Craig Ferguson: Hi everyone and thanks for joining us. My name is Craig Ferguson and I'm a director of Regional Sales for Europe at Evergent. Again, thanks for tuning in to this podcast where we're going to be focusing on addressing the ongoing trends around churn and retention, exploring new ways of streamlining operations to reduce overhead, and uncovering the strategies behind smarter content acquisitions and licensing.

We're at a little bit of a pivotal moment where the organizations that are succeeding are those that are leaning into data-driven life cycle management, AI personalization, and integrated partner ecosystems. At Evergent, we are supporting these changes globally, managing over 1 billion subscribers on behalf of our customers across 180 countries for brands such as the NBA, Sony Live, Astro, and many more companies in the media space.

Joining me today to share their thoughts are Tom Dvorak, co-founder and CCO at XroadMedia. XroadMedia are experts in behavior-driven personalization and discovery, helping operators surface the right content at the right moment. 

I am also joined by Kevin Savina from AWS who supports media organizations in building scalable end-to-end media operations, including cloud-native data workflows that power personalization, churn modeling, and intelligence decision making. Thanks for joining Tom and Kevin. Would you like to take a moment to introduce yourselves and let us know about your companies?

Tom Dvorak: Thank you very much, Craig. My name is Tom Dvorak. I founded the company Xroadmedia Media about 13-14 years ago when we took a team out of Philips Electronics. Our focus is very much on personalization and we pride ourselves to personalize every touch point the user has with digital media services—this being user experiences, notifications, newsletters, and other touch points. We are active in the media and entertainment industry, have more than 50 deployments worldwide in the market and we serve more than 200 million users on our platform. 

Kevin Savina: I'm Kevin Savina. I've been at AWS for five years after a long career in tech for media. My role here at AWS is -I lead the AWS partner strategy for media, entertainment, games and sports in the EMEA region. Really working with all of the community of partners—system integrators, software vendors, managed service providers—that help service the media, entertainment, gaming and sports customers here in EMEA for their what we call the line-of-business needs. Basically running their operations. Thanks for inviting me for the podcast.

Craig Ferguson: Thank you both for being here. So, let's start with the big picture. As we begin to move through 2026, I feel that the streaming landscape is continuing to shift from growth at all costs to more profitability through precision. We’re seeing a little bit of a convergence crisis right now where consumers are increasingly hesitant to manage dozens of siloed subscriptions. Instead, I think that they are really demanding a unified high value experience from Evergent’s perspective.

There are a couple of story points. We're seeing a real rise in hybrid monetization. I think we saw that by 2026 over 81% of platforms are going to have adopted some kind of hybrid model, blending SVOD, AVOD, TVOD to really maximize that revenue per user. We're also seeing companies lean more towards global scaling. We’re seeing a lot of growth in regions such as APAC, it's predicted to reach 54 billion dollars in OTT revenue by the end of this year. And finally, there's the expectation of personalization; it's now considered the industry standard with 86% of users intending to keep or even expand their subscriptions if the value and the discovery experience remains high enough.

So again, just in general, I am beginning to see the shift more towards this mixed monetization, combining independent revenue streams to maximise every single customer touchpoint. Yeah, I mean, Kevin or Tom, Considering these movements, what are primary trends that you both foresee dominating the media and digital service ecosystem through the rest of the year? Furthermore, how are consumer expectations evolving in terms of content value and discovery?

Tom Dvorak: I mean, you are completely right. What we see in the last few years is three areas and we are going to see this in the last few years, and we are going to see this most likely for this year as well, is actually 3 areas to help the user discover content more easily. First  of all–need for Personalization used to be nice to have in a service, but now users expect it because the different services coming in like, Netflix and Disney plus and the others. For example -the State of Play 2025 report from Gracenote said that more than 45% of streaming viewers say that the saturation of streaming services and content is overwhelming for them.

So it means when you can do personalization, right? And I think we are far from the main stream providers, but if you do it right, I think you can really separate itself out of your competition.

Second thing is aggregation or unified user experiences. This use to be the play of typical telcos and PaytV service providers, but with more fragmentation of content, its getting more and more difficult, but we see a shift moving a little bit into the other direction, driven by two main reasons. One is obviously consolidation- we all know the WBD saga going on at the moment in the market, where basically content is converging now into single services.

