Programmatic Gets An Agentic Brain

0:12Scott Ensign: Welcome to Ad Exchanger Talks, the podcast devoted to examining the issues and trends in advertising and marketing technology that matter most to you. This episode is sponsored by Verve, an ad solution that helps brands activate consumer intent in real time across platforms. As AI reshapes how people plan vacations, Verve helps travel brands show up when it counts.

0:47I'm Alison Schiff, Editor-in-Chief of AdExchanger, and my guest this week is Scott Ensign, the Chief Strategy Officer at Butler Till, an independent performance agency that's fully employee-owned and women-owned and women-led, so hell yeah. Butler Till has been right there on the forefront of testing agentic AI in the programmatic supply chain, and we'll get into the details of what happens when you enlist AI agents to plan, activate, and optimize campaigns end-to-end. We'll also talk about the ad context protocol, why it's finally and for the love of God or money, time to move away from a cheap reach mindset, and why human media planners aren't going away, but their jobs are about to change a lot.

1:30But before we get started, if you haven't snagged your ticket to Programmatic I.O. yet, you should really get on that. It's taking place September 28th and 29th at the New York Marriott Marquis, and you're going to want to be in that room because we've got a seriously good agenda lined up for you with speakers, including Chris Kane, Martin Sorrell, Colleen Coulter, the global ad solutions lead over at OpenAI, Alexa Raff, the CMO of Gwyneth Paltrow's Goop, and Bob Lohr, the president of Horizon Media. We've also got folks from Yum Brands, Dentsu X, the Brand Tech Group, Razorfish, Zoom, Time, PMG, and The Guardian.

2:06I don't actually have time to list everyone. You'll just have to visit the event website to see for yourself. Podcast listeners get 10 % off with the promo code POD10. So what you waiting for? Get your ticket and see you there. Hey, Scott. Welcome to the podcast. Hi, Alison. Thanks for having me. So you know the drill. What is one thing about you that not a lot of other people already know and that I couldn't find on your LinkedIn or by Googling you or some other, I don't know, duck-duck-going you? Well, I think this qualifies, although I do talk about it with people a lot, but it's definitely not on my LinkedIn.

2:45During COVID, I have three boys who are older now, and we built an arcade in the basement. I have pictures. It's pretty cool. Started with a couple of things, an air hockey table and a vintage Double Dragon arcade cabinet, but it's kind of blossomed into a little skeeball uh pop a shot kind of a full complement of arcade fun down there so that's something that uh it's a good time it's always fun to just pop down there especially when i'm working from home and i need a a moment good good dad you know i try my favorite arcade game of all time is skeeball I love skeeball, but also knock hockey. It's really a time between the two.

3:35Yeah. Knock hockey is, and I'm quite good at it, I have to say. Nice. Yeah. Got basically all the games down there, one form or another. You know, it's a home arcade, so it's a little bit scaled down, but it's still a lot of fun. I'm assuming you collect quarters from your cauldron. the the funny thing is the arcade cabinet that i have the vintage one which is from 1987 saturn somebody's arcade or pizza shop or whatever for 30 35 years does still accept quarters and you can set it up to to take quarters so i've thought about bringing it into work and setting it up to have a little side hustle so that's going to be thank you my segue from home life into professional life.

4:21So I want to properly introduce you. You're the chief strategy officer at Butler Till, Rochester, New York-based independent performance agency. It's 100 % employee-owned, women-owned and led. And it's a certified B Corps, which is a for-profit business that's verified to meet high standards of social and environmental performance, public transparency, accountability, A lot of things programmatic is not, funnily enough. I mean, other examples of B Corps, so people have a frame of reference, like Patagonia, Tom's, Ben and Jerry's. But you've been over at Butler Till for a while, since 2018, lots of different hats, VP of digital media, then a lot of VP titles, VP of client solutions, VP of strategy and partnerships.

5:07Keep it interesting, you know. Well, in 02, you became a chief, so chief strategy officer. But going back to the B Corp status thing, what does that certification mean for you? It does hold you to a pretty high bar and it's not always easy to square something like that with the realities of a business, a media business, where some of the biggest vendors aren't always known as paragons of ethical virtue. I could put it that way. So how do you compete in a business where, and I don't really think I'm going out on a limb, where cutting corners is like often the easiest way to cut costs. And I also realize it's a little bit of a flawed question, because I don't want to imply that ethics and business performance are mutually exclusive things.

5:54They are not. But I do feel like there's some tension there. Yeah. And it's a great question. And we believe that those two things are not at odds, which is really fundamental to who we are as a company. And you make a really good point about the agency world and the media world, which is largely, has been historically largely centered on things like cost efficiency, which can be at odds with running a business as a force for good, which is what we believe in. And really beyond the B Corp certification, which we've had for a while. Those are, as an employee-owned company, as you mentioned, a women-owned company, women-led company, you know, these are the values that our people hold and care about.

