10 Minute Martech · 2026-07-07 · 14 min
Key moments - from our scoring
Substance score
55 / 100
Five dimensions, 20 points each
Connected television has fundamentally shifted how brands buy video inventory - from bulk linear TV purchases to granular, audience-targeted campaigns. Greg Joseph explores how AI amplifies this capability through tools like StackAdapt's AI video creator, which allows marketers to generate multiple video assets (20-30+) in minutes rather than weeks, with real-time creative optimization for live events. The platform addresses the AI hallucination problem through intentional input design and human oversight rather than fully automated workflows. Joseph illustrates this with the World Cup example, where brands can create regionally personalized messaging for Brazilian, Caribbean, and European audiences watching the same event. He emphasizes that successful AI adoption in marketing requires willingness to break legacy processes and embrace automation, paired with rigorous output validation. The conversation reveals a practical tension: AI automates operational, entry-level tasks, but human judgment remains critical for strategic decisions and quality assurance. This matters for marketing leaders evaluating CTV platforms, creative production workflows, and organizational readiness for AI-powered personalization.
StackAdapt's AI video creator allows buyers to upload a link and generate video assets for CTV in real-time within minutes, versus the traditional weeks or months of production. Marketers can now create 20-30+ video variants instead of 3-5, and adjust creative dynamically during live events like sports games without waiting for new asset approvals.
The platform uses intentional input design - carefully limiting what data the AI reads - combined with human oversight. Rather than forcing AI to process large databases, StackAdapt adds people to monitor and check outputs, ensuring the right messaging appears before ads run live.
The World Cup campaign demonstrates personalization where brands create distinct creative messaging for Brazilian, Caribbean, and European audience communities watching their respective countries, rather than using generic regional targeting. This requires 20-30+ video assets customized for each community's values and interests.
According to Greg Joseph, willingness to break established processes and embrace innovation is most critical. Companies must be willing to disrupt legacy workflows and adopt AI automation tools, even though it carries short-term risk, because the upside of innovation outweighs the cost of staying behind.
Linear TV required bulk purchasing of all users regardless of audience relevance, while CTV enables selective targeting of individual markets and audiences. This allows brands to find the right person at the right time with personalized creative, supported by detailed audience data unavailable in linear TV.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode contains some useful operational insights - specifically about CTV's shift from bulk buying to audience targeting, AI's role in creative production at scale, and the importance of intentional inputs to avoid hallucinations. However, much of the content is surface-level explanation of known trends (personalization, automation, AI in marketing) rather than novel or non-obvious claims. The advice on prompting and quality assurance is basic and somewhat self-evident.
CTV has allowed for all of that to change. Where you can be as selective as you see fit, where you can target individual markets, individual audiences
one reducing the cost, but creating more automation and more efficiency where buyers don't need to spend countless hours thinking about all the different assets they need
The thinking here is derivative of mainstream martech discourse - personalization at scale, AI reducing production costs, the need for human oversight. The Bethany Hamilton tangent and the 'willingness to break processes' framing are personal philosophy rather than original insights about marketing or technology. No contrarian perspectives or first-principles reasoning emerges.
the one thing I'd probably say is willingness. Um, so right now if you speak to many agencies, a lot of them follow processes on a day to day basis
AI is here to stay, but people will not go away. People will always be part of the equation
Greg Joseph is a VP of inventory development at StackAdapt, a real martech operator with relevant day-to-day experience in CTV and AI ad platforms. He has practical knowledge of how the tools work. However, he is also a vendor discussing his own company's features, which creates some inherent bias and limits his independence as a neutral practitioner expert.
That's Greg Joseph, VP of inventory development at StackAdapt
a few weeks ago we released our AI video creator, um, where essentially we're giving buyers the ability to put in a link and we'll create a video asset for CTV in real time
The episode includes some concrete examples (20-30 video assets vs. 3-5 historically; World Cup targeting for different geographic communities; Knicks vs. Spurs real-time sentiment shifts) but lacks hard data, metrics, or case studies with results. The World Cup example is illustrative but not backed by numbers. No specific dollar figures, conversion rates, or measurable campaign outcomes are provided. Client anonymity is cited as a reason for withholding details.
they would essentially use let's say three, four or five video assets. Now they're able to have as much as 20, 30 plus
you're watching the Knicks and Spurs go head to head and there's a blowout with the Knicks and beating out the spurs. You may want to change your creative asset in real time
Sarah asks reasonable setup questions but largely allows Greg to deliver prepared talking points without meaningful follow-up or pushback. Questions are often open-ended softballs ("Can you talk about that a little?") rather than sharp or challenging. There is no productive disagreement, no probing on limitations, and no tension in the conversation. The interview reads as friendly but unchallenging.