But the other thing as well is the start of the willingness from content owners to share their content also on other platforms. So what used to be a few years ago, from a completely aggregated point of view, let's say broadcast TV, for example, into a completely fragmented situation with every content owner launching their own consumer services is now coming back a little bit more towards a more aggregated view. So we see this, for example, in the UK, ITV sharing their content on Amazon Prime Video. Also, Sky just recently announced that they will be adding, in addition to the Netflix package, Disney+, and also HBO content to their subscribers in the Sky packages, too.

So, I think all of these are the right noises, I would say, going into the right direction for, obviously, better aggregation, better discoverability, and users, you know, leaving less to search and more to watch, so basically putting them first instead of content first.

Kevin Savina: it's interesting, because I agree with a lot of the trends that you mentioned, but from the AWS point of view, we see them slightly differently, especially with my role, where I go across media, entertainment, games, and sports, which is a slightly wider point of view.

So, coming back on the first topic around the diversification of the monetization models, this is part of breaking some of the silos for revenue, but also how audiences and consumers are watching . and I think you mentioned that some say, broadcasters are trying to work with other streamers to have different distribution platforms. I think one of the big disruptors that is causing that is the viewing patterns, especially the younger audience's viewing patterns. I mean, if you look at Gen Z, for example, over 58% of Gen Z viewers watch more video on social media than on television or streaming platforms. And that is forcing all the content producers, content distributors to rethink, in a much wider sense, the way that they are distributing their content and they're engaging with their audiences. Some that are on their platform, some that are not on the platform, or through a different platform. And so then you mentioned how do you then engage with these audiences that are more fragmented, and where you have touchpoints that might be through a wider variety of platforms, and that's where smart personalization and even hyper-personalization is starting to become a kind of a new baseline.

And as a content provider, you have to think of your viewership and your audience across all of these platforms, and serve them personalized content, the way they watch it, the way they want to watch it, at the time they want it, on the platform that they want it, and really think of an omni-channel strategy to engage them, using social media to bring younger audiences to your paying platform, or to a monetizable platform through AVOD, or through fast channels.

So, there's two competing forces here. The willingness of simplicity for the users, and not having 50 different subscriptions, but at the same time, the habits of wanting to see different content from different sources at different times on different platforms. So that's something that our customers have to grapple with

Craig Ferguson: I think there's a real fine balance across these services and trying to find that level of flexibility, but then also, being able to offer packages that allow these consumers to engage on their own terms, right? Like, there's so many choices out there for these viewers. But a lot of the services still today offer your standard packages, whether that's monthly or yearly, and then you've got your traditional PVOD approach there, but I think that we need to be looking into, or some services need to be looking into diversifying that a little bit more, realizing that people might want to have 3 months of access, or, a little bit more. They want to be able to pause their subscriptions while they go on vacation. So, I think it's finding a line between everything that you're both saying there as well. But yeah, please continue Kevin.

Kevin Savina: one of the other things that we are observing with the, thousands of customers we have across the world that are… that are serving their audiences through AWS platforms. Is that there's also a technology convergence between, say, traditional media, streaming, telco, but also the usage of data, which we'll talk about a lot today, I'm assuming, but also cloud to be able to operate distributed at scale, and leveraging, genitive AI or some of the native media services that are available on a platform like AWS. Because the media distributions, the distributors, the operators, they need to be able to experiment at speed, try new formats, try new ways of engaging their audience. And that technology, which is merging cloud IT, traditional video broadcast streaming capabilities. It's really also changing the way our customers are thinking about engaging their audiences with more personalized content, products, ways of distribution.

Craig Ferguson: it's a really valid point, actually, and something that's really relevant to a case that we had.  There's these customers that really want the ability to experiment with these go-to-market approaches. We had a really large sports organization working with us, and they wanted to experiment with a 10-minute pass for a match, where a user would come on, they would subscribe really quickly, frictionless experience, and they'd be on there for the 10… the last 10 minutes, the first 10 minutes of the game. Turns out it wasn't the approach that they wanted. I don't think the results were there to justify this, so we ended up rolling it back, but it's all about that experimentation, Kevin, that you're talking about. Some things are going to work, some things aren't going to work, but as an industry and the position that we're in, I think we need to think about flexibility. We need to think about changing, the status quo that we've got right now.