6:36And we are made up of employee owners. So we're constantly looking for ways to better live those values in the marketplace. And honestly, you know, agency life, agency business is a tough business, but we don't really find it challenging to live those values because it's honestly so straightforward for us to do business that way. It's how our people want to do business. We have to be transparent with our people because they are the owners of the business, but that really flows into the work that we do for clients in the world. So it really makes it a unique culture and differentiated place to work and to do business.

7:20Well, I want to get into the weeds with you, but I want to start like really upstream. And it ties into questions of transparency and accountability and ethical behavior. So you wrote an op-ed for The Current. It was sometime last spring with the title From Haggling to Harmony. By the way, kudos on the alliteration. I love alliteration. But from haggling to harmony, we need a new framework to find the collective win on the open web. And it was kind of a rallying cry. And the issue is, as you were arguing in your piece, it's that the buy side is totally addicted to cheap reach and the sell side hoards signals and like no one has enough incentive to fix the problem.

8:06because if you stick your head above the parapet, like you might lose money. Like why be first? Like someone else should be first and then other people can follow. And then a little while later, earlier this year, you wrote this follow-up piece responding to the analytics report about ad placements next to child sexual abuse material. And you pushed back like rightly on the industry's instinct to just blame the sell side and buyers do need to own more of the responsibility. And I mean, ads next to CSAM is just like, come on, guys. And there's this very good line in there that you wrote. I've often compared the problems with program inequality to the war on drugs in the 80s.

8:49There is only so much that you can do in the form of supply restrictions to solve the problem. You have to interrogate the root causes of the demand of the demand at the same time. And it's a very good analogy. But I mean, is that starting to happen or what? Like, what other revelations do you need? First of all, the level of homework here is staggering. So who knows to you on the show prep. You've got better receipts than I do on what I've said over the years. So good for you. Yeah, I think it is starting to happen. And it's interesting. I'm involved in some industry groups around AI, and we're talking about the ways that that's changing the industry.

9:31I think people are understandably and rightly concerned about what impact AI is going to have on the space. But I think there are also a lot of positive things that it has the potential to do. And I think making buys more sophisticated and helping us to move beyond this framework of per unit cost as really the only measure is one of the things that that technology can and should help us to do. I think the other thing that's really changing, you know, the line, this is a conversation I literally this afternoon was having in one of these industry groups. The lines are really blurring between like what media is, what paid media is, and that's only going to continue.

10:20We see that with things like influencer and creator. And now with the, I think, ongoing efforts by brands and agencies to influence what LRMs are picking up. I think all of that is just going to change the framework, the currency for how we think about media and how we think about marketing and reaching audiences. I feel like in 10 years when we're having this conversation, we would look back at my quaint little piece about per unit cost and look at things very differently because we are finally seeing that change and it's changing pretty dramatically. In 10 years, I won't be here. You'll be having a conversation with my digital twin.

11:04I'll be reading a book in a coffee shop somewhere. Let's hope. So I'm going to mix my metaphors a little bit. you were talking about blurring lines. So the lines are blurring at a crossroads in a way. And you brought up AI. I want to spend a lot of time talking about agentic AI, which is, I was going to say it's going to transform the programmatic media landscape, but it's already transforming how programmatic media is bought and sold is happening right now. So in December, you guys, Butler Till, together with Pubmatic, you launched what is, I guess, one of the first fully autonomous end-to-end, all the buzzwords, agentic advertising campaign running across CTV for a malt beverage brand made by Goloso Beverage Group.

11:54And I'll just briefly - Jelloso. I thought it was - Jelloso. Well, my bad. I'm so sorry, Jelloso Beverage Group. So there was no traditional DSP workflow involved. There was no manual optimization. The campaign was activated using a Gentic OS, which is Pubmatics' Gentic platform for buying. And then it was all done within Butler Till's preferred generative AI environment. You guys use Claude. So basically, you submitted a natural language brief. Pubmatics agents interpreted it, generated the media strategy, initiated the campaign setup. continuously optimize the pacing, the targeting, and the performance.

12:35And so you do have to forgive me for being a little cheeky with this question, but like after you submitted the brief, what, did you just go to lunch? Like, what did you do? Well, in this case in particular, I mean, we had eyes on this all the time because, you know, as you mentioned, this may have been the first time anyone's ever done it this way. So it was a pretty cool moment, honestly. I remember standing in the office of Ryan who leads activation on my team and, you know, using Claude and prompting, like you said, agent to go ask and, and watching it do this work that was like only theoretical up until that point.

13:15Um, so that was a lot of fun. I, I do think that there's, I don't necessarily like the phrase human in the loop, but I feel like what this is going to do is just empower people with the ability to do things that were previously impossible. And Pubmatic's been a really good partner for us in all of that, because we've sat down and just said, instead of talking about how we can replace what's already happening in Programmatic, that just seems like such an an unimaginative way to look at this technology. Let's think about all of the things that are really bad in our space right now and all of the things that we would love to change if we had our druthers.

