How Is AI Changing The Media And Advertising Industry?

All right, folks. Sorry we're starting a little late. That's my fault. My last panel ran long. Maybe this panel will run long too as well. All right, let's let's jump off. We are talking about this thing. I heard this agentic AI thing is like a it's like a thing. Like it's like it's it's going to transform the industry. And that's what we're here to talk about today. Where we are, where we're going, what does our agentic future look like? Where are we on the maturity curve of agentic? Um is a lot of the topics that we're going to cover today. But before we dive into that, I'll introduce myself. I am Tony Katsur, CEO of the IAB Tech Lab. Who's familiar with the IAB Tech Lab? This has been a great can cuz the show of hands has been really high. Even some brands are like we know who you are. I'm like you do? Oh my god. Um and I will let the rest of the panelists introduce themselves. Hi, hi I'm John Roberts. I'm the Chief Innovation Officer at People Incorporated. Hello everybody. I'm Kimberly Jones and I'm the CEO of Butler Till, we're a media agency that specializes in the messy types of clients. So lots of regulated industries, health care, financial services, and multi-local. Speaking of messy, I'm Chris Blackhead. I'm the founder and CEO of Deep Intent. We're a health care DSP. So know the space well. I am Tyler Romasco, EVP of Commercial at OpenX. So this is a great representation of the industry to a larger extent. We have a publisher, we have an agency, we have buy side technology, we have sell side technology. The only thing really missing is a brand but I would argue not all of the brands are necessarily deeply immersed in agentic as of yet. Maybe next year they would be on the panel with us. But this is a great representation of the industry. So quick show of hands in the room. Raise your hand if you're a publisher. Okay. Raise your hand if you're an agency or brand. Okay, great. Raise your hand IF YOU'RE ADTECH. [laughter] MY PEOPLE. OKAY. Our people. Okay. Our people. The qualifying question was asked. We know who's in the room. Okay, great. Let's start with a level setting question for the group. We're all A gentic is the theme. I think last year felt like it was heavy CTV measurement even and then AI was just entering the picture. And this year it's it's the it's the flip of that. Um but this is something that's been around for maybe eight months or or so the concept of a gentic on our ecosystem. The core protocols that power it are not even 18 months old. Where are we Uh John, let's start with you and go down the panel. Where are we in the maturity curve of a gentic in our industry? I think at CES people said everyone's talking about it, nobody's doing it. And I think now we're a little over the line to like real proof points, real case studies. Um in our case I think the moment it became very real internally is when we sat eight different agents in their own Slack channel who can now do full-blown analytics. They each have access to different data sets. And because it was built by an engineering team, they named them after Sesame Street characters. [laughter] Lesson to the room, how you name an agent determines its personality. If it's called Elmo [laughter] And and yes, Oscar is a Thank you. Grudge is a much better word than what I was going to say. [laughter] Um word to the wise. It is amazing to see it starting to work. We have sell-side agents. We built a lot of our tech in it. It's really exciting to see it start to become real and also find the place where it's still hard, difficult, and has to be hammered through. It's great when you get to the real problems cuz then you know you're doing real work. Yeah, I think that we are maybe just past the point of experimentation. We ran our first fully agentic campaign towards the end of 2025 and since then uh slowly but surely more and more of our clients I did mention that we work in healthcare so tends to be a little bit of a conservative uh category but more and more of our clients are interested in starting to test. Yeah, I want I want to dive into the healthcare aspect of this a little bit further down the down the panel questions. Where where are we? I we I know it's FIFA World Cup right now and we're in Europe but I only know baseball analogies. Um like are we are we first pitch, first inning, where are we in the in our journey? Yeah, probably uh sorry. Probably um I I think in terms of activation I think it's still early but to your point I think we're beyond the the proof of concept. I think like we have working prototypes of how it could work and how it is working. I think um I mean I think we all have used agentic you know some sort of AI assistant in some way. We see our clients using AI now a lot in the planning process and I think that's like depending on if your question scope to just activation or other parts of the uh of the process. Like I I actually I I see it fully being used on the planning process right now and we have tools that we've implemented uh with our clients and like we're seeing 50% plus improvement in the planning cycle times. Like they're just able to churn out really solid proposals faster and with better quality and precision. Is that because there's just more efficient