Yeah, that's very cool. I mean, when you think about it, we've been talking about personalization for a long time, right and so this is when you add CTV with AI, you actually have that opportunity for hyper personalization
Yeah. Ah, that's great. Now, would you have any recommendations for people?
Computed from the transcript - who did the talking, and the words that came up most.
The World Cup has US audiences from South America, the Caribbean, and Europe watching the same matches and caring about completely different things. Greg Joseph says brands finally have the tools to show up for each of those communities individually. As VP of Inventory Development at StackAdapt, he works where CTV inventory meets the AI tools rewriting creative production. He explains how buyers now run 20 - 30 tailored video assets per campaign instead of three, how an ad can change in real time during a blowout game, and why most AI hallucinations trace back to sloppy inputs rather than bad models. He also names the martech skill he values most: the willingness to break processes with millions of dollars attached.
Transcribed and scored by The B2B Podcast Index.
Speaker A: I'm Sarah Fatz and I lead community and awareness at Progress. This is 10 Minute MarTech.
Speaker B: The World cup is going to be very personalized to people all over the US Brands as they're thinking about showing up for these communities, they have to be as personalized as possible. To have that level of personalization has been great. And we're seeing a lot of buyers dabble with that, which I think gets me excited on the innovation that exists with artificial intelligence that historically hasn't existed in the past.
Speaker A: That's Greg Joseph, VP of inventory development at StackAdapt. Let's get started. All right, Greg, I actually want to start by asking you m a little bit more about Connected Television or ctv. How do you think that fits into a well thought out Martech strategy? Um, and how can marketers be thinking about it in a, you know, in an era of AI?
Speaker B: So maybe one of the things I'll do for the CTV and kind of answering your question is look back like 30 years ago. So when, when you looked at 30 years ago when folks were permanent buying linear TV, they'd essentially buy TV in bulk. So you bought every single user independent of whether or not those audiences were relevant, um, to your brand. CTV has allowed for all of that to change. Where you can be as selective as you see fit, where you can target individual markets, individual audiences, meaning that your brand can find the right person at the right time. I think one of the ways AI is helping CTV is when you think about the notion of creatives, uh, that are specific, more specific to substantive audiences, but being able to do that at scale, uh, CTV creative productions are very expensive costs. AI is now one reducing the cost, but creating more automation and more efficiency where buyers don't need to spend countless hours thinking about all the different assets they need for a CTV environment. They can essentially upload it to a platform like stack it up, create the ads within a few minutes and be able to run on, um, connected televisions. But the additive piece is really the audience, uh, component. And today CTV is one of the many devices that offer the ability to do so. Unlike Linear tv, which buyers had to just buy in bulk and had very little data on what audiences are reaching now they can use platforms like stacknap as well as others to find those audiences at the right moment and the right time to allow brands to resonate with some of the audiences they're trying to reach.
Speaker A: Yeah, that's very cool. I mean, when you think about it, we've been talking about personalization for a long time, right and so this is when you add CTV with AI, you actually have that opportunity for hyper personalization. Yeah. And I love your comment about creating content at scale. Can you talk about that a little bit? Because I know when we talk to guests here, there's often this concern that AI is going to create a slop. Can you give some examples maybe of how you've seen it done?
Speaker B: Well, what we're seeing, uh, a few weeks ago we released our AI video creator, um, where essentially we're giving buyers the ability to put in a link and we'll create a video asset for CTV in real time and that ad can also shift. So as we think about CTV and then live sports, they can be instances where you're watching the Knicks and Spurs go head to head and there's a blowout with the Knicks and beating out the spurs. You may want to change your creative asset in real time. And when I say in real time within that four hour window of the game, because this blowout game and you want to have a sentiment that more is more linked to the users watching it, now you have the ability to do that with AI where you don't have to wait weeks or months to get the right creatives in front of the users watching content. You can now do all of that in real time and really find a scalable solution for something that historically has not been scalable.
Speaker A: Yeah, yeah, that's great. And yeah, I mean to be able to do something that quickly, are you finding um, with your platform that you're able to eliminate? I mean hallucinations are a part of AI, but in that real time creation, are you seeing that that's not really an issue with your, with the creatives?