13:59And we really started from that point of view, like what's, if not impossible, what's at least, you know, infeasible at this point that we could solve using technology. And so I think that's going to be the future is people just like when you started to be able to edit our video and audio with digital technology you know technically a lot of the stuff you were doing was was potentially possible but that just gave you so much more capacity and the ability to solve so many more problems how good was it though like how much editing was required when you saw what was generated? Did you look at it and go, oh, yeah, press play, like we're good?

14:46Or you had to tweak things and there was a good framework for you to play with? Yeah, I mean, there was definitely some tweaking. I mean, the one thing I will say, and people ask me what was surprising about this experience, I think what was surprising was how good it was right from the start and how it came up with things that we wouldn't necessarily have thought of in terms of like what inventory, audiences or packages we might want and how far along it seemed right from the start. That's what really opened our eyes to say, okay, this is, we knew it was going to be an important advance in our space, but we really, I think, view that now as something that's going to be at the heart of how we do pretty much at some point all of the work.

15:34It's going to be right there along for the ride. That's an interesting observation. I was just talking to Julian Ranj. He's a measurement expert at Northwestern and he's going to be speaking at our programmatic IO event in September. And I was asking him about open source MMM and using OSMMM, that's the acronym I know, God help us, together with Agentic AI and how that's allowing people that don't have a data science background to basically do MMM, which is wild. and we veered off into like can you actually trust the output of of an LLM and he said something like I'd push back I'd push back on the framing of hallucinations as purely bad because when you have a brand new idea in a way you're seeing something that the current consensus on like knowledge hasn't really validated yet.

16:37So a hallucination may just be a brand new thought, actually, and something that doesn't necessarily exist yet in the world. It's perhaps a very positive spin on something that could also lead you astray. So you do have to stay firmly in the loop, although I guess we've decided, and I totally agree, that human in the loop is such a cliche phrase and I don't love it. Yeah. I mean, I think sometimes people might use that to just make themselves feel better that there's always going to be a human involved. Yeah. Thank you. Great. Yeah. It's interesting. And I feel like this kind of happened when digital media really became the primary delivery for our industry too.

17:23There's always with a new technology like this, it feels like it's always held to a different or higher standard at the beginning. You know, we trusted, you know, just one of my first jobs was in radio and we trusted a rating system that was based on people writing their radio listening habits in a little diary. And we took that as truth. And then when digital came along, we said, oh, well, this is, you know, look, there's some inaccuracies here. I'm like, okay, some inaccuracies versus maybe a majority of inaccuracies. And I think the same thing, we're kind of going through that same cycle with AI and agentic.

18:06And I think it's fine to point out where there are gaps, where there are things that we can solve. But to your point, with something like an MMM, that's a great example. We're trying to draw an inference from data. And that inference is not going to ever be 100 % accurate and correct. But if a technology like that can put that in the hands, as you said, of more people who may or may not be data scientists and a larger group of people and get them to answer us faster, I think you can see where that's going to just accelerate our ability to better optimize and measure media. Totally. So one more before we take a quick break and going back to the work that you were doing with Pubmatic, I think the results are worth calling out because they were like really good and hard to argue with.

19:01You got supply chain and tech costs down by like 80 % or higher, even a bit higher, 40 % more impressions on the same budget, 98 % video completion rates, less than 1 % waste. But like it is one thing for a test to perform well and then another thing for this to become how the industry runs programmatic media at scale. So what will this transition actually look like? Like, what do you think the biggest friction points will be? Is it going to be like legal, contractual, just clients being comfortable doing it, inertia, the fact that the programmatic industry's existing infrastructure like wasn't built for this, for agent to agent communication in mind, which I guess is why you kind of want to blow it up.

19:50It's a cocktail of all those things. Yeah. I mean, I think all of those things are part of it. I mean, we definitely see in our space the way that buying behavior impacts the adoption of technology over time. We see that with things like TV and CTV. And I think that that's going to certainly impact this as well. You know, when I think about programmatic, and I don't want to get too maybe pedantic about that, but what is and is not considered technically programmatic? I think one of the things that if we wanted to take a very purist view of what a programmatic campaign is, you would say, well, that's real time bidding.

20:33It's going to have to happen at the impression level. And that's happening in milliseconds today in the ecosystem. and an OLM, an AI agent is not going to be able to make impression level decisions at scale like that. So it would never work for programmatic. So I think some of those things are a technological challenge potentially, but everything that's happening around that real-time bidding environment can happen in an agentic framework. And I think that's what we're going to see, but it's going to take time. I mean, I think the other thing that holds a lot of promise that we're looking at, if you think about, you know, sell-side ad servers, so, you know, GAM for the most part, they're opening up inventory to use buyer agents to talk to seller agents through the ad server.

