FAST has exploded in recent years, but where does it go from here? In this interview from the StreamTV Show, Frequency CRO Jon Cohen says we’re entering a transition phase—one where smarter data use and AI-powered scheduling will define which platforms thrive. While the initial FAST boom was about scale, Cohen says success now depends on improving viewer experience and monetization. With new content owners entering the space and better channel origination tools, the quality is rising—but so are the stakes.
Cohen also shares a glimpse into Frequency’s LLM-based AI scheduling agent, which is already driving double-digit engagement gains for early adopters. But he’s clear: AI won’t be a magic fix. Fragmented media data and legacy systems must evolve first. Still, the promise is real—smarter scheduling, optimized runtimes, and tailored programming could redefine how and when audiences tune in. “It’s not about removing humans,” Cohen notes, “but letting machines help them make better, faster decisions.” If engagement grows, monetization follows—and FAST proves it's here to stay.
Bevin Fletcher:
Hi everyone. Thanks for tuning in. I'm Bevin Fletcher, Senior Editor of StreamTV Insider, and we're here live at the StreamTV Show in Denver. I'm with my colleague, Daniel Frankel, co-founder of Next TMT and Jon Cohen, CRO of Frequency, and we're excited to be here to talk today. Hey guys.
Jon Cohen:
Thanks for having us.
Daniel Frankel:
Yeah. Jon, thanks for showing up. Bevin has a lot of questions about monetization she'll get into, but I want to ask you first about something I'm hearing here at the show about FAST in general, if you don't mind.
Jon Cohen:
Sure.
Daniel Frankel:
Indulge me. I'm hearing that one executive says it's plateauing you. We saw expansion of panels and channels for so long. What's going on? What's the state of growth for FAST at this point, do you think?
Jon Cohen:
I think we're certainly in a bit of a transitionary phase. I think if you look at the attendance at the show, there's still a lot of interest and business going on, which is important. I do think that after 2020 and the growth in FAST, it was remarkable what we were seeing. At this point I feel that you've got a lot of downward pressure on a lot of the media companies trying to figure out what to do with their rights and distribution.
This is a lucrative endeavor, distributing their content, but they have a lot of considerations. We're seeing new types of content owners coming into the space. The channel origination and how channels are being synthesized is getting a bit easier, and you're seeing more broadcast quality stuff. So I continue to be bullish, and the most exciting part of it is every time I think FAST is taking, I think the FAST acronym is a little tough because viewers don't understand what that is, but I do see that every time we make some advancements, we kind of take a step back as an industry because it's just growing faster than what everybody's expecting. So you have to kind of reset and figure out what you're going to do to improve the viewer experience and make this industry as lucrative as possible. I think it's a very exciting time to be in FAST.
Bevin Fletcher:
Awesome. If we could switch to product and tech for a minute, looking to the future of FAST, I mean, I think a lot of people are talking about AI potentially for linear scheduling. I know Frequency become a pretty key player in this space, and I believe you have an LLM-based AI agent of your own for linear scheduling. Could you tell me a little bit about that and what you see for the future of how that will play out?
Jon Cohen:
The agent is in development, and hopefully we'll make continuous, give updates to the market on that. But AI is not a straight line to the top. You can't just take AI, put it on top of your existing business and expect everything just to be better overnight. There are a lot of legacy parts of the media industry that are encumbered by things, and it starts with the data. The foundational data in a lot of the media is very fragmented, and the data plays a key role in the evolution and growth of AI. So there's some infrastructure and some foundational things that need to be put in place. Then we're going to start introducing the things that people are using in their everyday lives like, "I want an assistant. I want an analyst to help me look at this data and tell me things that I can do," and then say, go do it.
So you've got humans completely in the loop, you'll have humans on the loop, and there's a future phase, maybe humans out of the loop. But I don't think humans will ever be out of the loop of this operation because you're creating these entertaining viewing experiences for audiences and that's human to human. It can't all be machine, but machine can certainly help assist and make the process of creating these packages a lot easier, save a lot of money and time and get great product to the viewer. I do see it as a, it's definitely additive, and it's gonna be part of everyone's daily lives, including what we're doing in the media technology space.
Bevin Fletcher:
Interesting. Also, I'm curious, how do you see, not just on kind of the back end process-wise, but I feel like what I've heard is deepening engagement, kind of getting that habitual use, how do you see that impacting the user experience on FAST?
Jon Cohen:
I think that just going back to the legacy TV model. HBO is amazing. You could watch movies from multiple studios, you could watch original content. That original content got a lot better over time, Sunday night, and it was one-to-many HBO East, HBO West, just using it as an example, the multiplex. Given that we have internet dynamics and speed, the data and the usage primetime may not mean what primetime typically means. The Roku audience may be very different than the Amazon audience.
The data that we're getting in working with a bevy of partners on is really going to inform things like asset fatigue and when is too much, too much? When are you overplaying things? Does an audience in a certain season look for certain types of programming? Being able to take that data and make decisions on the fly and update a schedule and then say it's going out now and testing that in real time and seeing how engagement is impacted, hopefully in a positive way, will increase engagement. You'll have viewers watching longer, and if everyone's doing their jobs, you're gonna have more ads filled, which means more monetization for the whole marketplace. So I think it's the defining moment of the maturity and future profitability of this business and it being here to stay.
Daniel Frankel:
The future profitability.
Bevin Fletcher:
Yeah, I mean, the monetization piece is so key. Now, I do wonder, AI is obviously very hot and new, but is there any concern or how can you be confident if AI is kind of taking over roles that could actually impact what the viewer's seeing, like scheduling?
Jon Cohen:
We're already seeing it.
Bevin Fletcher:
You are?
Jon Cohen:
We are. We have a software platform, Frequency Studio, where a lot of channel orchestration and packaging and all the sausage-making of creating a channel happens. But behind the scenes, when we're getting data, we're analyzing that data, and we're providing insights to some key customers. We're already seeing 15, 20% jumps on engagement in a normal seasonal basis, and those channels are now outperforming other channels on some platforms. I have to be a little cryptic about what that is, but we are seeing it, and it's just a matter of time before we take the automation and all the analyze that's going on behind the scenes and bringing that natively into our software platform and to manage expectations. Building those things, it takes a little bit of time, and you have to do it right, but we want it tightly integrated because we have to worry about rights and avails.
You can only get so many runs. You don't want to have asset fatigue. You want the viewer to continuously be entertained with new content, and you want to do it the right way. So it's gonna take a little bit of time to get it from where it is right now as a back end automation into the front end experience where the programmer, is just saying, "Give me my insights. Would you like us to perform this and schedule the month of June?" That's not very far off, and that's gonna improve efficiency and hopefully the engagement, which we truly believe.
Bevin Fletcher:
Very cool. So also some content rights management in there too.
Jon Cohen:
Absolutely.
Bevin Fletcher:
Well, all right. We look forward to seeing how this all plays out, and thanks for being here.
Jon Cohen:
Thank you for having me.
Bevin Fletcher:
Thanks for watching.