discovery of content, discovery of audience? Like it it the efficiency of brains and all that. We're putting data in the hands of people who need it and it's like a lot of it I think it's just like we're making decisions faster with more data than ever before, right? And I think it's kind of like the analogy uh that I tell my team is like there used to be a you know a job called computer, right? And like that if you guys watch there's like a I forget the name of the movie but there's a movie about like like in NASA they had like a bunch of women like doing like the actual calculations. I'm totally blanking on the name now. Hidden Figures, thank you. Uh that used to be that used to be a full-time job called computer and then they created the calculator and now the calculator it's allowed us all to do more strategic work and kind of up level the whole conversation, right? Like do more types of innovative work And like I think that's where we're kind of in that transition moment now with AI. Tyler, where are we? You from the cuz there's a lot of uh I I see buys This By the way, this is the first time I've ever seen this shift. Buy side aggressively moving into this where I see sell side a little more like let's test around with this. But buy side's like, "Let's go." What are you seeing? What are you seeing? I think that's true. I might go a little bit different. It's It's funny. I think the supply side is coming along quickly, right? Es- especially as you start to think about supply side decisioning and intelligence, which I think we'll get into a little bit later and things like that. I'm going to go a little bit different cuz first of all, I agree with every single person here. Um to answer your question directly, Tony, we are in the top of the third. Wow, you put us that far? I am. But [laughter] But it's it's for a very specific reason. Um so, the way I was ground people is when we think about programmatic, go back 15 years, go back more than that. We've always been an AI-first industry, right? So, machine learning, big data models for decisioning, right? For us, I'm giving us top of the third because our experience is further along, meaning how AI and humans interact to drive value for and outcomes for businesses, either your own or your partners or things like that. So, while agentic and generative AI is very different and is propelling us into new areas and and solving new things, I think we have an advantage over other industries because we've been through similar things before. And we've experienced how those things can last and and be uh in harmony together to drive business and outcomes. So, I'm giving us a little bit more based off of just the experience and our backgrounds and where we come from because I feel like we're a bit more adaptable overall. I think I can get behind that kind of analogy. Yeah. You like third inning, you shocked us, and then you laid out your argument well. Well done. to say something different. Um Butler, till you mentioned uh you did your first fully agentic buy in December of last year. Yes. When when did you when did it click that this is moving from this is not a buzzword anymore? Like this is true tangible impact on you and your clients business. When did Was there a moment where you were like there was there was the aha? Like oh my god, this is this is a thing. Well, I think as an independent agency we're always trying to be on the forward-leaning side of innovation. And so we're continually looking especially to our partners for opportunities for those test and learns for those beta opportunities. And so I I I guess I'm trying to think if there was ever a moment. I mean it's just in our nature to want to be first. And so part of it was was wanting to kind of lead the pack a little bit and make sure that we were learning so that we could apply. I mean there wasn't any expectation of higher performance necessarily out of the gate. But we knew if we didn't start somewhere, we would never get ahead. Right. How about a deep deep dive? Yeah, I mean I think just in terms of the adoption curve internally like we it was a very deliberate. I think after CES, I think when like the new Opus model, was it 4.6 at the time, came out? Like that was everything just became possible. Like we were able to now integrate this technology and how we were coding I mean we like I'm now actually developing again, which is fun. I mean not a lot, but like I can actually just did my first PR. Like I pushed some feature to to prod. CEO is coding. CEO is Wild, dude. Short the business. My engineering team hates me for it, but it's fine. No, but I I think it's actually more indicative of how it's like actually starting to become like all the barriers are collapsing. Like product teams, engineering teams, business teams, you need the context to feed the AI in order to do the job that it needs to do. And I think that like streamlining and integrating all these teams together, like our team sizes are moving into smaller more discrete units or kind of like smaller teams of developers and product engineering teams. And they're just moving way faster than our larger like monolithic way of building code. And it's just it's like time efficiency is just incredible right now. But did you did you have a moment where you're