21:29I think that's been well reported and documented. And if you can do that in a way that shortens time from like RFPs that you might be sending out to those publishers to happen instead of in days to happen in potentially seconds with those agents talking to one another. You can get a campaign launch faster, you can get results faster, you can start to optimize faster. And while that might not be, it's not going to be a real-time bidding environment, like at least it feels a lot like small p programmatic. And it's an automation of the media buying that we've been doing, which is what programmatic was meant to be, at least if I think about first principles from the early days of the technology.

22:17So I think all of those things are just going to come one thing after another. It's going to take some time and some cycles. People are going to have to learn it. Publishers are going to have to adopt it. Platforms are going to have to adopt it. Buyers are going to have to get comfortable with it and adopt it. I think those are the things that are just going to take time like any other technology that we've seen. And like so many other technologies, we tend to overestimate the short-term impact and underestimate the long-term impact. I think we're seeing that right now with some of our early buzz on Agentic and then people, you know, it kind of goes in those peaks and valleys of excitement around it.

22:58But there's no doubt in my mind that it's going to be a part of really every way of buying in a few years. And I really do like and described to the expanded and expansive definition of programmatic. I think it's more of an umbrella and less of a circumscribed term, and it's more useful in that way. It's kind of a mentality, really. Yes, we like to call it a way of life. I guess some people would tell us that we need to get a life. Yes, that's fine. We all make our choices, right? So I'm choosing to take a quick break. And when we're back, we'll talk about AdCP and smart curation. So stick with us, nerds.

23:54Hello, listeners. I'm Sarah Sluice, Editorial Director of AdExchanger, and I have with me today Anna Slater, VP of Global Brand Partnerships at Verve. Welcome, Anna. Thanks so much, Sarah. It's great to be here. So we know that AI is changing how people plan trips. I've heard the anecdotes from my friends. I've done it myself. So can you give some insight into traveler behavior right now? Yeah, absolutely. I really don't want to date myself, but I've been in the travel industry now for almost 20 years. And having worked so long in one particular industry, it's been amazing to see the overall progression.

24:32And I really feel now that the adoption of AI and this trip planning process is the most impactful transformation we've seen over that time. And looking at our own LLM data here at Verve, consumers are using it for general discovery. Around 28 % of the prompts that we're seeing relate to these types of inquiries. And this is growing. And AI use for travel inspiration has doubled in the past year, especially amongst the 25 to 34 year old audience, which has grown to around 18 % adoption. And we're also noticing there's a shift away from keyword searches like cheap flights to Spain and treating AI now more like a travel agent, which is adapting in real time.

25:12However, there's still an element of nervousness around utilizing AI to book, you know, since the prompts that we're seeing here are certainly dominated by that discovery, the timing, the itinerary building. But I'm sure if we're going to have this conversation this time next year, that will have all changed. So if I'm a brand and I know that AI is owning discovery, people are using it as a travel agent, but not for the booking part of the equation, where does that leave brands who are trying to reach travelers? Good question. Travellers are using AI to navigate every stage of their holiday planning.

25:45But the biggest challenge with this new way of researching is having on brands is the pace at which it takes consumers from dreaming to consideration. Prompts get incredibly specific and in a very fast way. Someone can go from give me trip ideas straight to a shortlist without ever touching a brand's website. You know, and we're also seeing really heavy vendor comparison, you know, with travelers using AI to sort of audit those travel providers and then benchmark prices. So by the time the brand actually shows up in the conversation, the shortlist may already be set. So it is a very difficult moment for travel brands at this time.

26:23And then furthermore, that intent is scattered across more and more devices and platforms. But these conversations, they're carrying some of the richest signals that we have ever seen on intent. And AI is forcing the conversation to accelerate the decision making. And if they can tap in to use these signals to target, they're going to get better performance overall and have a chance at winning that conversion of the traveler's itinerary. So you just described a very complicated, complex, scattered and rapidly changing set of consumer behavior. So what steps do you recommend to brands who are trying to capture this consumer intent at this at this moment in time?

27:04Yeah, absolutely. And this is conversations that we're having with our clients daily at the moment. And the answer really is twofold. Firstly, travel brands need to see the intent and then also follow up and ensure that they show up in it. But there are a lot of brands that are still missing out on both right now. So on the seeing side, you know, that's LLM data, the conversational signals that we're pulling from AI chat tools. We layer them on top of actually search and zero party signals. And when you combine this, it provides a bigger, fuller picture of consumer intent. And importantly, you know, the LLM data shouldn't be seen as a replacement.

27:40it's a real complimentary intent signal. So obviously we're encouraging brands to sort of see what's out there in market. And then on the showing up side, it means being present inside the AI conversation itself and not just hoping that the LLM happens to mention you in the brand. Well, thank you, Anna, for those concrete steps and recommendations. And thank you to Verve for supporting AdExchange Your Podcast. Thanks so much, Sarah.

