theaters near you.
- [24:18]
A
Okay, well, I'll have to watch that when it gets to my couch, to be honest, but it sounds like a good one. Okay. So we started talking a little bit about doom and gloom, so. Which is a great segue to. One of my last topics I want us to discuss is there's a lot of talk, right, about AI replacing jobs. I know at the agency level in particular, a lot of this work. What happens to FT models when you have AI doing all of these things, these 30 step processes that used to be done by humans. And I feel like people have abandoned the idea that, oh, it's going to free people up for higher level work. And that while that may be true, I think people now acknowledge that job loss is inevitable. So I see two nodding heads there for people who are. Fun video. So. So tell me a little bit about how you see companies kind of answering this question when you bring it up in your interviews and what is the partyline now? And do you buy it and do they buy it? What's the deal?
- [25:32]
B
I was really hoping you would bring this one up because one of the things that scares me the most is that I don't have to ask half the time. Companies are excited and eager and willing to say to me, and this, you know, is so effective and so quick and speeds things up so much that we can do the same work with a third of the people. And it dumbfounds me every single time. They are effectively now bragging about layoffs and bragging about the efficiency, because it's all efficiency. That's. That's the buzzword. That's the key word. And unlike other buzzwords, we know exactly what it means. The. It's not exactly jargony. You're not left wondering what it means. It means we can get more done in less time with fewer people. And I think people are looking at it so monetarily and so much from a financial perspective that they just think fewer people better because they are thinking about it in terms of the company's best interests rather than the individual's best interests. And that is something that scares me. But I also don't think that's something that's necessarily new to AI. I think that is something that's been very, very true in the way that, you know, us businesses have functioned for a Long, long time.
- [26:44]
C
Yeah, I mean kind of piggyback on that. Like at the end of the day like people who works in adopts, like of course you want to get the right ad to the right person, you want the creative to look good. But at the end of the day these are businesses and money at the end of the day is kind of what keeps these, you know, businesses open. So I do think the, the thought of like AI being more efficient and kind of taking away some of these jobs is just like appealing to them as like crass as it is to say. I do think that is a little bit appealing to them and I think specifically for entry level adopts jobs is where a lot of the brunt work is going to be taken. A lot of the, I mean I think I've had conversations with a lot of ad ops people over the last couple of years about how entry level adopts is kind of disappearing a little bit and how like, you know, more senior people are just given different roles and different things to do. So I think that's where a lot of the hurt is going to be. And I think like thinking on the publisher side, I don't know if they're outright saying this as maybe as you're talking to agency startups, like I think, you know, one of the things publishers always say to me, you know, like AI is not going to be writing this type of content, it's not going to be replacing this. But I think if we're reading in between the lines, like that's kind of the way the industry is going in my point of view.
- [27:52]
B
Yes, people love to say human in the loop and a lot of the time it's a loop of AI and then the last like centimeter of the loop is a human clicking approve.
- [28:02]
A
Well, I guess I'll throw this to both of you too since you're earlier in your careers. Like the whole narrative of AI removing entry level jobs, like how is that hitting among your peer group? I mean you guys aren't like fresh out of college, like you're kind of beyond that stage, but you're not. So it's not so far in the rear view mirror.
- [28:23]
C
Yeah, I mean I think for journalism in particular that is definitely a concern because I mean we already have like AI companies kind of scraping a lot of content training these AI tools kind of how to write. Like I think now we're still early days where like you can easily tell if an article is written by AI. There's just certain kind of tells that you can kind of see But I think as it continues to learn and train and like, you know, I think there are some companies, like I always go back to this example, I think it was the LA Times, maybe I'll do some research. But when they went public, like the new owner basically told the writers, like, you have to use AI when you're doing your work. So it's just like the fact that, you know, like companies are kind of thinking that way, it does kind of make me already think that when, like, even when I was in, when I was going to school, like some of my professors used to say, you know, journalism is dying. And that was even before the AI trend was, you know, happening.
- [29:16]
B
So people have been saying journalism's been dying ever since journalism was born. Journalism isn't gonna die. I feel confident, yes, I do think that I've seen a like radical shift in the job market in the like, what, four or so years since Andrew and I graduated though, because I would say like I graduated in 22 and most people I knew had a pretty easy job time getting a job out of college, even if it wasn't, you know, the job of their dreams. There was a job market and I would say by two or so years after that it was so much harder. And you know, I'm not saying that was solely single handedly AI's fault, but it definitely played a role and it definitely played a role in what people are looking for in entry level jobs now. Like my, my good friend is a project manager at a tech startup and she was recently telling me that she was having one on one conversations with people saying, you know, how are you spending your time? Do you have enough work to fill your day? And one woman was like, I actually don't have enough work to fill my day anymore because I've learned how to successfully automate it. And they were like, oh, that's awesome. This is actually what we want people to be doing. But the next step for her is apparently going to be to help train other people on how to automate their roles. And I'm like, okay, so once everyone's being ten times more effective in automating, like what's everyone gonna do with the extra time? I mean, maybe pick flowers and write poetry, but I don't know.
- [30:36]
A
Yeah, well this is really, yeah, it's something I think that no one really has the answer to right now. And I think a lot about how in some cases efficiency increases consumption of things. Like, I don't know, like most people used to be farmers and now they have other jobs. And there's other things that people do now that we don't have to spend all of our time growing food. Like, it kind of. It did literally free up time. So I don't really spend any time growing my food except for when it's as a hobby. Right. And same thing with, you know, factories and consumption. It made people, you know, just buy a new pair of shoes instead of repairing them at the cobbler, that kind of thing. But I don't really. I really don't know, like, what happens if that. If that pattern were to hold or, like, what new things we'd be spending our time doing. But I have a little bit of optimism in terms of the fact that we've figured this out before as the human race that's like, feel like we're really philosophical for the Big Story podcast, but to kind of bring it home for the marketing world. When you used to have, literally back in the day, you'd have one TV commercial, you'd run it for a year, you'd then do one measurement thing, one media mix modeling would happen on an annual basis, and then you'd, like, figure it out in terms of what you should be doing the next year. And that's all they could do in the age of, you know, paper inboxes and secretaries and, you know, you know, two martini lunches or whatever. Right. Like,
- [32:23]
B
those three things sounds great.
- [32:25]
A
Efficiency. Right. Whereas now, like, if you did that, you would die as a brand. Like, it just wouldn't work. And maybe there's the same amount of people doing that, but they're able to run this faster because they're using all these tools. So I think that standards are being raised and I think the competitiveness that AI creates, yes, maybe some people will let go and the job will be more efficient, but I also think there are other companies that are like, well, what if we just, you know, maybe we let go one person, but then we keep the rest, and all of a sudden we're supercharging our marketing and it's like, so much more complicated because that does give us an edge. So I think that people are going to be forced into or see opportunity by, you know, not just using it to replace things, but to make it truly additive. That's my. That's my super optimistic view. There's that, like, it's a good note to end on. Right?
- [33:26]
B
Yeah.
- [33:27]
A
Okay. Well, thanks for hearing our mix of doom and gloom and optimism and distilling how AI is being used for all of these different products in the marketing space. There's definitely more to come there, so keep listening and we'll see you next month at Programmatic AI.