like, "Oh, wow, this there's efficiency here." Was it Uh yeah, totally. And actually I think the efficiency the efficiency of what's happening now or where we're realizing it is actually the flow of information around our organization is flowing faster. Everything is flowing faster, decisioning is make is happening quicker, uh stakeholders are getting the the insights that they need to drive our business faster than ever before. We're we're just like across all aspects of our business, we're we're flying. It's It's really good. Tyler, I mean you mentioned earlier, we're in the World Series in terms of our phase right now. I'm kidding. Uh top of the third inning. Top of the third inning. Um what's been the aha moment for OpenX? You're like, "Wow, this this could be transformative." It's um for us it's it's a change in behavior, right? In our industry we've we've gone through so many different things where um things have taken off and then they've kind of pulled back. Um we think a lot about I think the end result and we talk about the end result, but it doesn't actually change the behavior. So over the course of programmatic I think too, um things like quality come top of mind for me. We're we still have a ton of quality issues. Those could be solved through behavioral changes. How do we spend money? Who do we invest in? This is one of the first times I remember that it's a full behavioral change. So like when you walk into internally when we walk in, we're changing teams and we're shaping teams around this, right? We're hiring talent around this. Um externally we're having those conversations about how should they staff their teams, how should they be thinking about innovative products. So it's about the change in behavior I think for us was the aha moment, not just about the end result, but people's actual actions representing where we're going. Yeah, that I'm hearing that behavioral change is happening quickly in organizations. Jared, what about in publishers? I mean you've got the cast of Sesame Street there. Yep. Yeah, oh it's telling you Elmo, Rosita, they're they're all helping out. Um so no, it's it's it's true. Like I think a lot of stuff in innovation and it's a very fuzzy title usually, right? It means I can do anything, It's great. Nobody knows what I do, which is wonderful, including my boss. He's fine with it. [laughter] I kind of I think I kind of know what you do. Yeah, no, you you know a lot of the stuff that I've done for sure cuz we've been in the weeds on a lot of it. You see in a a lot of people see innovation as like cracking big hard problems, right? And I think one of the things that AI does is it makes the cost of asking a question very cheap, [snorts] right? Which means suddenly things that were expensive are cheap. You're like, oh, we can now take on this big hard problem, right? That that's 100% true. The thing that I find really interesting is when people ask dumb questions. Cuz when you can do like dumb silly things and play with it, that is when you get like in invention, creation, like kicking stuff around. Like so like the Sesame Street example is a trivial one, but as soon as I saw that I was like, oh, we are comfortable with this. Like we are having fun with this. We're going to create some incredible things. I think from an ads perspective, we held off actually on the agenty buying thing for quite a while because it was a little unclear what the actual KPI was that we're trying to Like what is the problem we're solving? Um on the like ad targeting, ad creation, high impact stuff, we've been doing that for a couple of years and we've got some like astonishing ROI numbers. We're now in the world of like guaranteeing incremental ROAS to our customers for agenty buying paths, right? For for and it it's doing incredibly well. Um on the agenty buying side, the question we had was is this an opportunity to go more direct to the supply and cut out a bunch of the ad tech tax? Sorry to all the people here who are whose strips were paid for by the ad tech tax, but like a $10 buy is $5 when it gets to the publisher, which often falls below our open market and gets 1% win rates. Like that's a bad path, right? Is it efficiency, right? Which every agency in the world is getting asked for at the moment because you headline efficiency and people want lower prices, right? That's tough. And efficiency is hard, as like we all know, right? Making that shift in the company is difficult. Is it actually driving better performance? Or is it AI jazz hands? Right? Which I would put it as like also valuable, and there's a lot of marketers, and and it has huge value, right? And the answer is we found out all of the above, but some of those have like very concrete KPIs, and I think it's coming to clarity maybe even in the last like month from different partners which one they want to really go after. And now, to your point, we have the tools to chase each of those specifically, even the jazz hands. Um yeah, jazz jazz hands. Jazz hands are important. I mean, real buzz around the industry is is absolutely important. Um so, this panel is comprised of independent companies. People like an independent media Absolutely. Uh Butler Till, independent agency. The Independent DSP. Uh uh OpenX, an independent SSP. Uh now, compare that to