28:15All right, we're back. And as promised, I want to talk about AdCP because we didn't really talk about it before the break. You just kind of mentioned it without naming it in passing. So it's the Ad Context Protocol. It's an open protocol that came out of agenticadvertising.org. It's a new independent nonprofit trade group that wants to establish open standards, governance frameworks and education or AI driven advertising. I'm like reading right off the tin. Guys, Butler Till, your founding member. And so is PubMatic and Scope3, Brian O 'Kelly's thing. He's very involved and a whole slew of other companies that would be way, way too many to mention.

28:56And the idea with AdCP is that it's not just another way to pass bid requests. It supports It's something more like a multi-step conversation between agents. So a brand's AI agent can say whatever, like find me women likely to be interested in cycling in California. I don't know. And a publisher's AI agent can respond with matching inventory and then send the transaction directly to an ad server and then essentially bypass the whole traditional programmatic bid stream entirely. I don't know if I'm being reductive, but that's how I understand it. And that sounds like it would be really disruptive, obviously, to a lot of existing infrastructure.

29:38But is that actually the point? The point is to kind of not blow it up, but say, well, we can start afresh because we have something brand new here. yeah i mean and i'm not it's a it's an interesting question i mean i think the idea is like this is a this is a new technology that is that has the potential to change things dramatically and i don't think that's going to happen as we've seen it's not going to happen overnight i mean had cpu as announced late last year and here we are and i would say it's very much still in the not even in the early, early innings. And I'm not, I'm not interested in necessarily kind of blowing up the way things work.

30:19I'm interested in enhancing what we're able to do in the space. And so what you're talking about is really part of what I see as a longer journey of automation. And I think this is a really important waypoint on that journey, but there's so much happening and the media space that despite all of the programmatic and technology that we have, that is still happening pretty manually. And if you can take things like pulling together planning data and pulling together briefs in an automated way and shortening planning and campaign cycles and getting to results and information faster, I think that's going to make everything better.

31:02And I think there's so much time and opportunity for everyone in the space, including brands, agencies, platforms, both the buy and sell side and publishers to start to experiment and figure out how this technology is going to work for them. I don't think there's any kind of fatalistic point where certain groups or a segment of the value chain here is going to be left behind. I think that's why we decided very early that we wanted to be a part of this, that we think it's something that's really important because we owe it to our people to say like, hey, we want to be part of writing this future on our behalf.

31:45And some people have said like, well, why are you participating in this thing that could be a potential existential crisis for the agency world? And we just really don't see it that way. we see it as a tool and an opportunity that we've got to be a part of. I mean, it's a very measured response. I will say not everyone is a huge fan of ADCP. Like Tony Katzer, the CEO of the IB Tech Lab, he's been out there with a strong opinion. He called it deeply flawed, resource consuming. His argument is that the industry doesn't need another trade group or protocol. of course you know the ibeat tech club's leader would say such a thing but you know just creating fragmentation when you can just evolve the existing infrastructure so more of a like evolutionary approach than a revolutionary approach but i mean how would you respond to that and how do we make sure this doesn't turn into sort of the latest edition of the ad industry's favorite or most hated tv show which is like the standards wars and like fighting standards wars that consume resources and then no one really wins in the end.

32:56And it just ends up with people pitching op-eds to AdExchanger. And like, that's as far as it goes, right? It's just like thought leadership, dot, dot, dot, nothing changes. Yeah. Or pitching panels. Yeah. Just like pontificating from a stage. Exactly. Well, some of us make a living pontificating. Me too. You know, I think standards are very important. So we're involved. You know, we know Tony very well and we're involved in the IAB. We're involved in these other standards bodies as well. I think it is going to sort itself out. It's going to take a little bit of time. I don't think it necessarily has to be long-term one or the other.

33:40And we don't really, you had mentioned we're founding members of ADCP. We're involved. We're going to, we're going to intend to continue to be involved with that. We don't know where everything is going. We just know that it's important that we are embedding this technology into what we're doing and that we have a really serious and meaningful approach to experimentation at this phase of a technology like this. So I know that I fully expect and I know the IAB has released their own stuff. They have the ARTF and they're working on standards at the same time, which I think is great. I don't think it necessarily has to be an either or.

34:23And that's another thing that this technology in particular allows for is creating things faster, getting to answers faster. And it might be that there's a future where there are multiple ways to buy the same inventory, to do the same kinds of things in the space. And I think that may well be okay. We just really don't know yet. I mean, people love choices. So if they're both functional choices, then you get to choose your flavor, I guess. Yeah. And maybe they'll come together at some point. Maybe those things will converge. We don't have a, as I said, we don't really have a, you know, our, our bets in one place.