Tom Dvorak: Yeah, and maybe let me just quickly add to what just Kevin said before. I think which is super important, is what Kevin more or less pointed out is that, for quite a long time now, even content production is moving towards the user's needs, right? So, as you just said before, right, they need to change the way they produce content based on the different generations, for example, they would like to address.  Which, again, then I think we're going to discuss this later on a little bit more, probably, anyways, which will also change monetization. And I talk about not only content monetization, but also, for example, ad monetization, right? So what we see as well is the shift also away from the traditional 30-second spot into, okay, how can I make it more immersive? And I think Amazon Prime has started, very well. I just watched the game yesterday on Amazon Prime. Where you can now also immediately put something into your basket when you see an ad you like, for example. So, immersive is one thing. But then also, I think where we're still lacking behind a little bit in that sense is to get the right inventory in ads as well, to make those kind of experiences. So, I think this is another shift we will see, together with kind of data aggregation, monetization, and user-centric content production, which I think will make a huge shift in the next year and years to come.

Kevin Savina: Completely agree.

Craig Ferguson: Really good points. So moving on, I think as we scale, something else that we're really starting to see becoming a problem, and it's been a bit of a hot topic for a while now, is the idea of user churn. A problem that really only becomes more severe with a larger scope, more global growth is going to have a direct impact on that growth is essentially… can be meaningless if you don't have the capability to keep the customers that you've worked super hard to wire. I think there's some good news here, that some services are starting to realize that, and they're starting to pivot to meet this challenge. 

There was an example that we actually wanted to talk about with a major Japanese video gaming company, so it's going a little bit away from the video ecosystem, but it's something that really powers this idea about lightweight AI churn prediction deployments. Some stats here within 8 weeks, using over 2 million behavioral data points. Evergent and this company, they were actually able to hit a 94% accuracy for churn prediction, so we could see the users, the way that they were engaging, were more inclined to be dropping off, which allows service, or this particular service, to come up with a hyper-targeted offer or a retention campaign for these users, which really did result in significant uplift in user engagement.

Guys, from your point of view, it's a really big topic, but,, in terms of lifecycle-based personalization, in terms of a defense for churn, or other topics around about churn, from your perspectives, is there any kind of thoughts that you have around that point?

Tom Dvorak: What we see in that, addressability or personalization is, usually back in the days, you were segmenting your customers by typical things like demographic and these kind of things, and we see this, actually moving away a little bit, because as we said, it's more interest-driven, more user-driven now, and even within the user, depending on the time of the day, for example, the season, for example, interests shift. That's number one. 

Number two. within, let's say, households, which we usually talk about when we talk about media and entertainment, we don't talk about a single person, but sometimes, particularly on the TV sector, we talk about households, so you have multiple users you have to address simultaneously.  So, we see segmentation actually coming a little bit more into the picture again, but more from a behavioral perspective, rather than from a demographic perspective, right? So, from a lifecycle perspective, so what we see and what we advise our customers is, you have to address different user segments also differently. So you have to show, for example, more of a catalog to new users than you have to do, for example, for your power users, who are coming back with a purpose, with quite clear, ideas in mind, what they want to watch. So you have to address them differently, not only one-on-one, but also one-to-many.

 So basically, meaning that you just follow different use cases. So you say, for example, if a new user is, okay, when there is what we call a cold start, so we don't know anything about the user, there are different means to hook them into the service than, for example, as I said, for people who are coming back more regularly. So, we have seen this, and we do this with a number of customers where we do, like, one-on-one, and also one-on-many personalization. And when you follow this kind of lifecycle personalization, as you put it before, Craig, is that we have seen an increase in content consumption by more than 20% from those segments. Which is quite substantial, from a, let's say, video consumption perspective.