you've got the Metas, the Googles, the Amazons of the world, right? You know, that have the ability to operate their own uh LLMs to power their own proprietary agentic workflows. Let's start with Butler Till cuz I I I think you guys have been really pioneers here. Where do you see the advantage that an independent agency has versus non-independents or larger hold cos or uh or or or the Googles of the world in this AI race? Yeah, so I think pre-AI we used to pride ourselves still pride ourselves on just being absolutely agnostic in terms of the channel and the medium in which we're recommending to our clients because we, at the end of the day, want to do what's in their best interest and not through some other motivating factor, and the same is true today. So, we believe in an open architecture, we believe in creating the right tech stack that includes data partners, supplies partners, publishers certainly, that are going to solve their very unique needs. I mentioned earlier that we work in very complicated categories and so there's not often kind of a one-size-fits-all mentality when it comes to especially programmatic media. And so I think the advantage is we can use whoever we want. We can use whoever is the best partner and we can experiment with a lot of partners to find the absolute very best one for a particular use case and that's something that a lot of the holdco's can't do. It's a little bit of the hammer and the nail kind of mentality. If you have this thing then you always have to use this thing and we're not beholden to those same constraints. Right. So agentic workflows enable you to almost shrink wrap solutions on a on a bespoke client basis. going to use the same word. Yes, absolutely bespoke and and I think clients you mentioned before not a client up here not a brand up here because I think they're a little bit shy in some cases in AI especially our categories and so they really want to understand how it's all going to work and just saying that you have this big um solution. I mean they want to know but how's it going to work for me? What where exactly are you getting the data sources? How exactly are they being applied? Are they HIPAA compliant? You know, how are you getting around certain types of regulations in a very compliant way and you know, we're able to do that and actually demonstrate the use case and not just say hey we have this thing. Question same question for DeepIntent. Like what what's the advantage over big tech? Yeah. And the higher end. I think I think what one thing we pride ourselves on and and depends how you define independent, right? But I think the the benefit here is that we are not beholden to like a legacy media business. We're not a publisher, we don't have you know, we're not trying to like sell our media but also sell you technology. Like we are very clearly we know our swim lane and that's to be a software company, right? We are and increasingly it's turning into like this intelligence layer that our moat is the data. It's the health data assets that we've integrated directly into how we buy. We have the DSP which is today the activation platform. We're agnostic. I'm sure at some point we're going to be doing a lot more agentic buying as well and then A2A or whatever the the model is. But, the point being is that we have developed I think agentic is doing nothing but like accentuating or is that a word? Uh uh yeah, basically just making it clearer that our where we are so clearly differentiated in how fast we can move at it expressing that differentiation into solutions into insights that actually drive meaningful business and and results. Tyler, same question. Um I think that some of the disadvantage we've probably experienced as independent companies or maybe just against bigger ones are their solutions to problems were to just throw more bodies at it, right? FTE models, whatever it might be, walled gardens, things like that. Um the the difference now is this is to me feels like it's starting to level a playing field, right? It's about independent companies have done a phenomenal job I think hiring and securing talent, right? Like we talk about curiosity and driven when we hire. People who can be super adaptable, but they've been weighed down in the monotonous day-to-day work and the manual processes that we've had. So, when I think about agentic and generative AI and how it can help automate some of that, it's freeing the incredibly talented people we've hired up to go do the important stuff, which is the high-value work of relationship building, driving strategy, working with clients to deep understand their business. So, I think the advantage we have is finally freeing up people for what we hired them to do, which we haven't historically been able to do because they've been weighed down in everything else. Right. I'm hearing it's combination of agility, moving people out of rote tasks, and moving them more to creative or strategic. Like no one wakes up in the morning and says, "I can't wait to reconcile some campaigns today." Right? Like [laughter] Like I'm sure that person exists. There's a handful. They're not here. I can tell you that. They're they're likely not here. Um so, agility, repurposing of talent, putting in the highest and best use. Um Yeah, you want to