35:05And I, and I'm not sure ultimately that it's going to be really impactful, like which of these standards runs or which, which body kind of comes out on top. I think there's a role for IAB to play. I think there will be a continued role for agenticadvertising.org and NADCP to play. And we're happy as an agency to have the opportunity to work with both. Okay, so it's not the sharks and the jets. Not from my perspective. Others may see it differently. So I want to take a step back for a second, talk more about humans. So brands and agencies have spent years, I mean, people's whole careers building up programmatic competencies like relationships with DSPs and internal teams that are fluent in things like bid management and SPO and a billion other things.

36:00But like we were saying, I mean, agentic systems, we're not talking about blowing things up necessarily, but they do potentially route around a lot of that infrastructure. So what happens to all of that institutional knowledge? Like my feeling is that it actually matters more now because, you know, someone has to set the strategy and the guardrails that the agent is executing against. Otherwise, things could go off the rails like pretty quickly. But no matter what, like the job description of a media planner and buyer is changing just so dramatically. Yeah. And maybe the title is going to change.

36:35Just as we've seen, I've told this story in a few other contexts, so apologies if you've heard it before. But I remember early in my agency days, getting in my not so nice car at the time as a young person and driving a printed copy of an RFP response from Rochester to Syracuse. So I'm quite happy that technology exists to handle some of those things at this point. And I think we're just going to continue to see tools being put in the hands of people. And that's what Agenic is. It's a tool that in the hands of the right professionals is going to be able to do some really, really cool stuff. It's already happening.

37:20It's only going to happen more and more. And when I think about the relationship side of the business, I couldn't agree with you more. That is only going to be more important. And if you look at work we've done with Pubmatic as an example, like our relationship with them as a result of doing that work together between an agency and a supply side platform, that's only deepened our relationship, our collaboration, our desire to do things together. So I think all of that, it's going to change, the nature of it's going to change. But I think the biggest change that you'll see is the amount of stuff that the same amount of people will be able to do together as a result.

38:04And people also, they need to roll their sleeves up so they understand what's happening. And to do that, you need an organization that is interested in messing around and experimentation. So I read about the Butler Till Innovation Fund, which I think is kind of cool. So it's specifically set aside for testing emerging technologies early on behalf of clients. And I'd imagine that things like AdCP and Agentic ad buying, those qualify. So how does the fund work? How do you decide what to finance the early testing of? And when do you cut bait when it's not worth it? Like, for example, maybe you guys tested the metaverse.

38:50I don't know. We did not for the record, but that's a good example still. So the concept in the ad agency world, especially in the media agency world in particular, has existed for a while of added value for clients. And, you know, RFPs will ask, how much added value do you have from this provider or that one, or do you generally give to clients? And there's a spot for that. And of course, you know, we negotiate added value for our clients. But the way that we started thinking about the innovation fund is, how do we move from added value to created value? And if you think about planning cycles of any kind, but certainly in the paid media space, there's a natural resistance to innovation.

39:44And by that, I'm not saying Butler, Taylor, our people, our clients, or our partners are naturally resistant to innovation. But when you build a plan, you're generally building that on what's happened in the past. And that is where you get your tried and true tactics. Like we We know where our search works for this. We know that CTV works for this. The research from different third-party syndicated partners is telling us that this works for this. That's all great, part of what we do, core to what we do. But of course, if something is brand new, it's not going to show up on the radar of any of those things, either in our primary experience or secondary research or anything.

40:25So we wanted to find ways to break into that. And so instead of just negotiating with partners for lower rates or in the programmatic sense with SSPs in particular, like your post-auction discounts, we wanted to do it in a way that felt really tangible and real to our clients and allowed us to innovate alongside our partners and alongside them. And so what we've asked partners like SSP, but also platforms and data partners to do is to contribute unrestricted funds to these client specific innovation funds. And so based on a client spend, they will be apportioned a percentage of those dollars that we then accumulate on a quarterly basis and then go back and say, OK, hey, here's a handful of things that didn't make the last media plan because obviously everyone's budget is finite.

41:22and we kind of filled up the finite budget with the stuff we know is going to work, is tried and true. But here's a few things that we see kind of out in the future are going to be important. And, you know, the agentic testing is we've used funds for that, for the different tests that we've run there is a great example. And that really allows us, as I said, to break through those, any of the resistance in those planning cycles. And it doesn't have to be Just to be clear, it doesn't have to be specific to, you know, if we're getting a lot of those funds from programmatic partners, the funds don't have to be spent on programmatic media.

42:00It can be spent on any kind of innovation that we see. We've spent it on some dynamic creative execution for clients, which may enhance programmatic, but it's not necessarily, you know, one for one and dollars are going back to the partners that contributed. And that's a part of the understanding. Was the stuff that you did with Swim.ai part of this innovation process? I actually don't think that the work that we've done with Swim so far has used innovation funds, but I would have to look into that. It could be, and certainly that's a good example of a partner where we're doing innovative stuff that would qualify for that program.