Kevin Savina: The Second thing, if i can add something, you made an implication in what you said, which may not be obvious for everyone. You started tying together audience behavior, segmentation of that audience into different behavioral patterns. but also how you use the right content to address that and I think there's an underlying assumption that we're making here that is maybe not clear for everyon–, is that having the ability to bring all these many different sources of data, the content information, the consumption patterns, the behavioral data of your audience or audiences, however you segment them. Maybe some advertising information. That bringing together all of that data is actually really crucial To be able to be effective. In that, over the life cycle of…a certain viewer or a subscriber. And if you look at it from the AWS perspective, we've seen a lot of the projects dealing with personalization or churn management, or even ad optimization, delivering suboptimal results, because the customer hadn't thought about that global data view. And I think we'll talk about it later in the podcast, but I just wanted to mention it here, because it's kind of obvious, but it's not always implemented the right way,

Tom Dvorak: That’s a very important point and thank you for bringing that up, actually. I think the complication is a little bit even further than that, because if you have a service provider who has multiple business models,, a single subscriber might be in different life cycles in those different business models, right?, it can be, like, for a paid TV service provider, for example, they can be what's usually the case, a power user, if you like, on linear TV, but they're just an occasional user on SVOD, so you need to address them differently, depending on where they are. And that actually brings me to the next point, that It's also important not to stop with addressing users in a meaningful and personalized way when they leave the user experience. You also need to understand, when are they, willing to consume your service. when they're not in your service. So what I mean by that is, we can, for example, identify viewing opportunities. So, for example, users are commuting, so why don't you suggest something for them to download on their device, to watch, on the commute to work, for example. Kind of, expanding the personalization from a typical, dynamic user interface, personalized dynamic user interface, into personalized notifications, newsletters and these kinds of things.

And what we've seen there as well is, we have, for example, increased click-through rates on newsletters by personalizing them by more than 60%, and that then, increases retention rate. So I think that's another important point, is you can't just let the users go when they're not using your service, and you should also not do things like, typical services sending out,  by the way, it's prime time, don't you want to watch some TV right now? It's more kind of, be more precise, don't annoy them. and make them see the value, of your service. 

And I think the last point and I think we will probably discuss a little bit later and Kevin probably has more to say about that is– how you best use the data available to you. Kevin mentioned before, different data points you can use. And as you said, you're mixing the information, for example, Virgin have on the user's lifecycle to, what AWS has, based on the user, consumption and what we have based on the profiling information. If you put all of the data together, you have a very, very rich data set where you can pick and match, basically, what you would like to… or how you would like to address the users.

Kevin Savina: Yeah, and that leads into then how, once you start having that very rich data set with many different data sources? How do you leverage some of the, data wrangling capabilities? How do you use, for example, generative AI to,  To, personalize a given message, or to create a temporary segment and segmentation based on, different behavioral, data, what content you have, how these audiences that are not on your platform are behaving in other areas? And how do you use these emerging, agentic AI, to help drive that personalization? Not by automating 100%, but kind of thinking about how you extend your editorial intent across all of these with a scale that is manageable, so…

Craig Ferguson: Yeah, I think this is a topic, honestly, that we could have a podcast on its own for, because it just… the topic just keeps going on and on, and, even brought thoughts to my mind regarding once you have that data set, and I know, Kevin, that, Evergent and AWS work really closely together on building that data portfolio, and we have access to a lot in there. But there's so many steps of evolution you could go further. I mean, if you  identify a certain segment. Tom, you were talking about segmentation. You could identify a segmentation of users that are more inclined to be watching, lifestyle TV, or more TV that links to a social environment. And, you could even identify that people, and those people even come up with some kind of targeted offer to say, maybe you want to enjoy this content with a friend or a family member. That might make it more of an inclusive experience for  everyone, and then that can potentially open the door to bring more people in, increase stickiness in the consumers, because they're not going to drop off, because it's more of a social thing where you can bring people together. 