weigh in on that one? Yeah, I think that we um we have long and complicated relationships with platforms. I think I'm legally allowed to say all of those words in that that sentence. Um the platform advantage has always been having tens of thousands of engineers, right? I now have tens of thousands of engineers in my pocket. Right? We can now build a search engine, which I did, and I sent it to my friend at Microsoft to poke him because it was better than theirs, and like he got competitive. I was like, "Dude, I built this on a weekend, and you're a trillion-dollar hyperscaler in search. This is not a fair fight, but I now I know what really annoys you. But we can work together. Let's Let's make it happen." Um absolutely the the comment about bespoke, like artisanal, farm-to-table at scale. The ability to take the things we do incredibly well and scale it is unique now to to what was possible before. A good example of this, like we run through massive publisher. We run health, food, finance, tech, auto, beauty, like the whole spectrum, right? Um and we have all the collaborative filtering data over years and years of people who read this also read that, right? It's a treasure trove of insights. One of the things we're able to do once we threw like AI on top of this to go mine for insights, the best predictor you're going to go on to an arthritis medication, outside of endemic health content, right? Cuz that one's easy, right? You don't need systems to go find that. Is in Brides magazine. Right? And it's one article, and it's how do I resize my wedding ring? Now, that is a person who has a problem, is dealing with it in their life, not in their medical treatment, which is so much of people's lived experience, and they're finally taking action. So, they're at the point of they're dealing with the problem, which is exactly when they are primed to also make the healthcare decision. We have 1.2 million documents. We have about a trillion overlaps between those 1.2 million documents, and we can mine that for connections. It was really inefficient to do that with human researchers. It is way more efficient to do that with AI. So, that is a wealth of stories and storytelling and narratives that help us meet our clients where they need to be and bring them to their consumers at the point where they're actually making a choice. Good point. It was wild. I was like, "Oh, yeah, that'll do." Tyler, you were talking about how [clears throat] AI has transformed repurposing of talent within OpenX. How is it transforming how you're building OpenX? How you're How you're hiring? What skill sets matter today? What skill sets are going to matter 5 years from now? Yeah, it's a it's an interesting one. Um so, I think we mentioned 10,000 engineers in in your pocket, I believe. Um look, on the engineering product side, the skill set I think is definitely changing, right? People with deep experience here, people who come from that background. People who have experience coding Agentech, coding generative, that is extremely extremely valuable, right? Our teams are doing a great job, I think they're learning and adapting really quickly. Um so, yes, the power of many more in your pocket is definitely evident. Um the business side is interesting, though. I think that that's a fascinating one because at the end of the day when we've had our own internal panels on this and things like that with customers, a lot of them tell us, "Look, the the output of the Agentech is important. What it can do for my business is important, but that's not all I'm buying. I'm still buying people, and I'm still buying relationships, and I'm still buying strategy, and I'm buying trusted advising is what I need more than ever." So, on the business side, it's a little bit different. It's about business development people, solutions focused selling is going to continue to be a key there. but we still need them to do the thing they've always done, which is that's the part that's the thing about our industry is it always comes back to relationships at the end of the day. So, the profile there doesn't necessarily change, I don't think. It adapts, for sure. Um but that's going to be how I think we think about it over the next few years, at least. Same at Deep Focus. Has it changed how you're building the business, how you're hiring, what skills matter? Totally. So, uh we and this is not a perfect a perfect kind of rubric, but we kind of bucket everything into three levels of work. There's like routine work, knowledge work, and then judgment work, and like relationship work. I like that. And what's happening is the the the horizon of what is possible with AI has only moved up and up and up that kind of that rubric, if you will. Um so, we we think that rubric uh sorry, the um routine work is going to increasingly be done by AI. And the knowledge work, like research, is going to increasingly be augmented by AI. And the category that will be left really untouched, but like highly informed, is going to be the the relational work, like the judgment, like the the the parts of the business that can't be codified into a model. It'll be the way like that we relate to the world around us, where judgment, intuition, context, morality, ethics, like those are things that must be human by by