Speaker B: Uh, yes, at the moment. So one of the things we're very intentional about is the inputs. So quite often for many platforms input a number of different things into that platform where AI does hallucinate from time to time because it has too many things that it's looking at and folks are not being as descriptive as they need to within that platform. What we've done is we've not only added AI solutions, but we've added people to kind of monitor and check things along the way to minimize risk to the business. Because one of the worst things someone could ever see is running an ad, ah, that says it has the wrong messaging on it because it pulled the information from the wrong place. So one is how do you still add a people layer to things that you're working on and what artificial intelligence does solve with automation, quality assurance is just as important. So how do we make certain we're solving for making certain things are the way it's supposed to be, um, which I think has been key for us. But being very intentional with the inputs is the most important thing. What is AI reading, um, within your platform? And if it's reading the correct things, then great. When, uh, you have too many things, if you have a database of 200,000 items and you're forcing it to read it in real time, it will then start to make a lot of mistakes. So, uh, intentionality has been a big part.
Speaker A: Yeah. Ah, that's great. Now, would you have any recommendations for people? I mean, going back to the comment that you made about prompts, like, if your prompt is not right, obviously it's going to cause issues and humans are messy and communication can be all over the place. Do you have any guidance for people who are trying to learn to be better prompters, to be, you know, we're all becoming prompt engineers, whether we, you know, whether we want that title or not. How, how would you coach somebody, uh, to think about prompting?
Speaker B: One of the things that I think it's very important as people are, um, utilizing AI tools, um, is one. Are you willing to spend the extra time to kind of go through, um, what the data is pulling or what it's putting out? The output is just as important input. So how do we make certain that we're giving the extra eye on the output? So I think a lot of people miss that. And that and while it seems very, very basic, is actually a very big issue. Because if you're getting the output that's incorrect, the input is kind of useless. And I think that's one of the things I would recommend to anyone who is now dabbling into artificial intelligence and its uses for their business. Do not throw away kind of the additional additive part, which is the last layer. I think just double check things. And that's no different than double checking your ad and making certain the spelling is correct. Words are very important. So, uh, for me, it's the double checking of, uh, things which I think is the most important. Ah. As people are thinking about, um, how best to utilize, um, artificial intelligence.
Speaker A: You guys work with a lot of different, a lot of different brands, a lot of different customers, a lot of different clients. Are there any, um, interesting or unique things that you can share with us, uh, campaigns that people have run or programs that they've run recently that are showing that personalization at scale or have really, you know, they've moved the needle in A way that, that even surprised you guys?
Speaker B: Well, I can't share individual campaign examples, uh, for obvious reasons to protect our clients. What we're noticing is that a lot of our clients are now with um, our AI tools are creating like 20 different video assets where historically when we would run with them, they would essentially use let's say three, four or five video assets. Now they're able to have as much as 20, 30 plus to use for campaigns. And uh, what that's a lot of which I think has been successful in our platform and gets me excited is now they're being much more intentional on how they break out the campaign and how they think about showing up for communities. So I'll use the example of the upcoming World Cup. The World cup is going to be very personalized to people all over the US you have some people from South America, some people from the uh, Caribbean, which my parents are from, you have some people from Europe. And all of them are going to be watching their countries play and participate in the World cup brands. As they're thinking about showing up for these communities, they have to be as personalized as possible. So if you're showing up for a community, um, let's say a Brazilian community, there may be messaging that you want in the creative that's very different than showing up for a community in Western Europe and very different for showing up community in the Caribbean where all of them care about different things tied to the team. So to have that level of personalization has been great and we're seeing a lot of buyers dabble with that, which I think gets me excited on the innovation that exists with artificial intelligence that historically has, hasn't existed in the past.
Speaker A: You know, especially when you start thinking about global campaigns and global programs, really thinking about how can you be efficient as you mentioned, but also that, that hyper personalization that goes beyond the hello first name or you know, let me just target. Because you're in a, you're, you're in a specific region. What do you think right now is uh, the most important skill a marketer or a Martech professional could have today? I mean we've touched on a number of different things, but do you have, if you could narrow it down to one thing, what do you think that would be?
Speaker B: Uh, the one thing I'd probably say is willingness. Um, so right now if you speak to many agencies, a lot of them follow processes on a day to day basis. Most of them don't break those processes. So if anything falls out of those processes, they'll just not do in order for any company to be successful today is that one they have to realize that innovation is a real thing and then processes need to change and they have to be willing to break those cycles in order to improve their businesses. So while it sounds like very something very easy and simplistic, it actually is very hard for companies to do because sometimes you have not only like tens of thousands, maybe millions of dollars associated with processes that you have in place and breaking away from that comes at a risk. But the benefit, there's always upside to risk that you're willing to take. And for me it's always that willingness or can you break away from um, legacy processes? And then I'll kind of go back to like artificial intelligence as well. A lot of companies are finding that they need to change in this age of innovation and they need to start embracing more technology. The companies that will be left behind is companies unwilling to take the step forward and willing to break processes and drive innovation and automation through some of the tools that um, artificial intelligence is helping to build. Um, so my biggest recommendation is break away from cycles and be willing to um, embrace change. And that willingness alone goes so far in most companies and businesses.