So, I think there's just so many steps beyond that you could explore. probably enough to have a topic on its own about.  So, I guess in terms of all that, we can predict who's staying, who's leaving.  But doing that at scale, I think that, that's one part of the problem, 

Craig Ferguson:  we were talking earlier about the technology side of things as well. There’s the idea of modern efficiency. We have seen firsthand with,a recently with a major Pay TV provider who actually replaced a 20-year-old legacy ecosystem with Evergent in just 9 months.  It wasn't just a modernization project. They really went as far as to really reduce the total cost ownership of this platform by 30%. And something that was really unique, a story, a really good news story that I've not heard enough of, I think, in our industry, was the fact that the cost savings that this the rider managed to get, as opposed to just feeding that back into the organization, they actually passed it on to their subscribers and sent a notification out to say that they were going to drop their fee. That's something, I don't know about you, Tom, or Kevin, but I constantly get messages and emails about how my service is going to go up by one euro. The complete opposite I've just not seen, so this was really cool.   And then thinking about that going forwards, I think that type of architecture lends itself to providers to move away from this kind of just cost-driving, heavy legacy approach. I've heard you guys talking about the single API philosophy. Maybe someone could go into that in a little bit more detail.

Kevin Savina:  AWS is an infrastructure provider that's helping our customers in the media industry move forward and leverage technology to better serve their audiences and their subscribers.  And we do that by helping in multiple ways. 

One of the main things that we actually do is we have an opinion on what good architecture looks like. And when we're thinking about these large, siloed data problems, one of the recommended architecture patterns that we really try to encourage our customers and our partners to go for is leveraging cloud-native API-first composable architectures. Because that's what will allow the customers, in the end, to be agile, to move fast, to scale in a way that's aligned with their business. your example, they can reduce their operational cost and pass some of that on to their viewers, And so, one of the big things that we really do when we're focusing on the needs of the…media industry is come with this opinion of what good architecture looks like. And this API-first composable architecture is really key. 

The second thing is we, as a cloud provider, we provide services, in… within the AWS, infrastructure that align to these principles that help reduce operational cost. And the last thing that we do, because we recognize that it's not always easy for customers to plug all these, different solutions in a, composable architecture, it's kind of work upfront with partners, like you guys, to help pre-integrate these, to bring these operational and architecture patterns into execution across our, different partners in the AWS partner community to kind of pre-test, harden. Help scale and secure these integrations, so that we're kind of proactively de-risking that for the end customers, and it's easier to onboard on already a matureSet of, different components that talk to each other, that give you that end-to-end view of all of your data… data silos. And then that's when a Xroadmedia really fits in with… with what you're doing.

Tom Dvorak: exactly, and I think I would go one step further. I think there is no solution that has a future without being cloud nowadays. I mean, cloud offers so much, opportunity, and also So many benefits, which you just can't do, on bare metal or, on-premise nowadays.  I mean, not to mention what you just said before, TCO, time to market, scalability, those are the most prominent benefits.  But when you talk about operational challenges, those sometimes come with challenges, because what we see in our industry, particularly is, sometimes providers of their technology are bound to a certain, cloud provider, a certain infrastructure way, right?

 So this is what just Kevin said before, is kind of,  helping the customer understanding what the best architecture for their, service is, is probably more helpful than just, throwing over a cloud over the fence and saying, look, here's the service we offer, right? Because Then, they deploy everything in AWS, and then they say, but we have to run it in our own cloud, then you generate more traffic costs to increase latency.

So I think So we went one step further there, and we used this with a number of our clients, particularly with AWS, is where we say, we are actually cloud agnostic. We know AWS has very, very strong incentive programs for their customers as well. We realize that, strength is not so much in reselling infrastructure, we're not a cloud provider like AWS is. So why don't we allow our customers to work exactly with those and participate in the best way in those kind of incentive programs, and we just run it in their cloud. So, they open up an instance in AWS, and we run our servers on their cloud. And this, brings back, the solutions to the challenges. We say, we're running in exactly the same region, or regions, usually, because you're geo-redundant. You have no additional latency, etc, So I think that is, that is super important From my point of view, to add, maybe, to the operational side.

Kevin Savina: Actually, what you're describing is a pattern that we see with a lot of our larger customers, is this, willingness to really treat their data as strategic asset, have it in their environment, and bring all of the tools that they need from their different vendors, partners, integrators, to really work around that, but keep the high-level governance and strategic maintenance of that data large data sets as a strategic asset for them. So, and I mean, we recognize not everyone's going to do everything on AWS, obviously, but having a strategic governance from the customer's perspective on where all my data is, how it flows, how it's governed. And having them be, in complete control as a customer of that is really important.