nature. Abs- absolutely, yeah. So, so we are we are very much, I think, in this idea that we are going to be hiring way more people who are interfacing with uh with our market place, more so than kind of hiring for people, staffing for like internal operations. Yeah. At this end. Going to Butler, Tell, I I I when I fir- you know, when I first started exploring the agency as an industry, I was like, "My god, this is going to transform agencies." Like I I feel like it's it's it it probably impacts, I think, agencies and your brand partners the most. How has that changed how you build Butler, Tell? How has that changed how you hire and organize the agency? Well, there are two key stakeholders in our business, our employee owners and our clients. And so anytime we're embracing new technology, we have to make sure that it makes sense for those two key constituencies. And so for our clients, you know, we're adopting things that are in their best interest, that create greater efficiency, higher quality, faster speed. And for our employee owners, it often comes down to the experience. And so we embraced a lot of the AI tools early on because we wanted to make the experience of our employee owners better. And so we put it in their hands and we said we want you to start playing around with this with the right guardrails, of course. And it's really interesting how quickly they all were to adopt. And so in terms of hiring, I think the two key traits for most companies are going to be agility and curiosity. So The reality is the roles that you're hiring today probably won't exist, not at least in the same shape or form, even a couple of years from now. So you need to be hiring people that are comfortable learning new skills. I mean, I don't even know what some of those skills are likely to be I like anyone that 24 months from now. And so I think it's more about like the mindset of the people that you hire and not necessarily the experiences they've had, because we're all in uncharted territory at the moment. Yeah, that that interview question is like, how comfortable are you with the rapid state of change? How do you feel about constant change? Yeah, so I often ask people that I'm interviewing, especially young people, you know, how they have they dealt with adversity, when have they learned something new. What do they do in their spare time to like make sure that there's some evidence that they want to continue to learn, because it's people that are stagnant that are ultimately going to be left behind. And what's your assessment of the current crop of candidates entering the industry that you're interviewing? Are they Are they up for it? Are they up for the challenge? Generally, like are they We've been really happy. I mean, we've hired over 100 people year-to-date at Butler, Till. And so it's yeah, yeah, we've been keeping our recruiters really, really busy. And you know, I think that we've been pleasantly surprised. And it's a mix of all levels, but for entry-level people, um yes, I think it's really interesting because they are true AI natives, which I think gives them an advantage. Yes, cuz sometimes it's harder to kind of retrain, you know, the middle layer of an organization on new tools than to hire people who just have known it since the beginning. Yeah. You're living proof of the counter narrative that AI doesn't eliminate jobs. It's actually creating new jobs, just different types of jobs. Totally different jobs, different experiences, but you know, ultimately as an employee-owned company, we have an incentive to continue to grow without necessarily having to hire at the same rate of growth, if you will. And so I think it creates a really great advantage for us and all of our employee owners to want to adopt AI tools because ultimately, you know, we're all sharing in the financial success of the company. We need to socialize that story. Yes. That's a great story. What about our publishers? How's it changing how you're how you're growing people, you're building people? Ink. Yeah, so people as well. People is obviously a very large part of the business. We got 41 other brands we all love equally, right? So that's how we got over a million documents, and you might say like, okay, well, that's lovely cuz I can now just write over a million documents in an afternoon with a couple of prompts, right? Um not in fact how good content gets written. Um we committed early that we will never publish content written by a machine, right? Which may seem counterintuitive given our scale. Um but AI is, as we all know now, mathematically average content, right? If you want a commodity fact, the average answer is the right answer, mostly. I mean, Reddit and 4chan are in the training data, so not always, right? Be careful what you go with, right? But we know that the thing that that we need more of going forward because inputs now power these systems, right? Your answer is only as good as what you put in. So, we need more and better information going forward than we've ever had in the past. We have the biggest content production process and business in the US by far. So, this is a very, very exciting world to get into when there's a greater demand for more and more interesting questions than ever before. And because we're a large language business, right? We