42:42Well, let's talk about it. So curation and the work that you're doing with them or that you've done with them, it feels like an interesting bridge to me between like more traditional, the more traditional programmatic world and like where the industry is headed. So a couple of years ago, you partnered with Swim to use their tool. It's called Scale, which is also very clever. Everyone's very clever. It stands for smart curation and learning engine. Well, they like they like acronyms. Our friends at Swim. I don't know if you know, but that stands for stop wasting your money. Oh, no, no. So why? So a little plug for them.

43:22Yeah, I didn't know that. So using this smart curation tool, you ran a display campaign for a financial services client. You ended up seeing good results, 56 % increase in their conversion rate, 26 % decrease in their cost per conversion compared to a control group. And then you were also able to cut down the number of domains the campaign ran on by like a ton, like more than half. So like less waste. So is this like the way everything is going, like smart everything? And was this just like agentic buying before we had a name for it? Well, there were certainly agentic and natural language components to that platform and technology, which was a big part of why we decided to partner with Swim as our self-service curation tool.

44:12So we do curation with a number of partners, including SSP-specific tools, but chose Swim as our partner of choice for the self-service side of that. And I mean, I think the short answer to your question is yes, things are going to continue to get smarter. And that kind of goes to a lot of what we've been talking about with the rise of the Gentic and AI technology overall. It's like, how do we make the stuff that we've already been doing better, smarter, faster, more effective? And that's been a big part of our relationship with SWIM. And I think also, if I think about it as a more of a macro trend in the space, it's about customization as well, which is why we like the self-service part of curation.

44:58And we see that happening with our clients in general, like less and less. It's about like, all right, we're Butler Till. We're going to bring kind of best of breed offering of all these partners together. And we've got good rates and that's our value proposition. And we'll save you money. And that's still part of what we do and we'll continue to be part of what we do. But I think a much more compelling story is this technology is allowing us to not just to kind of resell you something or to package something that somebody else made. But we can now create something, client, that is very, very tailored and specific to you.

45:35So if you think about that in terms of smart curation, we can have a very detailed natural language brief. And maybe that brief was created in part by different LLMs or different agentic technology, but then is going to put together a really, really nuanced, sophisticated, targeted, smart curation scenario where you're appearing in the right context, the right type of content that's not just brand safe, not just brand suitable, but is brand fortifying and brand affirming in ways that are much more aligned to what you're trying to do in the space than a standard buy would be. And so I see that as a big part of the future is you're going to see agencies in particular going to clients and saying, we are going to build something that is highly customized to your business need.

46:31And it's going to look different than the work that we do for the client to one side or the other of you. It's going to be very, very specifically designed and created for you. And the technology, which basically had us buying in the same way for everybody at one point in time, is going to continue to get more and more and more specific and targeted to you, which I think is good news. Well, how do DSPs have to evolve in a world where this is possible? because you essentially built a curated PMP of the impressions most likely to convert, updated it every day using signals from the buy side and the sell side.

47:11And I mean, the takeaway for me is that it seems like DSP algorithms, they're trained to find the impressions you'll bid on, but not necessarily the impressions that are going to perform. Like those aren't the same thing. And that, that feels like, not like a problem. It just feels like old school. So if that's the case, like how do DSPs play here? Well, I think it goes back to the question that you asked, like, does this make everything smarter? And I think it's going to, it has to make DSPs smarter too. And I'm not suggesting DSPs or anyone else are not already smart. But when you think about the way we started this conversation with your crack research into some of the articles I've written, the one about from Hagelin to Harmony is about how do we find smarter signal to transact on?

48:11And, you know, Chris Kane at Jones likes to talk about feeding DSPs what they eat, which historically has been cheap impressions and has been about efficiency and value. I think there's an opportunity for all of us to come to the table with better signal and data. And I want to give not just SSPs, but ultimately publishers better signal than just give me more of the low cost stuff that you have. I want to give them signal. And increasingly, you can do this, as you mentioned, with the PMP that's updated daily, almost a real time or a much more real time feed of like, hey, this is the stuff we're interested in.

48:58This is the stuff that's working. and that's going to have, you know, I don't mean to sound just too optimistic about the whole thing, but that should be in the, if done right, a really virtuous cycle to say like, hey, these are the contexts that our brands want to appear in. This is the performant inventory that we're seeing. I'm going to give you a rich signal to tell you that. So you're not just seeing, I want something for less money. You're seeing a rich array of signals that are telling you the kind of stuff that the buy side is looking for that should result in higher quality, better content over time.

49:36I think the reason that we've seen this race to the bottom and that we do continue to see cheap reach is that we're really only transacting on that per unit cost and trying to boil everything down to a CPM in our space. Yeah. I mean, less can be more. 100 i i often say and this is i i like to caveat this these are these are my opinions and not necessarily regardless of butler till okay it's subsidiaries of which we have none but um i think the future which is which goes to everything that we're talking about like there will be a future where at least in a in a relative sense there will be fewer ad opportunities so fewer impression opportunities.