Tom Dvorak: Let me maybe again, add to the point that Kevin just said, this is also for example, why we have deployed a one single standards-based API. So we just have one pipe if you like going to our solution, and we can hook up our solution with different ecosystem providers, be it like a monetization platform like Evergent, or be it like, a marketing automation tool, or a middleware, etc, 

But we have always the same API, and we can dock into those as we see fit and the advantage of those two things together, So being in the cloud and, having a single API is basically that our customers can roadmap their service and as Kevin said also before, it's kind of may be dealing with some legacy, and maybe have some parts running locally, or maybe in a different cloud. before, having 15 problems coming together into one, we can say, why don't we just, focus on one problem at a time, and implement with a solution, and then you can think about how you can optimize the implementation, rather than trying to do five things at once.

Craig Ferguson: It's really valid, and, that point that you were talking about, Tom, I'm really interested about the next steps beyond all this, having done all of this, avoiding this kind of siloed data approach and making sure that we have a really rich data library set. In your mind, what are some, like, the top two things that customer is going to be able to go forward and do with this type of data? For us, we've been really focusing this year on once the customer has that rich data set. what can they really use that for to go forward to present to customers? 

For us, it's giving, the right offer at the right time to prevent cancellation, or to bring more people on board, but in your minds, what do you think are the main things? And I think that we need to wrap up soon, but what are the main things that we can finish with in terms of that?

Tom Dvorak: I mean, from my point of view, as far as data is concerned, what we see very much is the siloed approach still, with some of our customers and I think where particularly also AI comes in is, it's becoming much easier to aggregate the data into one dataset with the help of AI. So, automating these kind of processes is at the forefront of everybody's work at the moment. 

So what we also do in our solution is, as I said before, we're predominantly, or actually exclusively in media and entertainment, but that does not limit it to one single, type of media. So it's usually, video-based, text-based, or audio-based services or maybe even games. So those are our, main kind of services we serve. And what we see there as well is, sometimes you have, like, one service for audio, one service for video, one service maybe for e-books or news articles and these kind of things. 

So we are trying to bring that together for our customers. So we have, for example, in Europe with a number of public broadcasters. What we have done is we have brought together, their three services, news articles.podcasts and also video content into one, not service, but into one experience. So basically, they started off with text-based and then, what we learned on the text-based side, we could automatically aggregate also then and apply to the other sites as well, even when the services were separated. So, focus on one problem at a time, is kind of try to build, the knowledge and the profile about the users across the different silos, and then start recommending in the other silos by what you've learned on the other side. So I think this is super interesting.

I mean, one of the big results we've seen there is, like, click-through rates on those media assets went up by a factor of 2.5, which is actually quite dramatic if you think about it, when you consume suddenly, more than twice as much content as you've done before. And we see, particularly on the data aggregation side, we don't see ourselves as kind of, the data aggregator per se, but we see, a lot of third-party services emerging there as well. AWS have a lot of those services as well. And the beauty, again, about, the cloud is that you can easily bring them together, because, you are in the same environment, and in a scalable environment.

Kevin Savina: Yeah, so you're correct, and there's a lot of tools on the AWS platform to help with that. But I'd like to widen the scope even a bit more. I mean, we've been talking about data, personalization, discovery, monetization. But I think what we're really talking about here is a component of a complete transformation of how media operations and digital services really operate as a business. and it's beyond this audience engagement and retention. It's about applying data-driven operations from content production all the way to generating revenue. And it does all start with good data hygiene. We've been about that quite a lot. 

But good data hygiene, it is an architectural and technical problem, and yes, we do have a bunch of AWS services, like AWS Glue, Clean Rooms, and reference architectures that will help with that. But its beyond that architectural problem, It's an organizational challenge for a lot of legacy, established companies, and you actually need a leadership push to change the way you're thinking about applying data, breaking these silos, applying data smartly in your workflows. And again, continuing to widen that scope. It's widening it beyond the… the audience user engagement, but also from the beginning of the production, and content preparation,  lifespan of content. 