move a lot of words around every day. These models are incredibly helpful in the back office helping with a lot of the grunt work. It's very AI assisted, very little AI replaced. Um and it's allowed us to like really lean and build new and better experiences in ways we have we haven't done before. Um it does freak out our engineering team that non-engineers are building prototypes. They're getting over it. They're learning. Like we're figuring out how that collaboration works. And as the CIO, I have also been committing code to Bitbucket and that scares the hell out of them. So, CIOS are hurting CIOS careers. fun. Uh but no, it's good. Uh [clears throat] last question, we'll close with this so people can go back with some action items and there were a few brands and agencies in the room. For those that For those that are curious about uh kicking off some agentic test or trying AI in their organization, Jared, we'll start with you and just go down the line. Um next Monday when everyone's back in the office, um what's what's a pragmatic step they should take in terms of starting that journey? I think the [clears throat] people are all stages, right? If this is something you're dipping your toe into, get a spreadsheet with data that you would hack through on a daily basis. Everybody here has one. Throw it into Claude, ask questions, figure out what works, and find where it fails. Much more useful to break it, right? Cuz then you're going to start to get confident in how to deal with it. If you're further down the learning curve, like [snorts] get data sets from other people or get a standard that you haven't worked with. I hear the IAB has a few of those, and just see how quick it is to build against the standard. Because you don't need to understand at all. You're just like, build me build me an agent that can go work with the content monetization protocol, or the Ad Gentek buying protocol, or any of these protocols. It's less than 30 minutes. See how like you will have something up and running in no time. It will then lie to you. So then figure out how to make it not lie. Right. Yeah, for me it's experimentation. You got to put the tools in the people's hands, and for us we started with the leadership team. And so we were the first ones to adopt a lot of the the tools, and we used to have um I would say like almost contest. We'd bring together and say, "Hey, what have you done lately?" And it could be personal, it could be professional. But we wanted to create this environment of experimentation, and kind of lead by example. You start you start a culture change. Absolutely. Absolutely. Yeah, we we very early on committed to opening up access to everybody, to Anthropic, and ChatGPT, and just getting them like just using the tools. So that experimentation phase, and then we I think myself included, we were modeling the behavior of like experimentation, putting stuff out there, getting reactions, kind of seeing and kind of discovering ways to actually integrate it and implement it. And actually the benefit, the upshot is that we're developing AI-driven products. We also have a we're also retooling how we operate the company today with AI. And actually the the similarities between the two different like the internal and the external-facing stuff, the knowledge transfer and in like just the ability for us to learn and kind of share notes about how we're changing our business around AI, and how our clients are we think our clients are changing their businesses as well, it's actually allowing us to build better products for our customers. So I think I think just experimentation, and then like modeling it and finding those like champions of it, super important. Ty, take us home. Uh experimentation for sure, but if I had to say something different, I would say um challenge your partners and ask questions. Um, this is coming from an ad tech person speaking to agencies and brands or publishers, but um every company is an AI company now. They're all pitching you something, right? Meaning, how do you cut through that that BS state to get to what's underneath it? Is super important, right? Because we always say, "Look, more agents are not going to save your business, but better inputs into that will." Right? So, it's about the foundation and what's going into those. It's not about people out doing the latest demo that isn't built on a foundation of of transparency, trust, and solid tech. What I hear there is data cleanliness, data fidelity. Like, now's the time to Now's the time to clean that up. Garbage in, garbage out. Exactly. I mean, I I said that on several panels today. GIGO, garbage in, garbage out. We We actually hired a data ontologist to like for that exact problem. So, that's like a new role that like And so, that person's job is just to clean up, build taxonomies, and actually build the connections between data sets. And like, 1 year from now, mark my word, it's going to be the hottest job. It's like the most important job. AI will make really good bad decisions on bad data with confidence. With confidence. It'll accelerate it. And and humans have never done that. [laughter] True. It'll just do it faster than us, right? So, I want to thank my panelists. Thank you so much for joining us. Thank you for hosting. Have a great camp.