50:25They will be smarter, as we talked about. They will be much more precise and targeted and relevant to the consumer, to the end user. And therefore, and this is where an agency person probably shouldn't say this publicly, they will cost more on a per unit basis because they will actually be more valuable. The other thing that we, I think, tend to look past in our space is bad advertising. It's like pollution. It has exhaust that we're pumping into the atmosphere. You saw that with things like ad blockers some years ago, which, of course, persist, which is a reaction to bad advertising, intrusive advertising, advertising that is not only not adding value, but is actually subtracting value from the experience.

51:21And if we are out there polluting our own space, it's a short-term thinking proposition. I think we all need to be conscious and aware of that. And we all have to live in the world. So if some people don't care about the environment, it's not good for anybody. We all need recipe sites, right? At some point. I definitely do. I don't know what I'm doing. So I'm going to cap off the episode if you're game with a little lightning round, but short answers only. Thank you. Five seconds or less. Nobody listens to me when I ask for short answers. So five seconds. I'm going to try. OK, I can only ask for you to try.

52:04So if you had to bet five years from now, are there still human media planners running programmatic campaigns from start to finish? Or will that be the exception or not even happening anymore? Five years from now, which is like a long time, actually. Human media buyers running campaigns, but heavily augmented by agentic technology. That was less than five seconds. Thank you. Thank you, Scott. Okay. DSPs, will they successfully adapt to an agentic world or will they become like the blockbuster of ad tech? DSPs will adapt and persist in the new world. Will walled gardens like Google and Meta embrace ad CP and open agentic protocols or will they build their own versions and keep them closed?

52:58I know which of those I would like to have happen, but I think they will build their own versions and keep them relatively closed. You are so right. I agree with you. What's something, and it could include something in your personal life, that you would let an AI agent run unsupervised?

53:19Oh, boy. I think shopping. I hate shopping. I hate buying stuff. I, again, was just on a call. Same. Yeah. With, yeah, an industry call and someone from a big retailer. I'm being vague about this because it's kind of Chatham House rules in this group. But the retailer was talking about this idea of shopping missions being powered by a gentic. So you basically say, I'm throwing a birthday party. Just go out and buy all this stuff. All right, please take it from me, agents. You can have it all day long. So I'm going to scoop and I still have a few lightning round questions, but sidebar. So our our executive editor, Sarah Sluice, she's working on an op ed that we're going to publish this week.

54:14that is about her wistful feelings when she realized that she could outsource a lot of back-to-school shopping to AI. It was able to digest the lists of supplies that her kids needed and put them into a shopping cart and all she had to do was check. But there was something missing from the experience because, you know, you can't smell a pencil. I mean, whatever. You're not smelling pencils anymore. or you're buying it on Amazon, but there was just something lacking from the experience that didn't feel like the beginning of back to school season. And she was a little bit sad about it, but also had to acknowledge that it was so much easier.

55:01Yeah. No, I feel the same way. I'm nostalgic for the way we used to do things. And I'm, you know, getting magazines once a week in the mail instead of having the information at your fingertips, strolling the aisles of a, you mentioned Blockbuster, of a video store with your friends. But at the same time, I can recognize, wow, it's so much better now than it used to be. And I think we're going to largely feel the same way. What is something you'd never let an AI agent do without human supervision, no matter how good it got?

55:43yeah um i think wait and then this is kind of goes against what i said about shopping i think like choosing meals or like recipes and cooking for me like i really cooking is one of my zen activities like i that whole process is something that i don't think i could let go of And you shouldn't have to. Well, I mean, once the campaign is running on its own, you'll have tons of time for cooking and trying out your recipes. Are clients into the idea of agentic buying or are they feeling some hesitance? I mean, I think both, they're both into it and feeling hesitant, but I think largely clients as they are with a lot of tech are excited about the possibilities and what it could solve for them.

56:37And then last one, what is the industry buzzword that you would most like to see die, never hear again, kill it with fire? Oh my God, please God, don't use this buzzword in front of me anymore, says Scott Ensign. this is a dangerous one because i am like 1000 likely to use any buzzword in the future and totally be a hypocrite um i would go back to what i would go back to what we talked about with human in the loop i think we need to find better language about how people are going to be involved in um in this new world and i don't think human in the loop is it so i'd be fine with getting rid of that one.

57:19Well, from one human in the loop to another human in the loop, it's been a pleasure. This has been great. It's a lot of fun. Thanks so much again for having me.

57:38This episode was sponsored by Verve, an ad solution that helps brands activate consumer intent in real time across platforms. As AI reshapes how people plan vacations, Verve helps travel brands show up when it counts. Find out more at www.verve.com. That's V-E-R-V-E.com.

58:12Thank you.