So, I mean, some of the big… trends that have started in 2025, and I think wil, be accelerating in 2026. I see 3 big ones. 

The first one is hyper-personalization at scale. We mentioned that a few times. But how you generate, maintain, and personalize the right personalized content, whether that content is recommendations, or a summary, or a invitation to consume your services. at scale. And one way of doing that is really leveraging Agentic AI. 

And so, a second big trend that we're seeing is using agentic AI to drive content operations. And so that's where you start taking some of the editorial intent of your team or your different teams, and start amplifying that and scaling that out by Delegating some of the work of, you know, defining and fine-tuning the different segments, or producing some short versions of content.And start doing that at scale. One example, actually, is the NFL. It was quite interesting. They did that over the past few years. They were using 500 million data points, across historic footage from the NFL and they use that to restore historic content and propose a new viewing experience to their audiences using natural language search across all of that archive. And then deploying, kind of, AI assistance to generate content that is automatically generated by summarizing relevant content for a given person. So, that's generating a new way of personalizing the experience, which is Again, a different way of consuming content for the customer, a different way of leveraging smart, rich data, and delegating parts of the execution of a new service that they want to put in place to a set of AI agents that are leveraging that. 

Coming back to the question of the data models, what we were talking about at the beginning, diversification of these business models, going from, say, a traditional SVOD to AVOD and FAST and so on.how do you then leverage the data that you have to maximize the value of your advertising inventory? And again, we're seeing a lot of use of machine learning to kind of match the right segments to the right content that might have been generated just before, through personalization, through AI agents, to maximize the value of that inventory, and really, you know get across…revenues… a diversified revenue stream where you maximize each of them, avoid cannibalization. So, kind of the bottom line, I think what we're going to see is 2026 is that the media companies that are really going to perform well are the ones that can have personalized content that fits their audience, but that can do that at scale, deploying, the use of data, AI, agentic AI, not as a feature, but kind of rethinking the operating business, the operating Way of running their, their business.

Craig Ferguson: I love those points. I really do, and something that you really mentioned as well, Kevin, at the start of what you were just saying there, regarding there needs to be a management focus on doing the new things in our industry, right? And if I think about the part of our industry that's making the most dynamic shift, the most transformation, it's the sports space. Exactly like you mentioned there with the NFL, Kevin.  We see that also with our customers, someone like the NBA, who we've been working with now. We really see those guys as being, standout in our industry to say, we're going to try new things, we're going to try the new packages, we're going to try a different go-to-market approach that's maybe not been handled elsewhere. And I really think that the rest of the industry could stand to learn a lesson from these types of real agile organizations who are leaning into this Agentic AI, to these new offer types, to really meeting customers on their own terms, because I still feel as if there's a big part of our industry that thinks that they can just,, lean back and just rest on their walls regarding the approach that it's worked for so long.  I don't think that's gonna work for much longer, and we need to realize that it's a really saturated market, the customers' expectations have really never, ever been higher and being able to go to that customer and say, listen, we know you want to engage this way, we're going to allow that to happen. Sports are the part of our industry that I see really doing that. 

So, I think that management decision and that management approach could warrant a stand for warrant from that part.

Kevin Savina: Absolutely.

Tom Dvorak : But, I think, change management requires a sponsor in management, and that is change. It is how to implement different tools into your workflow, particularly with Agentic AI.  And there are a lot of myths also, in the industry about what AI isn't going to destroy, and these kind of things, but I think it's more to understand how can it help you. And what we've seen, by using AI for many, many years, but now also by implementing, Gen and agentic AI into some of our products. 

What we've seen is that particularly editorial teams, they now see the opportunity, how the kind of the boring stuff can be automated with AI, but kind of the quality, human quality control becomes more important now. So, their roles might have shifted slightly, but I think if done right and correctly, I think the importance of the roles are getting bigger and higher like they use to be before.

Kevin Savina:  I like the quote, and I don't remember who it comes from, so I'm sorry, but it's how you combine editorial work and AI. It's about, like, AI handles scale, editors handle trust. And working together, that's how you scale the trust to your audiences.