Speaker A: Who do you follow uh, for inspiration or information right now?
Speaker B: Okay, uh, I'm going to take a slight pivot to this one because a few weeks ago with stacking up how it's conversion event and we had a phenomenal speaker on stage that most people would expect. Her name was Bethany Hamilton. Uh, she was a surfer. Um, our arm um, bit off as a young girl. But one of the things that impacted me and I actually did some research on her was the way she kind of spoke about her perspective on life through the lens of gratefulness, um, putting things into perspective and then resiliency which I think for all of us as we think about like our day to day life one is um, gratefulness is such a great way to think about life because yes a number of different things could go wrong. But finding positive outlooks to things around us is probably the most important thing that we, we had in life. And then resiliency. I think when we think about work today most of us run into challenges, whether it's with work, our families, with a number of different things. Resiliency is such a good thing to have like top of mind where we think through now how do we get past this issue? How do we overcome obstacles and really think, think about that on a day to day basis versus quite often it can be easy to things because um, there's a number of different routes we can go. I'm not, I'm a big person. I like to focus on things. So one of my hobbies during the weekends is I love to farm. I love to grow vegetables. There are times where it takes me six hours just to do one thing. But it is great for me because one, I get to think about it a little more and understand what I'm doing wrong. I get to understand what I'm doing better. So my tomato plant will actually grow tomatoes and not just die out within a few weeks. I think the notion of resiliency is so important, so I'll probably say Bethany Hamilton. But she comes top of mind and her perspective has been great.
Speaker A: Yeah, that is fantastic. Wow. Yeah. And she is incredibly inspirational. Um, but I do love, I love the sentiment of, uh, gratitude and resiliency. I think we talked a lot about resiliency during the pandemic and I haven't heard people talk about it quite as much since then. And I think it's one of those skill sets, especially when it's going hand in hand with gratitude, that really will take you far. Um, so that's wonderful. One last question for you. Um, yes. Do you have a Martech hot take?
Speaker B: The amount of people coming out of college and going to the workforce today? Um, I think one of the things that's going to help them with a lot is how do they adult experience before they go into the workforce. So artificial intelligence. Yes. Is able to do a lot of kind of day to day responsibilities that's more operational in nature and that's focused on entry level positions. But one of the benefits to that is maybe some people do spend more times going for the masters in schooling or spend more time taking internships with companies which stack it up. Would love because, um, there's a lot of value in training up the next generation. So my hot take is AI is here to stay, but people will not go away. People will always be part of the equation, which I think is important to know.
Speaker A: Yeah, uh, I love that and I've really enjoyed our conversation today. Thank you so much, Greg.
Speaker B: Uh, thank you for having me. Appreciate it.
Speaker A: Listeners, thanks for tuning in. Make sure you like and subscribe wherever you get your podcast. Until next time, I'm Sarah Fat and this is 10 Minute MarTech.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.
Burnout: Why Thought Load, Not Workload, Is Draining You with Liane DaveyBeyond the To-Do List · on AI hallucinations95 / 100
Why the performance marketing era is overBrandformance · on Linear TV88 / 100
Episode 7: AI & the Power of a "Thin Core"Architecting the AI Enterprise · on Prompt engineering82 / 100
477. The Nitty Gritty of AI From an Attorney and AI Expert with Mike BrownThe Game Changing Attorney Podcast with Michael Mogill · on Prompt engineering81 / 100
The Hidden Risk in AI-Generated Tests and Requirements - Olivier DenooSoftware Testing Unleashed · on Prompt engineering79 / 100
John Suarez on How Businesses Should Actually Use AI Without Losing Human JudgementMarketing for SMEs · on Prompt engineering78 / 100Neil Tewari: Marketing Can't Be the Company's AI Bottleneck61 / 100Bob Regular: The AI Ad Fads That Won't Survive64 / 100Allan Malcolm: The Future is Marketing Generalists, not Specialists54 / 100Kabir Daya: Human Connection Still Drives Discovery66 / 100Heike Young: Taste Gets You in the Room; Results Keep You There64 / 100